Datasets:

librarian-bots

/

card_with_first_commit

like 0

AkshaySg/langid

audio-classification

--- tags: - ru - 4ulan --- Забавное для дискордика))00)) https://discord.gg/HpeadKH Offers work@4ulan.fun

Akuva2001/SocialGraph

--- language: - ru tags: - PyTorch - Transformers - 4ulan --- **Fork of https://huggingface.co/sberbank-ai/rugpt3large_based_on_gpt2** Забавное для дискордика))00)) ROADMAP: - Собираю датасетик из книжек про попаданцев. <------------------------- Сейчас тут. - Дообучаю. - Выбрасываю в дискордик. https://discord.gg/HpeadKH

AlbertHSU/BertTEST

--- language: - ru tags: - summarization - mbart datasets: - IlyaGusev/gazeta license: apache-2.0 inference: parameters: no_repeat_ngram_size: 4 widget: - text: "Высота башни составляет 324 метра (1063 фута), примерно такая же высота, как у 81-этажного здания, и самое высокое сооружение в Париже. Его основание квадратно, размером 125 метров (410 футов) с любой стороны. Во время строительства Эйфелева башня превзошла монумент Вашингтона, став самым высоким искусственным сооружением в мире, и этот титул она удерживала в течение 41 года до завершения строительство здания Крайслер в Нью-Йорке в 1930 году. Это первое сооружение которое достигло высоты 300 метров. Из-за добавления вещательной антенны на вершине башни в 1957 году она сейчас выше здания Крайслер на 5,2 метра (17 футов). За исключением передатчиков, Эйфелева башня является второй самой высокой отдельно стоящей структурой во Франции после виадука Мийо." example_title: "Википедия" - text: "С 1 сентября в России вступают в силу поправки в закон «О банкротстве» — теперь должники смогут освобождаться от непосильных обязательств во внесудебном порядке, если сумма задолженности составляет не менее 50 тыс. рублей и не превышает 500 тыс. рублей без учета штрафов, пени, процентов за просрочку платежа и прочих имущественных или финансовых санкций. У физлиц и индивидуальных предпринимателей появилась возможность пройти процедуру банкротства без участия суда и финансового управляющего — достаточно подать соответствующее заявление через МФЦ. Сумму задолженности и список всех известных заявителю кредиторов нужно предоставить самостоятельно. Если все условия соблюдены, сведения внесут в Единый федеральный реестр в течение трех рабочих дней. При этом на момент подачи заявления в отношении заявителя должно быть окончено исполнительное производство с возвращением исполнительного документа взыскателю. Это значит, что у потенциального банкрота не должно быть имущества, которое можно взыскать. Кроме того, в отношении гражданина не должно быть возбуждено другое исполнительное производство. В период всей процедуры заявитель не сможет брать займы, кредиты, выдавать поручительства, совершать иные обеспечительные сделки. Внесудебное банкротство будет длиться шесть месяцев, в течение которых также будет действовать мораторий на удовлетворение требований кредиторов, отмеченных в заявлении должника, и мораторий об уплате обязательных платежей. Кроме того, прекращается начисление неустоек и иных финансовых санкций; имущественные взыскания (кроме алиментов) также будут приостановлены. По завершению процедуры заявителя освободят от дальнейшего выполнения требований кредиторов, указанных в заявлении о признании его банкротом, а эта задолженность признается безнадежной. В прошлом месяце стало известно, что за первое полугодие 2020 года российские суды признали банкротами 42,7 тыс. граждан (в том числе индивидуальных предпринимателей) — по данным единого реестра «Федресурс», это на 47,2% больше показателя аналогичного периода 2019 года. Рост числа обанкротившихся граждан во втором квартале по сравнению с первым замедлился — такая динамика обусловлена тем, что в период ограничений с 19 марта по 11 мая суды редко рассматривали банкротные дела компаний и меньше, чем обычно, в отношении граждан, объяснял руководитель проекта «Федресурс» Алексей Юхнин. Он прогнозирует, что во втором полугодии мы увидим рост показателя, когда суды рассмотрят все дела, что не смогли ранее в режиме ограничений. По его данным, уже в июне число личных банкротств выросло до 11,5 тыс., что в два раза превышает показатель аналогичного периода 2019 года." example_title: "Новости" - text: "Актуальность проблемы. Электронная информация играет все большую роль во всех сферах жизни современного общества. В последние годы объем научно-технической текстовой информации в электронном виде возрос настолько, что возникает угроза обесценивания этой информации в связи с трудностями поиска необходимых сведений среди множества доступных текстов. Развитие информационных ресурсов Интернет многократно усугубило проблему информационной перегрузки. В этой ситуации особенно актуальными становятся методы автоматизации реферирования текстовой информации, то есть методы получения сжатого представления текстовых документов–рефератов (аннотаций). Постановка проблемы автоматического реферирования текста и соответственно попытки ее решения с использованием различных подходов предпринимались многими исследователями. История применения вычислительной техники для реферирования насчитывает уже более 50 лет и связана с именами таких исследователей, как Г.П. Лун, В.Е. Берзон, И.П. Cевбо, Э.Ф. Скороходько, Д.Г. Лахути, Р.Г. Пиотровский и др. За эти годы выработаны многочисленные подходы к решению данной проблемы, которые достаточно четко подразделяются на два направления: автоматическое реферирование, основанное на экстрагировании из первичных документов с помощью определенных формальных признаков «наиболее информативных» фраз (фрагментов), совокупность которых образует некоторый экстракт; автоматическое реферирование, основанное на выделении из текстов с помощью специальных информационных языков наиболее существенной информации и порождении новых текстов (рефератов), содержательно обобщающих первичные документы." example_title: "Научная статья" --- # MBARTRuSumGazeta ## Model description This is a ported version of [fairseq model](https://www.dropbox.com/s/fijtntnifbt9h0k/gazeta_mbart_v2_fairseq.tar.gz). For more details, please see [Dataset for Automatic Summarization of Russian News](https://arxiv.org/abs/2006.11063). ## Intended uses & limitations #### How to use Colab: [link](https://colab.research.google.com/drive/1wdo_nPZPk6dWAn1J8nGx4Z5Ef82jCCob) ```python from transformers import MBartTokenizer, MBartForConditionalGeneration model_name = "IlyaGusev/mbart_ru_sum_gazeta" tokenizer = MBartTokenizer.from_pretrained(model_name) model = MBartForConditionalGeneration.from_pretrained(model_name) article_text = "..." input_ids = tokenizer( [article_text], max_length=600, padding="max_length", truncation=True, return_tensors="pt", )["input_ids"] output_ids = model.generate( input_ids=input_ids, no_repeat_ngram_size=4 )[0] summary = tokenizer.decode(output_ids, skip_special_tokens=True) print(summary) ``` #### Limitations and bias - The model should work well with Gazeta.ru articles, but for any other agencies it can suffer from domain shift ## Training data - Dataset: [Gazeta](https://huggingface.co/datasets/IlyaGusev/gazeta) ## Training procedure - Fairseq training script: [train.sh](https://github.com/IlyaGusev/summarus/blob/master/external/bart_scripts/train.sh) - Porting: [Colab link](https://colab.research.google.com/drive/13jXOlCpArV-lm4jZQ0VgOpj6nFBYrLAr) ## Eval results * Train dataset: **Gazeta v1 train** * Test dataset: **Gazeta v1 test** * Source max_length: **600** * Target max_length: **200** * no_repeat_ngram_size: **4** * num_beams: **5** | Model | R-1-f | R-2-f | R-L-f | chrF | METEOR | BLEU | Avg char length | |:--------------------------|:------|:------|:------|:-------|:-------|:-----|:-----| | [mbart_ru_sum_gazeta](https://huggingface.co/IlyaGusev/mbart_ru_sum_gazeta) | **32.4** | 14.3 | 28.0 | 39.7 | **26.4** | 12.1 | 371 | | [rut5_base_sum_gazeta](https://huggingface.co/IlyaGusev/rut5_base_sum_gazeta) | 32.2 | **14.4** | **28.1** | **39.8** | 25.7 | **12.3** | 330 | | [rugpt3medium_sum_gazeta](https://huggingface.co/IlyaGusev/rugpt3medium_sum_gazeta) | 26.2 | 7.7 | 21.7 | 33.8 | 18.2 | 4.3 | 244 | * Train dataset: **Gazeta v1 train** * Test dataset: **Gazeta v2 test** * Source max_length: **600** * Target max_length: **200** * no_repeat_ngram_size: **4** * num_beams: **5** | Model | R-1-f | R-2-f | R-L-f | chrF | METEOR | BLEU | Avg char length | |:--------------------------|:------|:------|:------|:-------|:-------|:-----|:-----| | [mbart_ru_sum_gazeta](https://huggingface.co/IlyaGusev/mbart_ru_sum_gazeta) | **28.7** | **11.1** | 24.4 | **37.3** | **22.7** | **9.4** | 373 | | [rut5_base_sum_gazeta](https://huggingface.co/IlyaGusev/rut5_base_sum_gazeta) | 28.6 | **11.1** | **24.5** | 37.2 | 22.0 | **9.4** | 331 | | [rugpt3medium_sum_gazeta](https://huggingface.co/IlyaGusev/rugpt3medium_sum_gazeta) | 24.1 | 6.5 | 19.8 | 32.1 | 16.3 | 3.6 | 242 | Predicting all summaries: ```python import json import torch from transformers import MBartTokenizer, MBartForConditionalGeneration from datasets import load_dataset def gen_batch(inputs, batch_size): batch_start = 0 while batch_start < len(inputs): yield inputs[batch_start: batch_start + batch_size] batch_start += batch_size def predict( model_name, input_records, output_file, max_source_tokens_count=600, batch_size=4 ): device = "cuda" if torch.cuda.is_available() else "cpu" tokenizer = MBartTokenizer.from_pretrained(model_name) model = MBartForConditionalGeneration.from_pretrained(model_name).to(device) predictions = [] for batch in gen_batch(inputs, batch_size): texts = [r["text"] for r in batch] input_ids = tokenizer( batch, return_tensors="pt", padding="max_length", truncation=True, max_length=max_source_tokens_count )["input_ids"].to(device) output_ids = model.generate( input_ids=input_ids, no_repeat_ngram_size=4 ) summaries = tokenizer.batch_decode(output_ids, skip_special_tokens=True) for s in summaries: print(s) predictions.extend(summaries) with open(output_file, "w") as w: for p in predictions: w.write(p.strip().replace("\n", " ") + "\n") gazeta_test = load_dataset('IlyaGusev/gazeta', script_version="v1.0")["test"] predict("IlyaGusev/mbart_ru_sum_gazeta", list(gazeta_test), "mbart_predictions.txt") ``` Evaluation: https://github.com/IlyaGusev/summarus/blob/master/evaluate.py Flags: --language ru --tokenize-after --lower ### BibTeX entry and citation info ```bibtex @InProceedings{10.1007/978-3-030-59082-6_9, author="Gusev, Ilya", editor="Filchenkov, Andrey and Kauttonen, Janne and Pivovarova, Lidia", title="Dataset for Automatic Summarization of Russian News", booktitle="Artificial Intelligence and Natural Language", year="2020", publisher="Springer International Publishing", address="Cham", pages="122--134", isbn="978-3-030-59082-6" } ```

AlbertHSU/ChineseFoodBert

feature-extraction

--- language: - ru license: apache-2.0 --- # NewsTgRuBERT Training script: https://github.com/dialogue-evaluation/Russian-News-Clustering-and-Headline-Generation/blob/main/train_mlm.py

Alberto15Romero/GptNeo

--- language: - ru tags: - summarization - token-classification - t5 datasets: - IlyaGusev/gazeta license: apache-2.0 inference: false widget: - text: "С 1 сентября в России вступают в силу поправки в закон «О банкротстве» — теперь должники смогут освобождаться от непосильных обязательств во внесудебном порядке, если сумма задолженности составляет не менее 50 тыс. рублей и не превышает 500 тыс. рублей без учета штрафов, пени, процентов за просрочку платежа и прочих имущественных или финансовых санкций.[SEP]У физлиц и индивидуальных предпринимателей появилась возможность пройти процедуру банкротства без участия суда и финансового управляющего — достаточно подать соответствующее заявление через МФЦ.[SEP]Сумму задолженности и список всех известных заявителю кредиторов нужно предоставить самостоятельно.[SEP]Если все условия соблюдены, сведения внесут в Единый федеральный реестр в течение трех рабочих дней.[SEP]При этом на момент подачи заявления в отношении заявителя должно быть окончено исполнительное производство с возвращением исполнительного документа взыскателю.[SEP]Это значит, что у потенциального банкрота не должно быть имущества, которое можно взыскать.[SEP]Кроме того, в отношении гражданина не должно быть возбуждено другое исполнительное производство.[SEP]В период всей процедуры заявитель не сможет брать займы, кредиты, выдавать поручительства, совершать иные обеспечительные сделки.[SEP]Внесудебное банкротство будет длиться шесть месяцев, в течение которых также будет действовать мораторий на удовлетворение требований кредиторов, отмеченных в заявлении должника, и мораторий об уплате обязательных платежей.[SEP]Кроме того, прекращается начисление неустоек и иных финансовых санкций; имущественные взыскания (кроме алиментов) также будут приостановлены.[SEP]По завершению процедуры заявителя освободят от дальнейшего выполнения требований кредиторов, указанных в заявлении о признании его банкротом, а эта задолженность признается безнадежной.[SEP]В прошлом месяце стало известно, что за первое полугодие 2020 года российские суды признали банкротами 42,7 тыс. граждан (в том числе индивидуальных предпринимателей) — по данным единого реестра «Федресурс», это на 47,2% больше показателя аналогичного периода 2019 года.[SEP]Рост числа обанкротившихся граждан во втором квартале по сравнению с первым замедлился — такая динамика обусловлена тем, что в период ограничений с 19 марта по 11 мая суды редко рассматривали банкротные дела компаний и меньше, чем обычно, в отношении граждан, объяснял руководитель проекта «Федресурс» Алексей Юхнин.[SEP]" example_title: "Новости" --- # RuBERTExtSumGazeta ## Model description Model for extractive summarization based on [rubert-base-cased](DeepPavlov/rubert-base-cased) ## Intended uses & limitations #### How to use Colab: [link](https://colab.research.google.com/drive/1Q8_v3H-kxdJhZIiyLYat7Kj02qDq7M1L) ```python import razdel from transformers import AutoTokenizer, BertForTokenClassification model_name = "IlyaGusev/rubert_ext_sum_gazeta" tokenizer = AutoTokenizer.from_pretrained(model_name) sep_token = tokenizer.sep_token sep_token_id = tokenizer.sep_token_id model = BertForTokenClassification.from_pretrained(model_name) article_text = "..." sentences = [s.text for s in razdel.sentenize(article_text)] article_text = sep_token.join(sentences) inputs = tokenizer( [article_text], max_length=500, padding="max_length", truncation=True, return_tensors="pt", ) sep_mask = inputs["input_ids"][0] == sep_token_id # Fix token_type_ids current_token_type_id = 0 for pos, input_id in enumerate(inputs["input_ids"][0]): inputs["token_type_ids"][0][pos] = current_token_type_id if input_id == sep_token_id: current_token_type_id = 1 - current_token_type_id # Infer model with torch.no_grad(): outputs = model(**inputs) logits = outputs.logits[0, :, 1] # Choose sentences logits = logits[sep_mask] logits, indices = logits.sort(descending=True) logits, indices = logits.cpu().tolist(), indices.cpu().tolist() pairs = list(zip(logits, indices)) pairs = pairs[:3] indices = list(sorted([idx for _, idx in pairs])) summary = " ".join([sentences[idx] for idx in indices]) print(summary) ``` #### Limitations and bias - The model should work well with Gazeta.ru articles, but for any other agencies it can suffer from domain shift ## Training data - Dataset: [Gazeta](https://huggingface.co/datasets/IlyaGusev/gazeta) ## Training procedure TBD ## Eval results TBD Evaluation: https://github.com/IlyaGusev/summarus/blob/master/evaluate.py Flags: --language ru --tokenize-after --lower

AlchemistDude/DialoGPT-medium-Gon

--- language: - ru tags: - summarization license: apache-2.0 inference: parameters: no_repeat_ngram_size: 4 --- # RuBertTelegramHeadlines ## Model description Example model for [Headline generation competition](https://competitions.codalab.org/competitions/29905) Based on [RuBERT](http://docs.deeppavlov.ai/en/master/features/models/bert.html) model ## Intended uses & limitations #### How to use ```python from transformers import AutoTokenizer, EncoderDecoderModel model_name = "IlyaGusev/rubert_telegram_headlines" tokenizer = AutoTokenizer.from_pretrained(model_name, do_lower_case=False, do_basic_tokenize=False, strip_accents=False) model = EncoderDecoderModel.from_pretrained(model_name) article_text = "..." input_ids = tokenizer( [article_text], add_special_tokens=True, max_length=256, padding="max_length", truncation=True, return_tensors="pt", )["input_ids"] output_ids = model.generate( input_ids=input_ids, max_length=64, no_repeat_ngram_size=3, num_beams=10, top_p=0.95 )[0] headline = tokenizer.decode(output_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True) print(headline) ``` ## Training data - Dataset: [ru_all_split.tar.gz](https://www.dropbox.com/s/ykqk49a8avlmnaf/ru_all_split.tar.gz) ## Training procedure ```python import random import torch from torch.utils.data import Dataset from tqdm.notebook import tqdm from transformers import BertTokenizer, EncoderDecoderModel, Trainer, TrainingArguments, logging def convert_to_tensors( tokenizer, text, max_text_tokens_count, max_title_tokens_count = None, title = None ): inputs = tokenizer( text, add_special_tokens=True, max_length=max_text_tokens_count, padding="max_length", truncation=True ) result = { "input_ids": torch.tensor(inputs["input_ids"]), "attention_mask": torch.tensor(inputs["attention_mask"]), } if title is not None: outputs = tokenizer( title, add_special_tokens=True, max_length=max_title_tokens_count, padding="max_length", truncation=True ) decoder_input_ids = torch.tensor(outputs["input_ids"]) decoder_attention_mask = torch.tensor(outputs["attention_mask"]) labels = decoder_input_ids.clone() labels[decoder_attention_mask == 0] = -100 result.update({ "labels": labels, "decoder_input_ids": decoder_input_ids, "decoder_attention_mask": decoder_attention_mask }) return result class GetTitleDataset(Dataset): def __init__( self, original_records, sample_rate, tokenizer, max_text_tokens_count, max_title_tokens_count ): self.original_records = original_records self.sample_rate = sample_rate self.tokenizer = tokenizer self.max_text_tokens_count = max_text_tokens_count self.max_title_tokens_count = max_title_tokens_count self.records = [] for record in tqdm(original_records): if random.random() > self.sample_rate: continue tensors = convert_to_tensors( tokenizer=tokenizer, title=record["title"], text=record["text"], max_title_tokens_count=self.max_title_tokens_count, max_text_tokens_count=self.max_text_tokens_count ) self.records.append(tensors) def __len__(self): return len(self.records) def __getitem__(self, index): return self.records[index] def train( train_records, val_records, pretrained_model_path, train_sample_rate=1.0, val_sample_rate=1.0, output_model_path="models", checkpoint=None, max_text_tokens_count=256, max_title_tokens_count=64, batch_size=8, logging_steps=1000, eval_steps=10000, save_steps=10000, learning_rate=0.00003, warmup_steps=2000, num_train_epochs=3 ): logging.set_verbosity_info() tokenizer = BertTokenizer.from_pretrained( pretrained_model_path, do_lower_case=False, do_basic_tokenize=False, strip_accents=False ) train_dataset = GetTitleDataset( train_records, train_sample_rate, tokenizer, max_text_tokens_count=max_text_tokens_count, max_title_tokens_count=max_title_tokens_count ) val_dataset = GetTitleDataset( val_records, val_sample_rate, tokenizer, max_text_tokens_count=max_text_tokens_count, max_title_tokens_count=max_title_tokens_count ) model = EncoderDecoderModel.from_encoder_decoder_pretrained(pretrained_model_path, pretrained_model_path) training_args = TrainingArguments( output_dir=output_model_path, per_device_train_batch_size=batch_size, per_device_eval_batch_size=batch_size, do_train=True, do_eval=True, overwrite_output_dir=False, logging_steps=logging_steps, eval_steps=eval_steps, evaluation_strategy="steps", save_steps=save_steps, learning_rate=learning_rate, warmup_steps=warmup_steps, num_train_epochs=num_train_epochs, max_steps=-1, save_total_limit=1, ) trainer = Trainer( model=model, args=training_args, train_dataset=train_dataset, eval_dataset=val_dataset ) trainer.train(checkpoint) model.save_pretrained(output_model_path) ```

Ale/Alen

--- language: - ru tags: - text-classification license: apache-2.0 --- # RuBERTConv Toxic Classifier ## Model description Based on [rubert-base-cased-conversational](https://huggingface.co/DeepPavlov/rubert-base-cased-conversational) model ## Intended uses & limitations #### How to use Colab: [link](https://colab.research.google.com/drive/1veKO9hke7myxKigZtZho_F-UM2fD9kp8) ```python from transformers import pipeline model_name = "IlyaGusev/rubertconv_toxic_clf" pipe = pipeline("text-classification", model=model_name, tokenizer=model_name, framework="pt") text = "Ты придурок из интернета" pipe([text]) ``` ## Training data Datasets: - [2ch]( https://www.kaggle.com/blackmoon/russian-language-toxic-comments) - [Odnoklassniki](https://www.kaggle.com/alexandersemiletov/toxic-russian-comments) - [Toloka Persona Chat Rus](https://toloka.ai/ru/datasets) - [Koziev's Conversations](https://github.com/Koziev/NLP_Datasets/blob/master/Conversations/Data) with [toxic words vocabulary](https://www.dropbox.com/s/ou6lx03b10yhrfl/bad_vocab.txt.tar.gz) Augmentations: - ё -> е - Remove or add "?" or "!" - Fix CAPS - Concatenate toxic and non-toxic texts - Concatenate two non-toxic texts - Add toxic words from vocabulary - Add typos - Mask toxic words with "*", "@", "$" ## Training procedure TBA

Aleenbo/Arcane

--- language: - ru tags: - token-classification license: apache-2.0 widget: - text: Ёпта, меня зовут придурок и я живу в жопе --- # RuBERTConv Toxic Editor ## Model description Tagging model for detoxification based on [rubert-base-cased-conversational](https://huggingface.co/DeepPavlov/rubert-base-cased-conversational). 4 possible classes: - Equal = save tokens - Replace = replace tokens with mask - Delete = remove tokens - Insert = insert mask before tokens Use in pair with [mask filler](https://huggingface.co/IlyaGusev/sber_rut5_filler). ## Intended uses & limitations #### How to use Colab: [link](https://colab.research.google.com/drive/1NUSO1QGlDgD-IWXa2SpeND089eVxrCJW) ```python import torch from transformers import AutoTokenizer, pipeline tagger_model_name = "IlyaGusev/rubertconv_toxic_editor" device = "cuda" if torch.cuda.is_available() else "cpu" device_num = 0 if device == "cuda" else -1 tagger_pipe = pipeline( "token-classification", model=tagger_model_name, tokenizer=tagger_model_name, framework="pt", device=device_num, aggregation_strategy="max" ) text = "..." tagger_predictions = tagger_pipe([text], batch_size=1) sample_predictions = tagger_predictions[0] print(sample_predictions) ``` ## Training data - Dataset: [russe_detox_2022](https://github.com/skoltech-nlp/russe_detox_2022/tree/main/data) ## Training procedure - Parallel corpus convertion: [compute_tags.py](https://github.com/IlyaGusev/rudetox/blob/main/rudetox/marker/compute_tags.py) - Training script: [train.py](https://github.com/IlyaGusev/rudetox/blob/main/rudetox/marker/train.py) - Pipeline step: [dvc.yaml, train_marker](https://github.com/IlyaGusev/rudetox/blob/main/dvc.yaml#L367) ## Eval results TBA

Aleksandar/bert-srb-base-cased-oscar

fill-mask

--- language: - ru tags: - causal-lm - summarization datasets: - IlyaGusev/gazeta license: - apache-2.0 inference: false widget: - text: "Высота башни составляет 324 метра (1063 фута), примерно такая же высота, как у 81-этажного здания, и самое высокое сооружение в Париже. Его основание квадратно, размером 125 метров (410 футов) с любой стороны. Во время строительства Эйфелева башня превзошла монумент Вашингтона, став самым высоким искусственным сооружением в мире, и этот титул она удерживала в течение 41 года до завершения строительство здания Крайслер в Нью-Йорке в 1930 году. Это первое сооружение которое достигло высоты 300 метров. Из-за добавления вещательной антенны на вершине башни в 1957 году она сейчас выше здания Крайслер на 5,2 метра (17 футов). За исключением передатчиков, Эйфелева башня является второй самой высокой отдельно стоящей структурой во Франции после виадука Мийо.<s>" example_title: "Википедия" --- # RuGPT3MediumSumGazeta ## Model description This is the model for abstractive summarization for Russian based on [rugpt3medium_based_on_gpt2](https://huggingface.co/sberbank-ai/rugpt3medium_based_on_gpt2). ## Intended uses & limitations #### How to use Colab: [link](https://colab.research.google.com/drive/1eR-ev0Y5ISWIwGnzYYoHyGMaSIUz8GTN) ```python import torch from transformers import AutoTokenizer, AutoModelForCausalLM model_name = "IlyaGusev/rugpt3medium_sum_gazeta" tokenizer = AutoTokenizer.from_pretrained(model_name) model = AutoModelForCausalLM.from_pretrained(model_name) article_text = "..." text_tokens = tokenizer( article_text, max_length=600, add_special_tokens=False, padding=False, truncation=True )["input_ids"] input_ids = text_tokens + [tokenizer.sep_token_id] input_ids = torch.LongTensor([input_ids]) output_ids = model.generate( input_ids=input_ids, no_repeat_ngram_size=4 ) summary = tokenizer.decode(output_ids[0], skip_special_tokens=False) summary = summary.split(tokenizer.sep_token)[1] summary = summary.split(tokenizer.eos_token)[0] print(summary) ``` ## Training data - Dataset: [Gazeta](https://huggingface.co/datasets/IlyaGusev/gazeta) ## Training procedure - Training script: [train.py](https://github.com/IlyaGusev/summarus/blob/master/external/hf_scripts/train.py) - Config: [gpt_training_config.json](https://github.com/IlyaGusev/summarus/blob/master/external/hf_scripts/configs/gpt_training_config.json) ## Eval results * Train dataset: **Gazeta v1 train** * Test dataset: **Gazeta v1 test** * Source max_length: **600** * Target max_length: **200** * no_repeat_ngram_size: **4** * num_beams: **5** | Model | R-1-f | R-2-f | R-L-f | chrF | METEOR | BLEU | Avg char length | |:--------------------------|:------|:------|:------|:-------|:-------|:-----|:-----| | [mbart_ru_sum_gazeta](https://huggingface.co/IlyaGusev/mbart_ru_sum_gazeta) | **32.4** | 14.3 | 28.0 | 39.7 | **26.4** | 12.1 | 371 | | [rut5_base_sum_gazeta](https://huggingface.co/IlyaGusev/rut5_base_sum_gazeta) | 32.2 | **14.4** | **28.1** | **39.8** | 25.7 | **12.3** | 330 | | [rugpt3medium_sum_gazeta](https://huggingface.co/IlyaGusev/rugpt3medium_sum_gazeta) | 26.2 | 7.7 | 21.7 | 33.8 | 18.2 | 4.3 | 244 | * Train dataset: **Gazeta v1 train** * Test dataset: **Gazeta v2 test** * Source max_length: **600** * Target max_length: **200** * no_repeat_ngram_size: **4** * num_beams: **5** | Model | R-1-f | R-2-f | R-L-f | chrF | METEOR | BLEU | Avg char length | |:--------------------------|:------|:------|:------|:-------|:-------|:-----|:-----| | [mbart_ru_sum_gazeta](https://huggingface.co/IlyaGusev/mbart_ru_sum_gazeta) | **28.7** | **11.1** | 24.4 | **37.3** | **22.7** | **9.4** | 373 | | [rut5_base_sum_gazeta](https://huggingface.co/IlyaGusev/rut5_base_sum_gazeta) | 28.6 | **11.1** | **24.5** | 37.2 | 22.0 | **9.4** | 331 | | [rugpt3medium_sum_gazeta](https://huggingface.co/IlyaGusev/rugpt3medium_sum_gazeta) | 24.1 | 6.5 | 19.8 | 32.1 | 16.3 | 3.6 | 242 | Evaluation script: [evaluate.py](https://github.com/IlyaGusev/summarus/blob/master/evaluate.py) Flags: --language ru --tokenize-after --lower

Aleksandar/bert-srb-ner-setimes-lr

--- language: - ru tags: - summarization license: apache-2.0 widget: - text: "Комиссия Совета Федерации по информационной политике и взаимодействию со СМИ совместно с заинтересованными ведомствами думает над разработкой национального законодательства в области налогообложения глобальных интернет-компаний, таких как Google и Facebook. Об этом сообщил ТАСС председатель комиссии Алексей Пушков. «В настоящее время по линии ОЭСР [Организация экономического сотрудничества и развития] ведется разработка международной конвенции, однако работа над ней еще не завершена. В этих условиях мы исходим из того, что самая разумная позиция - начать разработку национального законодательства, не дожидаясь конвенции», — пояснил сенатор. Пушков отметил, что по такому пути пошли еще несколько стран, в числе которых Франция, Австралия и Турция. По его словам, в России важно задействовать в этой работе Минфин, ФНС, МИД РФ и Роскомнадзор. «Интернет-платформы не фигурируют у нас сейчас как отдельный объект налогообложения. Когда они откроют в России свои представительства в рамках закона о «приземлении», возникнет вопрос: как их официальное присутствие на территории России, которого сейчас нет, будет соотноситься с нашим налоговым режимом. Мы сейчас продумываем, как установить эту взаимосвязь», — сказал Пушков, добавляя, что вопрос внесения изменений в российское законодательство в части налогообложения крупных IT-компаний находится «на первой стадии изучения». Сам сенатор выступает за введение прогрессивной ставки налога в зависимости от прибыли IT-компаний на территории страны. При этом, подчеркнул он, одна из задач национальной системы налогообложения будет заключаться в подсчете налогооблагаемой базы. Сейчас крупные ИТ-компании самостоятельно отчитываются о своей прибыли. Однако России нужна собственная система подсчета их доходов, которая позволит определить их «реальную налогооблагаемую базу», считает Пушков. (https://www.gazeta.ru/tech/news/2021/12/17/n_17024239.shtml)" example_title: "Новость про налоги в IT" - text: "Первую многоножку, у которой более тысячи ног, обнаружили в австралийских пещерах биологи, изучавшие там подземные воды. Предыдущей рекордсменкой по количеству ног была 700-ногая многоножка. Новый вид имеет длинное тонкое тело, похожее на нить, и большое количество конечностей, по-видимому, дает преимущества для быстрого перемещения и проникновения в труднодоступные места — ученые полагают, такая многоножка может спокойно перемещаться по трещинам в камнях. Австралия известна своими огромными и жутковатыми животными вроде 25-сантиметровых пауков. Теперь список пугающих членистоногих пополнился самой «многоногой» в мире многоножкой, у которой более тысячи ног. Необычное животное обнаружила группа исследователей из Австралии и США в пещерах на западе страны. Подробнее многоножку ученые описали в статье в журнале Scientific Reports. Исследователи занимались оценкой воздействия подземных вод на окружающую среду в зоне добычи полезных ископаемых на западе страны, когда наткнулись на новый вид многоножек. В отличие от большинства сородичей, живущих на поверхности, эти многоножки обитали в пещерах на глубине до 60 метров. Новый вид исследователи назвали Eumillipes persephone, в честь Персефоны — древнегреческой богини подземного мира. У многоножки оказалось 1306 ног — больше, чем у любого другого известного вида. Предыдущей рекордсменкой была калифорнийская Illacme plenipes, у которой насчитывалось до 750 ног. «Эти животные были настолько уникальны, — говорит биолог Бруно Бузатто. — Как только я понял, какой длины они были... Стало ясно, что это что-то совершенно новое». У Е. persephone нитевидное тело длиной около 9,5 см и шириной всего миллиметр, состоящее из 330 сегментов, короткие ноги и конусообразная голова. Как и другие животные, живущие в постоянной темноте, эти многоножки бледны и слепы. Энтомолог Пол Марек сравнивает ее с белой нитью, выдернутой из рубашки. Чтобы посчитать количество ног, ученым пришлось сначала снять многоножку в высоком разрешении, а затем закрашивать на фото каждый десяток ног другим цветом. (https://www.gazeta.ru/science/2021/12/17_a_14325355.shtml)" example_title: "Новость про многоножку" - text: "Высота башни составляет 324 метра (1063 фута), примерно такая же высота, как у 81-этажного здания, и самое высокое сооружение в Париже. Его основание квадратно, размером 125 метров (410 футов) с любой стороны. Во время строительства Эйфелева башня превзошла монумент Вашингтона, став самым высоким искусственным сооружением в мире, и этот титул она удерживала в течение 41 года до завершения строительство здания Крайслер в Нью-Йорке в 1930 году. Это первое сооружение которое достигло высоты 300 метров. Из-за добавления вещательной антенны на вершине башни в 1957 году она сейчас выше здания Крайслер на 5,2 метра (17 футов). За исключением передатчиков, Эйфелева башня является второй самой высокой отдельно стоящей структурой во Франции после виадука Мийо." example_title: "Википедия" --- # RuT5TelegramHeadlines ## Model description Based on [rut5-base](https://huggingface.co/cointegrated/rut5-base) model ## Intended uses & limitations #### How to use ```python from transformers import AutoTokenizer, T5ForConditionalGeneration model_name = "IlyaGusev/rut5_base_headline_gen_telegram" tokenizer = AutoTokenizer.from_pretrained(model_name) model = T5ForConditionalGeneration.from_pretrained(model_name) article_text = "..." input_ids = tokenizer( [article_text], max_length=600, add_special_tokens=True, padding="max_length", truncation=True, return_tensors="pt" )["input_ids"] output_ids = model.generate( input_ids=input_ids )[0] headline = tokenizer.decode(output_ids, skip_special_tokens=True) print(headline) ``` ## Training data - Dataset: [ru_all_split.tar.gz](https://www.dropbox.com/s/ykqk49a8avlmnaf/ru_all_split.tar.gz) ## Training procedure - Training script: [train.py](https://github.com/IlyaGusev/summarus/blob/master/external/hf_scripts/train.py)

Aleksandar/bert-srb-ner-setimes

token-classification

--- language: - ru tags: - summarization - t5 datasets: - IlyaGusev/gazeta license: - apache-2.0 inference: parameters: no_repeat_ngram_size: 4 widget: - text: "Высота башни составляет 324 метра (1063 фута), примерно такая же высота, как у 81-этажного здания, и самое высокое сооружение в Париже. Его основание квадратно, размером 125 метров (410 футов) с любой стороны. Во время строительства Эйфелева башня превзошла монумент Вашингтона, став самым высоким искусственным сооружением в мире, и этот титул она удерживала в течение 41 года до завершения строительство здания Крайслер в Нью-Йорке в 1930 году. Это первое сооружение которое достигло высоты 300 метров. Из-за добавления вещательной антенны на вершине башни в 1957 году она сейчас выше здания Крайслер на 5,2 метра (17 футов). За исключением передатчиков, Эйфелева башня является второй самой высокой отдельно стоящей структурой во Франции после виадука Мийо." example_title: "Википедия" - text: "С 1 сентября в России вступают в силу поправки в закон «О банкротстве» — теперь должники смогут освобождаться от непосильных обязательств во внесудебном порядке, если сумма задолженности составляет не менее 50 тыс. рублей и не превышает 500 тыс. рублей без учета штрафов, пени, процентов за просрочку платежа и прочих имущественных или финансовых санкций. У физлиц и индивидуальных предпринимателей появилась возможность пройти процедуру банкротства без участия суда и финансового управляющего — достаточно подать соответствующее заявление через МФЦ. Сумму задолженности и список всех известных заявителю кредиторов нужно предоставить самостоятельно. Если все условия соблюдены, сведения внесут в Единый федеральный реестр в течение трех рабочих дней. При этом на момент подачи заявления в отношении заявителя должно быть окончено исполнительное производство с возвращением исполнительного документа взыскателю. Это значит, что у потенциального банкрота не должно быть имущества, которое можно взыскать. Кроме того, в отношении гражданина не должно быть возбуждено другое исполнительное производство. В период всей процедуры заявитель не сможет брать займы, кредиты, выдавать поручительства, совершать иные обеспечительные сделки. Внесудебное банкротство будет длиться шесть месяцев, в течение которых также будет действовать мораторий на удовлетворение требований кредиторов, отмеченных в заявлении должника, и мораторий об уплате обязательных платежей. Кроме того, прекращается начисление неустоек и иных финансовых санкций; имущественные взыскания (кроме алиментов) также будут приостановлены. По завершению процедуры заявителя освободят от дальнейшего выполнения требований кредиторов, указанных в заявлении о признании его банкротом, а эта задолженность признается безнадежной. В прошлом месяце стало известно, что за первое полугодие 2020 года российские суды признали банкротами 42,7 тыс. граждан (в том числе индивидуальных предпринимателей) — по данным единого реестра «Федресурс», это на 47,2% больше показателя аналогичного периода 2019 года. Рост числа обанкротившихся граждан во втором квартале по сравнению с первым замедлился — такая динамика обусловлена тем, что в период ограничений с 19 марта по 11 мая суды редко рассматривали банкротные дела компаний и меньше, чем обычно, в отношении граждан, объяснял руководитель проекта «Федресурс» Алексей Юхнин. Он прогнозирует, что во втором полугодии мы увидим рост показателя, когда суды рассмотрят все дела, что не смогли ранее в режиме ограничений. По его данным, уже в июне число личных банкротств выросло до 11,5 тыс., что в два раза превышает показатель аналогичного периода 2019 года." example_title: "Новости" - text: "Актуальность проблемы. Электронная информация играет все большую роль во всех сферах жизни современного общества. В последние годы объем научно-технической текстовой информации в электронном виде возрос настолько, что возникает угроза обесценивания этой информации в связи с трудностями поиска необходимых сведений среди множества доступных текстов. Развитие информационных ресурсов Интернет многократно усугубило проблему информационной перегрузки. В этой ситуации особенно актуальными становятся методы автоматизации реферирования текстовой информации, то есть методы получения сжатого представления текстовых документов–рефератов (аннотаций). Постановка проблемы автоматического реферирования текста и соответственно попытки ее решения с использованием различных подходов предпринимались многими исследователями. История применения вычислительной техники для реферирования насчитывает уже более 50 лет и связана с именами таких исследователей, как Г.П. Лун, В.Е. Берзон, И.П. Cевбо, Э.Ф. Скороходько, Д.Г. Лахути, Р.Г. Пиотровский и др. За эти годы выработаны многочисленные подходы к решению данной проблемы, которые достаточно четко подразделяются на два направления: автоматическое реферирование, основанное на экстрагировании из первичных документов с помощью определенных формальных признаков «наиболее информативных» фраз (фрагментов), совокупность которых образует некоторый экстракт; автоматическое реферирование, основанное на выделении из текстов с помощью специальных информационных языков наиболее существенной информации и порождении новых текстов (рефератов), содержательно обобщающих первичные документы." example_title: "Научная статья" --- # RuT5SumGazeta ## Model description This is the model for abstractive summarization for Russian based on [rut5-base](https://huggingface.co/cointegrated/rut5-base). ## Intended uses & limitations #### How to use Colab: [link](https://colab.research.google.com/drive/1re5E26ZIDUpAx1gOCZkbF3hcwjozmgG0) ```python from transformers import AutoTokenizer, T5ForConditionalGeneration model_name = "IlyaGusev/rut5_base_sum_gazeta" tokenizer = AutoTokenizer.from_pretrained(model_name) model = T5ForConditionalGeneration.from_pretrained(model_name) article_text = "..." input_ids = tokenizer( [article_text], max_length=600, add_special_tokens=True, padding="max_length", truncation=True, return_tensors="pt" )["input_ids"] output_ids = model.generate( input_ids=input_ids, no_repeat_ngram_size=4 )[0] summary = tokenizer.decode(output_ids, skip_special_tokens=True) print(summary) ``` ## Training data - Dataset: [Gazeta](https://huggingface.co/datasets/IlyaGusev/gazeta) ## Training procedure - Training script: [train.py](https://github.com/IlyaGusev/summarus/blob/master/external/hf_scripts/train.py) - Config: [t5_training_config.json](https://github.com/IlyaGusev/summarus/blob/master/external/hf_scripts/configs/t5_training_config.json) ## Eval results * Train dataset: **Gazeta v1 train** * Test dataset: **Gazeta v1 test** * Source max_length: **600** * Target max_length: **200** * no_repeat_ngram_size: **4** * num_beams: **5** | Model | R-1-f | R-2-f | R-L-f | chrF | METEOR | BLEU | Avg char length | |:--------------------------|:------|:------|:------|:-------|:-------|:-----|:-----| | [mbart_ru_sum_gazeta](https://huggingface.co/IlyaGusev/mbart_ru_sum_gazeta) | **32.4** | 14.3 | 28.0 | 39.7 | **26.4** | 12.1 | 371 | | [rut5_base_sum_gazeta](https://huggingface.co/IlyaGusev/rut5_base_sum_gazeta) | 32.2 | **14.4** | **28.1** | **39.8** | 25.7 | **12.3** | 330 | | [rugpt3medium_sum_gazeta](https://huggingface.co/IlyaGusev/rugpt3medium_sum_gazeta) | 26.2 | 7.7 | 21.7 | 33.8 | 18.2 | 4.3 | 244 | * Train dataset: **Gazeta v1 train** * Test dataset: **Gazeta v2 test** * Source max_length: **600** * Target max_length: **200** * no_repeat_ngram_size: **4** * num_beams: **5** | Model | R-1-f | R-2-f | R-L-f | chrF | METEOR | BLEU | Avg char length | |:--------------------------|:------|:------|:------|:-------|:-------|:-----|:-----| | [mbart_ru_sum_gazeta](https://huggingface.co/IlyaGusev/mbart_ru_sum_gazeta) | **28.7** | **11.1** | 24.4 | **37.3** | **22.7** | **9.4** | 373 | | [rut5_base_sum_gazeta](https://huggingface.co/IlyaGusev/rut5_base_sum_gazeta) | 28.6 | **11.1** | **24.5** | 37.2 | 22.0 | **9.4** | 331 | | [rugpt3medium_sum_gazeta](https://huggingface.co/IlyaGusev/rugpt3medium_sum_gazeta) | 24.1 | 6.5 | 19.8 | 32.1 | 16.3 | 3.6 | 242 | Predicting all summaries: ```python import json import torch from transformers import AutoTokenizer, T5ForConditionalGeneration from datasets import load_dataset def gen_batch(inputs, batch_size): batch_start = 0 while batch_start < len(inputs): yield inputs[batch_start: batch_start + batch_size] batch_start += batch_size def predict( model_name, input_records, output_file, max_source_tokens_count=600, batch_size=8 ): device = "cuda" if torch.cuda.is_available() else "cpu" tokenizer = AutoTokenizer.from_pretrained(model_name) model = T5ForConditionalGeneration.from_pretrained(model_name).to(device) predictions = [] for batch in gen_batch(input_records, batch_size): texts = [r["text"] for r in batch] input_ids = tokenizer( texts, add_special_tokens=True, max_length=max_source_tokens_count, padding="max_length", truncation=True, return_tensors="pt" )["input_ids"].to(device) output_ids = model.generate( input_ids=input_ids, no_repeat_ngram_size=4 ) summaries = tokenizer.batch_decode(output_ids, skip_special_tokens=True) for s in summaries: print(s) predictions.extend(summaries) with open(output_file, "w") as w: for p in predictions: w.write(p.strip().replace("\n", " ") + "\n") gazeta_test = load_dataset('IlyaGusev/gazeta', script_version="v1.0")["test"] predict("IlyaGusev/rut5_base_sum_gazeta", list(gazeta_test), "t5_predictions.txt") ``` Evaluation script: [evaluate.py](https://github.com/IlyaGusev/summarus/blob/master/evaluate.py) Flags: --language ru --tokenize-after --lower

Aleksandar/bert-srb-ner

token-classification

--- language: - ru license: apache-2.0 widget: - text: Эта блядь меня заебала</s> Эта <extra_id_0> меня <extra_id_1> ---

Aleksandar/distilbert-srb-base-cased-oscar

fill-mask

--- language: - ru - en tags: - xlm-roberta-large datasets: - IlyaGusev/headline_cause license: apache-2.0 widget: - text: "Песков опроверг свой перевод на удаленку</s>Дмитрий Песков перешел на удаленку" --- # XLM-RoBERTa HeadlineCause Full ## Model description This model was trained to predict the presence of causal relations between two headlines. This model is for the Full task with 7 possible labels: titles are almost the same, A causes B, B causes A, A refutes B, B refutes A, A linked with B in another way, A is not linked to B. English and Russian languages are supported. You can use hosted inference API to infer a label for a headline pair. To do this, you shoud seperate headlines with ```</s>``` token. For example: ``` Песков опроверг свой перевод на удаленку</s>Дмитрий Песков перешел на удаленку ``` ## Intended uses & limitations #### How to use ```python from tqdm.notebook import tqdm from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline def get_batch(data, batch_size): start_index = 0 while start_index < len(data): end_index = start_index + batch_size batch = data[start_index:end_index] yield batch start_index = end_index def pipe_predict(data, pipe, batch_size=64): raw_preds = [] for batch in tqdm(get_batch(data, batch_size)): raw_preds += pipe(batch) return raw_preds MODEL_NAME = TOKENIZER_NAME = "IlyaGusev/xlm_roberta_large_headline_cause_full" tokenizer = AutoTokenizer.from_pretrained(TOKENIZER_NAME, do_lower_case=False) model = AutoModelForSequenceClassification.from_pretrained(MODEL_NAME) model.eval() pipe = pipeline("text-classification", model=model, tokenizer=tokenizer, framework="pt", return_all_scores=True) texts = [ ( "Judge issues order to allow indoor worship in NC churches", "Some local churches resume indoor services after judge lifted NC governor’s restriction" ), ( "Gov. Kevin Stitt defends $2 million purchase of malaria drug touted by Trump", "Oklahoma spent $2 million on malaria drug touted by Trump" ), ( "Песков опроверг свой перевод на удаленку", "Дмитрий Песков перешел на удаленку" ) ] pipe_predict(texts, pipe) ``` #### Limitations and bias The models are intended to be used on news headlines. No other limitations are known. ## Training data * HuggingFace dataset: [IlyaGusev/headline_cause](https://huggingface.co/datasets/IlyaGusev/headline_cause) * GitHub: [IlyaGusev/HeadlineCause](https://github.com/IlyaGusev/HeadlineCause) ## Training procedure * Notebook: [HeadlineCause](https://colab.research.google.com/drive/1NAnD0OJ0TnYCJRsHpYUyYkjr_yi8ObcA) * Stand-alone script: [train.py](https://github.com/IlyaGusev/HeadlineCause/blob/main/headline_cause/train.py) ## Eval results Evaluation results can be found in the [arxiv paper](https://arxiv.org/pdf/2108.12626.pdf). ### BibTeX entry and citation info ```bibtex @misc{gusev2021headlinecause, title={HeadlineCause: A Dataset of News Headlines for Detecting Causalities}, author={Ilya Gusev and Alexey Tikhonov}, year={2021}, eprint={2108.12626}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

Aleksandar/distilbert-srb-ner-setimes-lr

--- language: - ru - en tags: - xlm-roberta-large datasets: - IlyaGusev/headline_cause license: apache-2.0 widget: - text: "Песков опроверг свой перевод на удаленку</s>Дмитрий Песков перешел на удаленку" --- # XLM-RoBERTa HeadlineCause Simple ## Model description This model was trained to predict the presence of causal relations between two headlines. This model is for the Simple task with 3 possible labels: A causes B, B causes A, no causal relation. English and Russian languages are supported. You can use hosted inference API to infer a label for a headline pair. To do this, you shoud seperate headlines with ```</s>``` token. For example: ``` Песков опроверг свой перевод на удаленку</s>Дмитрий Песков перешел на удаленку ``` ## Intended uses & limitations #### How to use ```python from tqdm.notebook import tqdm from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline def get_batch(data, batch_size): start_index = 0 while start_index < len(data): end_index = start_index + batch_size batch = data[start_index:end_index] yield batch start_index = end_index def pipe_predict(data, pipe, batch_size=64): raw_preds = [] for batch in tqdm(get_batch(data, batch_size)): raw_preds += pipe(batch) return raw_preds MODEL_NAME = TOKENIZER_NAME = "IlyaGusev/xlm_roberta_large_headline_cause_simple" tokenizer = AutoTokenizer.from_pretrained(TOKENIZER_NAME, do_lower_case=False) model = AutoModelForSequenceClassification.from_pretrained(MODEL_NAME) model.eval() pipe = pipeline("text-classification", model=model, tokenizer=tokenizer, framework="pt", return_all_scores=True) texts = [ ( "Judge issues order to allow indoor worship in NC churches", "Some local churches resume indoor services after judge lifted NC governor’s restriction" ), ( "Gov. Kevin Stitt defends $2 million purchase of malaria drug touted by Trump", "Oklahoma spent $2 million on malaria drug touted by Trump" ), ( "Песков опроверг свой перевод на удаленку", "Дмитрий Песков перешел на удаленку" ) ] pipe_predict(texts, pipe) ``` #### Limitations and bias The models are intended to be used on news headlines. No other limitations are known. ## Training data * HuggingFace dataset: [IlyaGusev/headline_cause](https://huggingface.co/datasets/IlyaGusev/headline_cause) * GitHub: [IlyaGusev/HeadlineCause](https://github.com/IlyaGusev/HeadlineCause) ## Training procedure * Notebook: [HeadlineCause](https://colab.research.google.com/drive/1NAnD0OJ0TnYCJRsHpYUyYkjr_yi8ObcA) * Stand-alone script: [train.py](https://github.com/IlyaGusev/HeadlineCause/blob/main/headline_cause/train.py) ## Eval results Evaluation results can be found in the [arxiv paper](https://arxiv.org/pdf/2108.12626.pdf). ### BibTeX entry and citation info ```bibtex @misc{gusev2021headlinecause, title={HeadlineCause: A Dataset of News Headlines for Detecting Causalities}, author={Ilya Gusev and Alexey Tikhonov}, year={2021}, eprint={2108.12626}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

Aleksandar/distilbert-srb-ner-setimes

token-classification

--- tags: - conversational --- # Harry Botter Model

Aleksandar1932/gpt2-rock-124439808

text-generation

## Usage: ``` from sentence_transformers import models from sentence_transformers import SentenceTransformer word_embedding_model = models.Transformer('Cro-CoV-cseBERT') pooling_model = models.Pooling(word_embedding_model.get_word_embedding_dimension(), pooling_mode_mean_tokens=True, pooling_mode_cls_token=False, pooling_mode_max_tokens=False) model = SentenceTransformer(modules=[word_embedding_model, pooling_model], device='') ## device = 'gpu' or 'cpu' texts_emb = model.encode(texts) ``` ## Datasets: https://github.com/InfoCoV/InfoCoV ## Paper: Please cite https://www.mdpi.com/2076-3417/11/21/10442

AlekseyKorshuk/bert

text-classification

--- language: en --- # Sparse BERT base model fine tuned to MNLI without classifier layer (uncased) Fine tuned sparse BERT base to MNLI (GLUE Benchmark) task from [bert-base-uncased-sparse-70-unstructured](https://huggingface.co/Intel/bert-base-uncased-sparse-70-unstructured). <br> This model doesn't have a classifier layer to enable easier loading of the model for training to other downstream tasks. In all the other layers this model is similar to [bert-base-uncased-mnli-sparse-70-unstructured](https://huggingface.co/Intel/bert-base-uncased-mnli-sparse-70-unstructured). <br><br> Note: This model requires `transformers==2.10.0` ## Evaluation Results Matched: 82.5% Mismatched: 83.3% This model can be further fine-tuned to other tasks and achieve the following evaluation results: | Task | QQP (Acc/F1) | QNLI (Acc) | SST-2 (Acc) | STS-B (Pears/Spear) | SQuADv1.1 (Acc/F1) | |------|--------------|------------|-------------|---------------------|--------------------| | | 90.2/86.7 | 90.3 | 91.5 | 88.9/88.6 | 80.5/88.2 |

AlekseyKorshuk/comedy-scripts

text-generation

--- language: en --- # Sparse BERT base model fine tuned to MNLI (uncased) Fine tuned sparse BERT base to MNLI (GLUE Benchmark) task from [bert-base-uncased-sparse-70-unstructured](https://huggingface.co/Intel/bert-base-uncased-sparse-70-unstructured). <br><br> Note: This model requires `transformers==2.10.0` ## Evaluation Results Matched: 82.5% Mismatched: 83.3% This model can be further fine-tuned to other tasks and achieve the following evaluation results: | Task | QQP (Acc/F1) | QNLI (Acc) | SST-2 (Acc) | STS-B (Pears/Spear) | SQuADv1.1 (Acc/F1) | |------|--------------|------------|-------------|---------------------|--------------------| | | 90.2/86.7 | 90.3 | 91.5 | 88.9/88.6 | 80.5/88.2 |

AlekseyKorshuk/horror-scripts

text-generation

--- language: en --- # Sparse BERT base model (uncased) Pretrained model pruned to 1:2 structured sparsity. The model is a pruned version of the [BERT base model](https://huggingface.co/bert-base-uncased). ## Intended Use The model can be used for fine-tuning to downstream tasks with sparsity already embeded to the model. To keep the sparsity a mask should be added to each sparse weight blocking the optimizer from updating the zeros. ## Evaluation Results We get the following results on the tasks development set, all results are mean of 5 different seeded models: | Task | MNLI-m (Acc) | MNLI-mm (Acc) | QQP (Acc/F1) | QNLI (Acc) | SST-2 (Acc) | STS-B (Pears/Spear) | SQuADv1.1 (Acc/F1) | |------|--------------|---------------|--------------|------------|-------------|---------------------|--------------------| | | 83.3 | 83.9 | 90.8/87.6 | 90.4 | 91.3 | 88.8/88.3 | 80.5/88.2 |

AlekseyKulnevich/Pegasus-HeaderGeneration

text2text-generation

--- language: en --- # Sparse BERT base model (uncased) Pretrained model pruned to 70% sparsity. The model is a pruned version of the [BERT base model](https://huggingface.co/bert-base-uncased). ## Intended Use The model can be used for fine-tuning to downstream tasks with sparsity already embeded to the model. To keep the sparsity a mask should be added to each sparse weight blocking the optimizer from updating the zeros.

Amba/wav2vec2-large-xls-r-300m-turkish-colab

--- tags: - conversational --- # Thor DialogGPT Model

Andrija/SRoBERTa

fill-mask

--- language: en widget: - text: "I am going to buy 100 shares of cake tomorrow" --- # roberta-ticker: model was fine-tuned from Roberta to detect financial tickers ## Introduction This is a model specifically designed to identify tickers in text. Model was trained on transformed dataset from following Kaggle dataset: https://www.kaggle.com/omermetinn/tweets-about-the-top-companies-from-2015-to-2020 ## How to use roberta-ticker with HuggingFace ##### Load roberta-ticker and its sub-word tokenizer : ```python from transformers import AutoTokenizer, AutoModelForTokenClassification tokenizer = AutoTokenizer.from_pretrained("Jean-Baptiste/roberta-ticker") model = AutoModelForTokenClassification.from_pretrained("Jean-Baptiste/roberta-ticker") ##### Process text sample from transformers import pipeline nlp = pipeline('ner', model=model, tokenizer=tokenizer, aggregation_strategy="simple") nlp("I am going to buy 100 shares of cake tomorrow") [{'entity_group': 'TICKER', 'score': 0.9612462520599365, 'word': ' cake', 'start': 32, 'end': 36}] nlp("I am going to eat a cake tomorrow") [] ``` ## Model performances ``` precision: 0.914157 recall: 0.788824 f1: 0.846878 ```

AnonymousSub/cline-s10-AR

text-classification

--- tags: - generated_from_trainer model-index: - name: BertjeWDialData results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # BertjeWDialData This model is a fine-tuned version of [GroNLP/bert-base-dutch-cased](https://huggingface.co/GroNLP/bert-base-dutch-cased) on the None dataset. It achieves the following results on the evaluation set: - Loss: 2.2608 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | No log | 1.0 | 297 | 2.2419 | ### Framework versions - Transformers 4.13.0.dev0 - Pytorch 1.10.0+cu111 - Datasets 1.15.1 - Tokenizers 0.10.3

AnonymousSub/cline-s10-SR

--- tags: - generated_from_trainer model-index: - name: BertjeWDialDataALL results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # BertjeWDialDataALL This model is a fine-tuned version of [GroNLP/bert-base-dutch-cased](https://huggingface.co/GroNLP/bert-base-dutch-cased) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.9469 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:-----:|:---------------:| | 2.1739 | 1.0 | 1542 | 2.0150 | | 2.0759 | 2.0 | 3084 | 1.9918 | | 2.0453 | 3.0 | 4626 | 2.0132 | | 1.9936 | 4.0 | 6168 | 1.9341 | | 1.9659 | 5.0 | 7710 | 1.9140 | | 1.9545 | 6.0 | 9252 | 1.9418 | | 1.9104 | 7.0 | 10794 | 1.9179 | | 1.8991 | 8.0 | 12336 | 1.9157 | ### Framework versions - Transformers 4.13.0.dev0 - Pytorch 1.10.0 - Datasets 1.16.1 - Tokenizers 0.10.3

AnonymousSub/roberta-base_squad2.0

question-answering

`LOREN` is an interpretable fact verification model trained on [FEVER](https://fever.ai), which aims to predict the veracity of a textual claim against a trustworthy knowledge source such as Wikipedia. `LOREN` also decomposes the verification and makes accurate and faithful phrase-level veracity predictions without any phrasal veracity supervision. This repo hosts the following pre-trained models for `LOREN`: - `fact_checking/`: the verification models based on BERT (large) and RoBERTa (large), respectively. - `mrc_seq2seq/`: the generative machine reading comprehension model based on BART (base). - `evidence_retrieval/`: the evidence sentence ranking models, which are copied directly from [KGAT](https://github.com/thunlp/KernelGAT). More technical details can be found at [this GitHub Repo](https://github.com/jiangjiechen/LOREN). Please check out our AAAI 2022 paper for more details: "[LOREN: Logic-Regularized Reasoning for Interpretable Fact Verification](https://arxiv.org/abs/2012.13577)".

AnonymousSub/rule_based_roberta_twostage_quadruplet_epochs_1_shard_1_squad2.0

question-answering

--- license: apache-2.0 language: es tags: - spanish datasets: - catalonia_independence metrics: - accuracy model-index: - name: roberta-base-bne-finetuned-mnli results: - task: name: Text Classification type: text-classification dataset: name: catalonia_independence type: catalonia_independence args: spanish metrics: - name: Accuracy type: accuracy value: 0.7880893300248138 - task: type: text-classification name: Text Classification dataset: name: catalonia_independence type: catalonia_independence config: catalan split: test metrics: - name: Accuracy type: accuracy value: 0.4592039800995025 verified: true - name: Precision Macro type: precision value: 0.6104489964825159 verified: true - name: Precision Micro type: precision value: 0.4592039800995025 verified: true - name: Precision Weighted type: precision value: 0.6167123723406555 verified: true - name: Recall Macro type: recall value: 0.4146479268294389 verified: true - name: Recall Micro type: recall value: 0.4592039800995025 verified: true - name: Recall Weighted type: recall value: 0.4592039800995025 verified: true - name: F1 Macro type: f1 value: 0.33416407167650636 verified: true - name: F1 Micro type: f1 value: 0.4592039800995025 verified: true - name: F1 Weighted type: f1 value: 0.34549318538357193 verified: true - name: loss type: loss value: 3.393402099609375 verified: true widget: - text: "Junqueras, sobre la decisi\xF3n judicial sobre Puigdemont: La justicia que\ \ falta en el Estado llega y llegar\xE1 de Europa" - text: "Desconvocada la manifestaci\xF3n del domingo en Barcelona en apoyo a Puigdemont" --- # roberta-base-bne-finetuned-catalonia-independence-detector This model is a fine-tuned version of [BSC-TeMU/roberta-base-bne](https://huggingface.co/BSC-TeMU/roberta-base-bne) on the catalonia_independence dataset. It achieves the following results on the evaluation set: - Loss: 0.9415 - Accuracy: 0.7881 <details> ## Model description The data was collected over 12 days during February and March of 2019 from tweets posted in Barcelona, and during September of 2018 from tweets posted in the town of Terrassa, Catalonia. Each corpus is annotated with three classes: AGAINST, FAVOR and NEUTRAL, which express the stance towards the target - independence of Catalonia. ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | No log | 1.0 | 378 | 0.5534 | 0.7558 | | 0.6089 | 2.0 | 756 | 0.5315 | 0.7643 | | 0.2678 | 3.0 | 1134 | 0.7336 | 0.7816 | | 0.0605 | 4.0 | 1512 | 0.8809 | 0.7866 | | 0.0605 | 5.0 | 1890 | 0.9415 | 0.7881 | </details> ### Model in action 🚀 Fast usage with **pipelines**: ```python from transformers import pipeline model_path = "JonatanGk/roberta-base-bne-finetuned-catalonia-independence-detector" independence_analysis = pipeline("text-classification", model=model_path, tokenizer=model_path) independence_analysis( "Junqueras, sobre la decisión judicial sobre Puigdemont: La justicia que falta en el Estado llega y llegará de Europa" ) # Output: [{'label': 'FAVOR', 'score': 0.9936726093292236}] independence_analysis( "El desafío independentista queda adormecido, y eso que el Gobierno ha sido muy claro en que su propuesta para Cataluña es una agenda de reencuentro, centrada en inversiones e infraestructuras") # Output: [{'label': 'AGAINST', 'score': 0.7508948445320129}] independence_analysis( "Desconvocada la manifestación del domingo en Barcelona en apoyo a Puigdemont" ) # Output: [{'label': 'NEUTRAL', 'score': 0.9966907501220703}] ``` [](https://colab.research.google.com/github/JonatanGk/Shared-Colab/blob/master/Catalonia_independence_Detector_(SPANISH).ipynb#scrollTo=uNMOXJz38W6U) ### Framework versions - Transformers 4.11.3 - Pytorch 1.9.0+cu111 - Datasets 1.12.1 - Tokenizers 0.10.3 ## Citation Thx to HF.co & [@lewtun](https://github.com/lewtun) for Dataset ;) > Special thx to [Manuel Romero/@mrm8488](https://huggingface.co/mrm8488) as my mentor & R.C. > Created by [Jonatan Luna](https://JonatanGk.github.io) | [LinkedIn](https://www.linkedin.com/in/JonatanGk/)

AnonymousSub/unsup-consert-base_copy_wikiqa

text-classification

--- language: en license: apache-2.0 tags: - sagemaker - roberta-base - text classification datasets: - emotion widget: - text: I am really upset that I have to call up to three times to the number on the back of my insurance card for my call to be answer model-index: - name: sagemaker-roberta-base-emotion results: - task: type: text-classification name: Multi Class Text Classification dataset: name: emotion type: emotion metrics: - type: accuracy value: 94.1 name: Validation Accuracy - type: f1 value: 94.13 name: Validation F1 - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion config: default split: test metrics: - type: accuracy value: 0.931 name: Accuracy verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNmM1ZmI0NjZhYjdlMWU4NWUwZmFjODFmMmM5MTlhMmEyMmQwOTk2NjQ5ZDNlYmFlMGEyMTY4Y2JiMTcwM2MwNiIsInZlcnNpb24iOjF9.haDbUk1y7nW1e_ext0s1xKefyOzep-XFa1HEkNQEcNV0cHCSRb-0YFakMf5Iee6q_EWFUS-QYxNkgEBlbw3fCQ - type: precision value: 0.8833042147663716 name: Precision Macro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNjZkOTQyMzkwYjE1ZWQ5YjJkMTEzNmIyZmFlMjkwY2YxNzA3OWE0ZDk5YjJlOWVhOTU5Nzc4ZTk5Mzg5NDcxOCIsInZlcnNpb24iOjF9._XhknNSsiailHiMr1SH9ki7SRswR_b-embALunoCjhBssh9WERkv0z1xpsbw7ORo0wx7WCslZRdJWaQoXOmgDQ - type: precision value: 0.931 name: Precision Micro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMGY0MTc0ZDBiYmZlYmFmMTcyYjk5MWM0MTRmYTlhY2U1ODY5NTQzNTQ5YjAzN2U0YjljNDAzZDQ5NDBkZDUwYyIsInZlcnNpb24iOjF9.313HYKetR4S4kjcMvEk9Yj2J-Ox8ZqvVk4FLrF6UmxlXYZ4F3put-89BEOxGl_ScugjjAWhKY1pHLPYpKz9PAA - type: precision value: 0.9337002742192515 name: Precision Weighted verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiZjQ1ZDIzNmE3MjljMTk2NTBmNzcyMTEyOTUwZTljYTA2MjIwY2E4ZThkNGVjYjQwNzU3MTcxMzBiYzJkNWIzOSIsInZlcnNpb24iOjF9.6yXKQ9WS9AWdt1jxixtA5O2S1bcPTKQqIOw291Ytam8OI-zdTI2jwltT6JdU4lHdhTi5797zeNldJMCxGPR2DQ - type: recall value: 0.9087144572668905 name: Recall Macro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNzJhNTFmNGJkYTAxNzRiOWQ4YzQyMGY5NGQxMjBiMmRjZTA5OTM2ZjM0NWY0ZDJiOTIyODQzZTZkMzEzZmY4YSIsInZlcnNpb24iOjF9.Fy1gkGvRiyANGU6nYgc5QbhccqAfb4PjxEk1EkJAIAZJjs-f0hffwUDlJt_6gRY3KKnoU2kKg1XxpWjybRY7BQ - type: recall value: 0.931 name: Recall Micro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiYTgwYWJmZDAzM2VkOGNjNjY3NjViOTFiMTYyZDc4ZDIzY2VhNTcwMDg3MjdiOTI4Nzc5ODI4N2ExYzY5ODAzMyIsInZlcnNpb24iOjF9.bEW-tZ-5JqkPDDfqkrdvzlzTGEJtYqRACZI1Jv7C8fWkJ8uJj0eQ8TDhcdGGDnFML-q1z3tnkO6PJuK9V2IxAg - type: recall value: 0.931 name: Recall Weighted verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiYTM2ZDk4NDQ2YWIwM2VjNzUxZjQ0YzU4MzViZGMzYzA3YjlhMTI1NjQwOTM3M2U4NGJhNTMxYzllMjRkMzU2NSIsInZlcnNpb24iOjF9.k9yprOWEoB0-k306GyDGF-g4uw3kABLc8iE_3E5ZYfVbo9VHPo61GuSsWJyYJ7_aq6zWbzgfOFEwUeVjcmnaDA - type: f1 value: 0.8949974527433656 name: F1 Macro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiODg0ZDllYWJkYWZkMWY2NjEzYWIxMWIwMWUyZDhmNWEzM2FmN2E0MWEwOTIyMTM2YTI1MDdmYmRmZWQ5ZmVmNCIsInZlcnNpb24iOjF9.DUD3dfb4vRu-Z9YxvDErJaPLuZIEDBNsdqzkf4ee6dkOCOnYtUhGAybnxtGN1xSYsynXYhU-ymCajWcrVKUCAA - type: f1 value: 0.931 name: F1 Micro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiOGU0MTYyOTNjOTBmNzAxNjVlZmQxYmRkMmE5MWY2NzhlNjg0ZGZkMmNmZmI3Zjk1NjJlYTdjMGRhMDMwYzAzNCIsInZlcnNpb24iOjF9.h0wCmhwRT4qRZJcc2zGP3T7dF0_wKdKzTtSVoVWFOUzQZ3RoeY2Hfjl3XA7yyw9KnoDWnLiW8DU_5kOBX-peCQ - type: f1 value: 0.9318434300647934 name: F1 Weighted verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiZmU4OGY4M2NkYWExNjI3Yjk0YmYzNWJjZGQ5ZGNmYzc4ZDk4YzRmZDRiNmRkN2VlNDZhOGIwZDc3MzcxYjVlYiIsInZlcnNpb24iOjF9.qhwi7AV-7NSm1yVd8v1Ea3nTRAFXfqLMwUJ5PUbPSa11jJ0tZNOQVDXHMAD8fVmoueLgZNRUpPVIB881Sq3EBg - type: loss value: 0.17379647493362427 name: loss verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNDdjODE2MjA5ODg2MmM2OWJmMjMzMzUzNGU1ZDc5NjRkNGU4N2VmNmM2NWE0YTEyYWMxNGUzN2M3YTkxNzUyMCIsInZlcnNpb24iOjF9.qcQWfHuRnfiluicR7gke3vm9u701hB4Bp0YaX2opaxL6d5DRCzuqAg-2kdmhhOL-8DW5JhY6gTrF14AEuEE9Cw --- ## roberta-base This model is a fine-tuned model that was trained using Amazon SageMaker and the new Hugging Face Deep Learning container. - Problem type: Multi Class Text Classification (emotion detection). It achieves the following results on the evaluation set: - Loss: 0.1613253802061081 - f1: 0.9413321705151999 ## Hyperparameters ```json { "epochs": 10, "train_batch_size": 16, "learning_rate": 3e-5, "weight_decay":0.01, "load_best_model_at_end": true, "model_name":"roberta-base", "do_eval": True, "load_best_model_at_end":True } ``` ## Validation Metrics | key | value | | --- | ----- | | eval_accuracy | 0.941 | | eval_f1 | 0.9413321705151999 | | eval_loss | 0.1613253802061081| | eval_recall | 0.941 | | eval_precision | 0.9419519436781406 |

AnonymousSub/unsup-consert-base_squad2.0

question-answering

--- tags: - asteroid - audio - ConvTasNet - audio-to-audio datasets: - Libri1Mix - enh_single license: cc-by-sa-4.0 --- ## Asteroid model `JorisCos/ConvTasNet_Libri1Mix_enhsignle_16k` Description: This model was trained by Joris Cosentino using the librimix recipe in [Asteroid](https://github.com/asteroid-team/asteroid). It was trained on the `enh_single` task of the Libri1Mix dataset. Training config: ```yml data: n_src: 1 sample_rate: 16000 segment: 3 task: enh_single train_dir: data/wav16k/min/train-360 valid_dir: data/wav16k/min/dev filterbank: kernel_size: 32 n_filters: 512 stride: 16 masknet: bn_chan: 128 hid_chan: 512 mask_act: relu n_blocks: 8 n_repeats: 3 n_src: 1 skip_chan: 128 optim: lr: 0.001 optimizer: adam weight_decay: 0.0 training: batch_size: 6 early_stop: true epochs: 200 half_lr: true num_workers: 4 ``` Results: On Libri1Mix min test set : ```yml si_sdr: 14.743051006476085 si_sdr_imp: 11.293269700616385 sdr: 15.300522933671061 sdr_imp: 11.797860134458015 sir: Infinity sir_imp: NaN sar: 15.300522933671061 sar_imp: 11.797860134458015 stoi: 0.9310514162434267 stoi_imp: 0.13513159270288563 ``` License notice: This work "ConvTasNet_Libri1Mix_enhsignle_16k" is a derivative of [LibriSpeech ASR corpus](http://www.openslr.org/12) by Vassil Panayotov, used under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/); of The WSJ0 Hipster Ambient Mixtures dataset by [Whisper.ai](http://wham.whisper.ai/), used under [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/) (Research only). "ConvTasNet_Libri1Mix_enhsignle_16k" is licensed under [Attribution-ShareAlike 3.0 Unported](https://creativecommons.org/licenses/by-sa/3.0/) by Joris Cosentino

Anorak/nirvana

text2text-generation

--- tags: - asteroid - audio - ConvTasNet - audio-to-audio datasets: - Libri3Mix - sep_clean license: cc-by-sa-4.0 --- ## Asteroid model `JorisCos/ConvTasNet_Libri3Mix_sepclean_8k` Description: This model was trained by Joris Cosentino using the librimix recipe in [Asteroid](https://github.com/asteroid-team/asteroid). It was trained on the `sep_clean` task of the Libri3Mix dataset. Training config: ```yml data: n_src: 3 sample_rate: 8000 segment: 3 task: sep_clean train_dir: data/wav8k/min/train-360 valid_dir: data/wav8k/min/dev filterbank: kernel_size: 16 n_filters: 512 stride: 8 masknet: bn_chan: 128 hid_chan: 512 mask_act: relu n_blocks: 8 n_repeats: 3 n_src: 3 skip_chan: 128 optim: lr: 0.001 optimizer: adam weight_decay: 0.0 training: batch_size: 24 early_stop: true epochs: 200 half_lr: true num_workers: 4 ``` Results : On Libri3Mix min test set : ```yaml si_sdr: 8.581797049575108 si_sdr_imp: 11.977037288467368 sdr' 9.305885208641385 sdr_imp: 12.3943409734845 sir: 16.42030534048559 sir_imp: 19.508759460400984 sar: 10.641943911079238 sar_imp: -56.4345187842095 stoi: 0.8365148408724333 stoi_imp: 0.24401766199806396 ``` License notice: This work "ConvTasNet_Libri3Mix_sepclean_8k" is a derivative of [LibriSpeech ASR corpus](http://www.openslr.org/12) by Vassil Panayotov, used under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). "ConvTasNet_Libri3Mix_sepclean_8k" is licensed under [Attribution-ShareAlike 3.0 Unported](https://creativecommons.org/licenses/by-sa/3.0/) by Cosentino Joris.

AnthonyNelson/DialoGPT-small-ricksanchez

conversational

--- tags: - asteroid - audio - ConvTasNet - audio-to-audio datasets: - Libri3Mix - sep_noisy license: cc-by-sa-4.0 --- ## Asteroid model `JorisCos/ConvTasNet_Libri3Mix_sepnoisy_16k` Description: This model was trained by Joris Cosentino using the librimix recipe in [Asteroid](https://github.com/asteroid-team/asteroid). It was trained on the `sep_noisy` task of the Libri3Mix dataset. Training config: ```yml data: n_src: 3 sample_rate: 16000 segment: 3 task: sep_noisy train_dir: data/wav16k/min/train-360 valid_dir: data/wav16k/min/dev filterbank: kernel_size: 32 n_filters: 512 stride: 16 masknet: bn_chan: 128 hid_chan: 512 mask_act: relu n_blocks: 8 n_repeats: 3 n_src: 3 skip_chan: 128 optim: lr: 0.001 optimizer: adam weight_decay: 0.0 training: batch_size: 8 early_stop: true epochs: 200 half_lr: true num_workers: 4 ``` Results: On Libri3Mix min test set : ```yml si_sdr: 5.926151147554517 si_sdr_imp: 10.282912158535625 sdr: 6.700975236867358 sdr_imp: 10.882972447337504 sir: 15.364110064569388 sir_imp: 18.574476587171688 sar: 7.918866830474568 sar_imp: -0.9638973409971135 stoi: 0.7713777027310713 stoi_imp: 0.2078696167973911 ``` License notice: This work "ConvTasNet_Libri3Mix_sepnoisy_16k" is a derivative of [LibriSpeech ASR corpus](http://www.openslr.org/12) by Vassil Panayotov, used under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/); of The WSJ0 Hipster Ambient Mixtures dataset by [Whisper.ai](http://wham.whisper.ai/), used under [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/). "ConvTasNet_Libri3Mix_sepnoisy_16k" is licensed under [Attribution-ShareAlike 3.0 Unported](https://creativecommons.org/licenses/by-sa/3.0/) by Joris Cosentino

Anthos23/distilbert-base-uncased-finetuned-sst2

text-classification

--- tags: - asteroid - audio - DCCRNet - audio-to-audio - speech-enhancement datasets: - Libri1Mix - enh_single license: cc-by-sa-4.0 --- ## Asteroid model `JorisCos/DCCRNet_Libri1Mix_enhsignle_16k` Description: This model was trained by Joris Cosentino using the librimix recipe in [Asteroid](https://github.com/asteroid-team/asteroid). It was trained on the `enh_single` task of the Libri1Mix dataset. Training config: ```yml data: n_src: 1 sample_rate: 16000 segment: 3 task: enh_single train_dir: data/wav16k/min/train-360 valid_dir: data/wav16k/min/dev filterbank: stft_kernel_size: 400 stft_n_filters: 512 stft_stride: 100 masknet: architecture: DCCRN-CL n_src: 1 optim: lr: 0.001 optimizer: adam weight_decay: 1.0e-05 training: batch_size: 12 early_stop: true epochs: 200 gradient_clipping: 5 half_lr: true num_workers: 4 ``` Results: On Libri1Mix min test set : ```yml si_sdr: 13.329767398333798 si_sdr_imp: 9.879986092474098 sdr: 13.87279932997016 sdr_imp: 10.370136530757103 sir: Infinity sir_imp: NaN sar: 13.87279932997016 sar_imp: 10.370136530757103 stoi: 0.9140907015623948 stoi_imp: 0.11817087802185405 ``` License notice: This work "DCCRNet_Libri1Mix_enhsignle_16k" is a derivative of [LibriSpeech ASR corpus](http://www.openslr.org/12) by Vassil Panayotov, used under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/); of The WSJ0 Hipster Ambient Mixtures dataset by [Whisper.ai](http://wham.whisper.ai/), used under [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/) (Research only). "DCCRNet_Libri1Mix_enhsignle_16k" is licensed under [Attribution-ShareAlike 3.0 Unported](https://creativecommons.org/licenses/by-sa/3.0/) by Joris Cosentino

Anthos23/my-awesome-model

text-classification

--- tags: - asteroid - audio - DCUNet - audio-to-audio datasets: - Libri1Mix - enh_single license: cc-by-sa-4.0 --- ## Asteroid model `JorisCos/DCUNet_Libri1Mix_enhsignle_16k` Description: This model was trained by Joris Cosentino using the librimix recipe in [Asteroid](https://github.com/asteroid-team/asteroid). It was trained on the `enh_single` task of the Libri1Mix dataset. Training config: ```yml data: n_src: 1 sample_rate: 16000 segment: 3 task: enh_single train_dir: data/wav16k/min/train-360 valid_dir: data/wav16k/min/dev filterbank: stft_n_filters: 1024 stft_kernel_size: 1024 stft_stride: 256 masknet: architecture: Large-DCUNet-20 fix_length_mode: pad n_src: 1 optim: lr: 0.001 optimizer: adam weight_decay: 1.0e-05 training: batch_size: 2 early_stop: true epochs: 200 gradient_clipping: 5 half_lr: true num_workers: 4 ``` Results: On Libri1Mix min test set : ```yml si_sdr: 13.154035391645971 si_sdr_imp: 9.704254085786271 sdr: 13.568058873121435 sdr_imp: 10.065396073908367 sar: 13.568058873121435 sar_imp: 10.065396073908367 stoi: 0.9199373340235417 stoi_imp: 0.12401751048300132 ``` License notice: This work "DCUNet_Libri1Mix_enhsignle_16k" is a derivative of [LibriSpeech ASR corpus](http://www.openslr.org/12) by Vassil Panayotov, used under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/); of The WSJ0 Hipster Ambient Mixtures dataset by [Whisper.ai](http://wham.whisper.ai/), used under [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/) (Research only). "DCUNet_Libri1Mix_enhsignle_16k" is licensed under [Attribution-ShareAlike 3.0 Unported](https://creativecommons.org/licenses/by-sa/3.0/) by Joris Cosentino

Anthos23/sentiment-roberta-large-english-finetuned-sentiment-analysis

--- tags: - asteroid - audio - DPRNNTasNet - audio-to-audio datasets: - Libri1Mix - enh_single license: cc-by-sa-4.0 --- ## Asteroid model `JorisCos/DPRNNTasNet_Libri1Mix_enhsignle_16k` Description: This model was trained by Joris Cosentino using the librimix recipe in [Asteroid](https://github.com/asteroid-team/asteroid). It was trained on the `enh_single` task of the Libri1Mix dataset. Training config: ```yml data: n_src: 1 sample_rate: 16000 segment: 1 task: enh_single train_dir: data/wav16k/min/train-360 valid_dir: data/wav16k/min/dev filterbank: kernel_size: 2 n_filters: 64 stride: 1 masknet: bidirectional: true bn_chan: 128 chunk_size: 250 dropout: 0 hid_size: 128 hop_size: 125 in_chan: 64 mask_act: sigmoid n_repeats: 6 n_src: 1 out_chan: 64 optim: lr: 0.001 optimizer: adam weight_decay: 1.0e-05 training: batch_size: 2 early_stop: true epochs: 200 gradient_clipping: 5 half_lr: true num_workers: 4 ``` Results: On Libri1Mix min test set : ```yml si_sdr: 14.7228101708889 si_sdr_imp: 11.2730288650292 sdr: 15.35661405197161 sdr_imp: 11.853951252758595 sir: Infinity sir_imp: NaN sar: 15.35661405197161 sar_imp: 11.853951252758595 stoi: 0.9300461826351578 stoi_imp: 0.13412635909461715 ``` License notice: This work "DPRNNTasNet_Libri1Mix_enhsignle_16k" is a derivative of [LibriSpeech ASR corpus](http://www.openslr.org/12) by Vassil Panayotov, used under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/); of The WSJ0 Hipster Ambient Mixtures dataset by [Whisper.ai](http://wham.whisper.ai/), used under [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/) (Research only). "DPRNNTasNet_Libri1Mix_enhsignle_16k" is licensed under [Attribution-ShareAlike 3.0 Unported](https://creativecommons.org/licenses/by-sa/3.0/) by Joris Cosentino

Apisate/Discord-Ai-Bot

text-generation

# BART_Finetuned_CNN_dailymail The following repo contains a [bart-base](https://huggingface.co/facebook/bart-base) model that was finetuned using the dataset [cnn_dailymail](https://huggingface.co/datasets/cnn_dailymail)

Apoorva/k2t-test

text2text-generation

--- license: mit tags: - translation - generated_from_trainer datasets: - kde4 model-index: - name: m2m100_418M-fr results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # m2m100_418M-fr This model is a fine-tuned version of [facebook/m2m100_418M](https://huggingface.co/facebook/m2m100_418M) on the kde4 dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 32 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 ### Framework versions - Transformers 4.12.5 - Pytorch 1.9.0+cpu - Datasets 1.16.1 - Tokenizers 0.10.3

ArBert/bert-base-uncased-finetuned-ner-agglo

--- tags: - conversational --- # Morty DialoGPT Model

ArBert/bert-base-uncased-finetuned-ner-gmm

--- tags: - conversational --- # Rick Sanchez DialoGPT Model

ArashEsk95/bert-base-uncased-finetuned-stsb

FidicBERT is a pre-trained language model to analyze legal text. It is built by further training the Roberta language model in the legal domain, using an extensive legal and contract corpus and thereby fine-tuning for classifying and clustering contractual documents.

Aravinth/test

--- language: - zh - ja - en tags: - translation widget: - text: "ja2zh: 吾輩は猫である。名前はまだ無い。" license: cc-by-nc-sa-4.0 --- This model is finetuned from [mt5-base](https://huggingface.co/google/mt5-base). The model vocabulary is trimmed to ~1/3 by selecting top 85000 tokens in the training data. The code to trim the vocabulary can be found [here](https://gist.github.com/K024/4a100a0f4f4b07208958e0f3244da6ad). Usage: ```python from transformers import ( T5Tokenizer, MT5ForConditionalGeneration, Text2TextGenerationPipeline, ) path = "K024/mt5-zh-ja-en-trimmed" pipe = Text2TextGenerationPipeline( model=MT5ForConditionalGeneration.from_pretrained(path), tokenizer=T5Tokenizer.from_pretrained(path), ) sentence = "ja2zh: 吾輩は猫である。名前はまだ無い。" res = pipe(sentence, max_length=100, num_beams=4) res[0]['generated_text'] ``` Training data: ``` wikimedia-en-ja wikimedia-en-zh wikimedia-ja-zh wikititles-ja-en wikititles-zh-en wikimatrix-ja-zh news-commentary-en-ja news-commentary-en-zh news-commentary-ja-zh ted2020-en-ja ted2020-en-zh ted2020-ja-zh ``` License: [![CC BY-NC-SA 4.0][cc-by-nc-sa-image]][cc-by-nc-sa] [cc-by-nc-sa]: http://creativecommons.org/licenses/by-nc-sa/4.0/ [cc-by-nc-sa-image]: https://licensebuttons.net/l/by-nc-sa/4.0/88x31.png

Archie/myProject

--- tags: - conversational --- #Rick DialoGPT Model

Arpita/opus-mt-en-ro-finetuned-syn-to-react

text2text-generation

--- language: sv tags: - audio - automatic-speech-recognition - voxrex license: cc0-1.0 --- # Wav2vec 2.0 large VoxRex (C) **Please note:** The model hosted in this repository is a pretrained wav2vec2 without a CTC head, as such it cannot do speech-to-text. If you are interested in speech-to-text, see our finetuned version of this model, which can be found at [KBLab/wav2vec2-large-voxrex-swedish](https://huggingface.co/KBLab/wav2vec2-large-voxrex-swedish). The weights found in this repository are from the pure acoustic model after unsupervised pretraining. This model is suitable for anyone interested in i) continued wav2vec2-pretraining with your own unsupervised data, ii) a feature extractor for finetuning your own downstream tasks (e.g. if you want to train your own CTC head, or an audio classifier). **Disclaimer:** This is a work in progress.<br> **Update 2022-01-08:** Updated to VoxRex-C version, use git to get the older (B) version.<br> **Update 2022-05-16:** Paper is is [here](https://arxiv.org/abs/2205.03026). This model has been pretrained for 400,000 updates on the P4-10k corpus which contains 10 000 hours of swedish local public service radio as well as 1500 hours of audio books and other speech from KBs collections. 

Arpita/opus-mt-en-ro-finetuned-synthon-to-reactant

--- language: sv-SE datasets: - common_voice - NST Swedish ASR Database metrics: - wer - cer tags: - audio - automatic-speech-recognition - speech - xlsr-fine-tuning-week license: apache-2.0 model-index: - name: XLSR Wav2Vec2 Swedish by KBLab results: - task: name: Speech Recognition type: automatic-speech-recognition dataset: name: Common Voice sv-SE type: common_voice args: sv-SE metrics: - name: Test WER type: wer value: 14.298610 - name: Test CER type: cer value: 4.925294 --- # Wav2Vec2-Large-XLSR-53-Swedish Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) in Swedish using the [NST Swedish Dictation](https://www.nb.no/sprakbanken/en/resource-catalogue/oai-nb-no-sbr-17/). When using this model, make sure that your speech input is sampled at 16kHz. ## Usage The model can be used directly (without a language model) as follows: ```python import torch import torchaudio from datasets import load_dataset from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor test_dataset = load_dataset("common_voice", "sv-SE", split="test[:2%]"). processor = Wav2Vec2Processor.from_pretrained("KBLab/wav2vec2-large-xlsr-53-swedish") model = Wav2Vec2ForCTC.from_pretrained("KBLab/wav2vec2-large-xlsr-53-swedish") resampler = torchaudio.transforms.Resample(48_000, 16_000) # Preprocessing the datasets. # We need to read the aduio files as arrays def speech_file_to_array_fn(batch): speech_array, sampling_rate = torchaudio.load(batch["path"]) batch["speech"] = resampler(speech_array).squeeze().numpy() return batch test_dataset = test_dataset.map(speech_file_to_array_fn) inputs = processor(test_dataset["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True) with torch.no_grad(): logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits predicted_ids = torch.argmax(logits, dim=-1) print("Prediction:", processor.batch_decode(predicted_ids)) print("Reference:", test_dataset["sentence"][:2]) ``` ## Evaluation The model can be evaluated as follows on the Swedish test data of Common Voice. ```python import torch import torchaudio from datasets import load_dataset, load_metric from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor import re test_dataset = load_dataset("common_voice", "sv-SE", split="test") wer = load_metric("wer") processor = Wav2Vec2Processor.from_pretrained("KBLab/wav2vec2-large-xlsr-53-swedish") model = Wav2Vec2ForCTC.from_pretrained("KBLab/wav2vec2-large-xlsr-53-swedish") model.to("cuda") chars_to_ignore_regex = '[,?.!\\-;:"“]' resampler = torchaudio.transforms.Resample(48_000, 16_000) # Preprocessing the datasets. # We need to read the aduio files as arrays def speech_file_to_array_fn(batch): batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower() speech_array, sampling_rate = torchaudio.load(batch["path"]) batch["speech"] = resampler(speech_array).squeeze().numpy() return batch test_dataset = test_dataset.map(speech_file_to_array_fn) # Preprocessing the datasets. # We need to read the aduio files as arrays def evaluate(batch): inputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True) with torch.no_grad(): logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits pred_ids = torch.argmax(logits, dim=-1) batch["pred_strings"] = processor.batch_decode(pred_ids) return batch result = test_dataset.map(evaluate, batched=True, batch_size=8) print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"]))) print("CER: {:2f}".format(100 * wer.compute(predictions=[" ".join(list(entry)) for entry in result["pred_strings"]], references=[" ".join(list(entry)) for entry in result["sentence"]]))) ``` **WER**: 14.298610% **CER**: 4.925294% ## Training First the XLSR model was further pre-trained for 50 epochs with a corpus consisting of 1000 hours spoken Swedish from various radio stations. Secondly [NST Swedish Dictation](https://www.nb.no/sprakbanken/en/resource-catalogue/oai-nb-no-sbr-17/) was used for fine tuning as well as [Common Voice](https://commonvoice.mozilla.org/en/datasets). Lastly only Common Voice dataset was used for final finetuning. The [Fairseq](https://github.com/fairseq) scripts were used.

AryanLala/autonlp-Scientific_Title_Generator-34558227

text2text-generation

--- license: apache-2.0 language: [ko, en] tags: - t5 eos_token: "</s>" widget: - text: 아버지가 방에 들어가신다.</s> --- # ke-t5 base Pretrained T5 Model on Korean and English. See [Github](https://github.com/AIRC-KETI/ke-t5) and [Paper](https://aclanthology.org/2021.findings-emnlp.33/) [Korean paper](https://koreascience.kr/article/CFKO202130060717834.pdf) for more details. ## How to use ```python from transformers import AutoModel, AutoTokenizer model = AutoModel.from_pretrained("KETI-AIR/ke-t5-large-newslike") tokenizer = AutoTokenizer.from_pretrained("KETI-AIR/ke-t5-large-newslike") ``` ## BibTeX entry and citation info ```bibtex @inproceedings{kim-etal-2021-model-cross, title = "A Model of Cross-Lingual Knowledge-Grounded Response Generation for Open-Domain Dialogue Systems", author = "Kim, San and Jang, Jin Yea and Jung, Minyoung and Shin, Saim", booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2021", month = nov, year = "2021", address = "Punta Cana, Dominican Republic", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2021.findings-emnlp.33", doi = "10.18653/v1/2021.findings-emnlp.33", pages = "352--365", abstract = "Research on open-domain dialogue systems that allow free topics is challenging in the field of natural language processing (NLP). The performance of the dialogue system has been improved recently by the method utilizing dialogue-related knowledge; however, non-English dialogue systems suffer from reproducing the performance of English dialogue systems because securing knowledge in the same language with the dialogue system is relatively difficult. Through experiments with a Korean dialogue system, this paper proves that the performance of a non-English dialogue system can be improved by utilizing English knowledge, highlighting the system uses cross-lingual knowledge. For the experiments, we 1) constructed a Korean version of the Wizard of Wikipedia dataset, 2) built Korean-English T5 (KE-T5), a language model pre-trained with Korean and English corpus, and 3) developed a knowledge-grounded Korean dialogue model based on KE-T5. We observed the performance improvement in the open-domain Korean dialogue model even only English knowledge was given. The experimental results showed that the knowledge inherent in cross-lingual language models can be helpful for generating responses in open dialogue systems.", } ```

Ashagi/Ashvx

--- license: apache-2.0 language: ko tags: - t5 eos_token: "</s>" widget: - text: 아버지가 방에 들어가신다.</s> --- # ke-t5 base Pretrained T5 Model on Korean and English. See [Github](https://github.com/AIRC-KETI/ke-t5) and [Paper](https://aclanthology.org/2021.findings-emnlp.33/) [Korean paper](https://koreascience.kr/article/CFKO202130060717834.pdf) for more details. ## How to use ```python from transformers import AutoModel, AutoTokenizer model = AutoModel.from_pretrained("KETI-AIR/ke-t5-small-ko") tokenizer = AutoTokenizer.from_pretrained("KETI-AIR/ke-t5-small-ko") ``` ## BibTeX entry and citation info ```bibtex @inproceedings{kim-etal-2021-model-cross, title = "A Model of Cross-Lingual Knowledge-Grounded Response Generation for Open-Domain Dialogue Systems", author = "Kim, San and Jang, Jin Yea and Jung, Minyoung and Shin, Saim", booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2021", month = nov, year = "2021", address = "Punta Cana, Dominican Republic", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2021.findings-emnlp.33", doi = "10.18653/v1/2021.findings-emnlp.33", pages = "352--365", abstract = "Research on open-domain dialogue systems that allow free topics is challenging in the field of natural language processing (NLP). The performance of the dialogue system has been improved recently by the method utilizing dialogue-related knowledge; however, non-English dialogue systems suffer from reproducing the performance of English dialogue systems because securing knowledge in the same language with the dialogue system is relatively difficult. Through experiments with a Korean dialogue system, this paper proves that the performance of a non-English dialogue system can be improved by utilizing English knowledge, highlighting the system uses cross-lingual knowledge. For the experiments, we 1) constructed a Korean version of the Wizard of Wikipedia dataset, 2) built Korean-English T5 (KE-T5), a language model pre-trained with Korean and English corpus, and 3) developed a knowledge-grounded Korean dialogue model based on KE-T5. We observed the performance improvement in the open-domain Korean dialogue model even only English knowledge was given. The experimental results showed that the knowledge inherent in cross-lingual language models can be helpful for generating responses in open dialogue systems.", } ```

Ashok/my-new-tokenizer

--- tags: - conversational --- # Clever bot DialoGPT Model

Ateeb/EmotionDetector

text-classification

--- tags: - conversational --- #Peralta DialoGPT Model

Ayham/robertagpt2_xsum

text2text-generation

--- tags: - generated_from_trainer model-index: - name: distilbert-undersampled-noweights results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-undersampled-noweights This model was trained from scratch on the None dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 33 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 100 - num_epochs: 5 - mixed_precision_training: Native AMP ### Framework versions - Transformers 4.16.2 - Pytorch 1.10.0+cu111 - Datasets 1.18.3 - Tokenizers 0.11.0

Ayham/robertagpt2_xsum2

text2text-generation

--- license: apache-2.0 tags: - generated_from_trainer metrics: - accuracy - f1 - recall - precision model-index: - name: distilbert-undersampled results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-undersampled This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.0826 - Accuracy: 0.9811 - F1: 0.9810 - Recall: 0.9811 - Precision: 0.9812 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 33 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 100 - num_epochs: 5 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Recall | Precision | |:-------------:|:-----:|:-----:|:---------------:|:--------:|:------:|:------:|:---------:| | 0.0959 | 0.2 | 2000 | 0.0999 | 0.9651 | 0.9628 | 0.9651 | 0.9655 | | 0.0618 | 0.41 | 4000 | 0.0886 | 0.9717 | 0.9717 | 0.9717 | 0.9731 | | 0.159 | 0.61 | 6000 | 0.0884 | 0.9719 | 0.9720 | 0.9719 | 0.9728 | | 0.0513 | 0.81 | 8000 | 0.0785 | 0.9782 | 0.9782 | 0.9782 | 0.9788 | | 0.0219 | 1.01 | 10000 | 0.0680 | 0.9779 | 0.9779 | 0.9779 | 0.9783 | | 0.036 | 1.22 | 12000 | 0.0745 | 0.9787 | 0.9787 | 0.9787 | 0.9792 | | 0.0892 | 1.42 | 14000 | 0.0675 | 0.9786 | 0.9786 | 0.9786 | 0.9789 | | 0.0214 | 1.62 | 16000 | 0.0760 | 0.9799 | 0.9798 | 0.9799 | 0.9801 | | 0.0882 | 1.83 | 18000 | 0.0800 | 0.9800 | 0.9800 | 0.9800 | 0.9802 | | 0.0234 | 2.03 | 20000 | 0.0720 | 0.9813 | 0.9813 | 0.9813 | 0.9815 | | 0.0132 | 2.23 | 22000 | 0.0738 | 0.9803 | 0.9803 | 0.9803 | 0.9805 | | 0.0136 | 2.43 | 24000 | 0.0847 | 0.9804 | 0.9804 | 0.9804 | 0.9806 | | 0.0119 | 2.64 | 26000 | 0.0826 | 0.9811 | 0.9810 | 0.9811 | 0.9812 | ### Framework versions - Transformers 4.16.2 - Pytorch 1.10.0+cu111 - Datasets 1.18.3 - Tokenizers 0.11.0

Ayjayo/DialoGPT-medium-AyjayoAI

conversational

--- license: apache-2.0 tags: - generated_from_trainer datasets: - pub_med_summarization_dataset metrics: - rouge model-index: - name: bart-base-finetuned-pubmed results: - task: name: Sequence-to-sequence Language Modeling type: text2text-generation dataset: name: pub_med_summarization_dataset type: pub_med_summarization_dataset args: document metrics: - name: Rouge1 type: rouge value: 9.3963 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bart-base-finetuned-pubmed This model is a fine-tuned version of [facebook/bart-base](https://huggingface.co/facebook/bart-base) on the pub_med_summarization_dataset dataset. It achieves the following results on the evaluation set: - Loss: 2.0277 - Rouge1: 9.3963 - Rouge2: 4.0473 - Rougel: 8.4526 - Rougelsum: 8.9659 - Gen Len: 20.0 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len | |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:------:|:---------:|:-------:| | 2.3706 | 1.0 | 4000 | 2.1245 | 9.1644 | 3.8264 | 8.2223 | 8.718 | 20.0 | | 2.2246 | 2.0 | 8000 | 2.0811 | 9.023 | 3.7716 | 8.1453 | 8.5998 | 20.0 | | 2.1034 | 3.0 | 12000 | 2.0469 | 9.4412 | 4.0783 | 8.4949 | 8.9977 | 20.0 | | 2.0137 | 4.0 | 16000 | 2.0390 | 9.2261 | 3.9307 | 8.3154 | 8.7937 | 20.0 | | 1.9288 | 5.0 | 20000 | 2.0277 | 9.3963 | 4.0473 | 8.4526 | 8.9659 | 20.0 | ### Framework versions - Transformers 4.16.2 - Pytorch 1.10.0+cu111 - Datasets 1.18.3 - Tokenizers 0.11.6

Aymene/opus-mt-en-ro-finetuned-en-to-ro

--- license: mit tags: - generated_from_trainer datasets: - pub_med_summarization_dataset metrics: - rouge model-index: - name: bart-large-cnn-finetuned-pubmed results: - task: name: Sequence-to-sequence Language Modeling type: text2text-generation dataset: name: pub_med_summarization_dataset type: pub_med_summarization_dataset args: document metrics: - name: Rouge1 type: rouge value: 40.4866 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bart-large-cnn-finetuned-pubmed This model is a fine-tuned version of [facebook/bart-large-cnn](https://huggingface.co/facebook/bart-large-cnn) on the pub_med_summarization_dataset dataset. It achieves the following results on the evaluation set: - Loss: 1.8416 - Rouge1: 40.4866 - Rouge2: 16.7472 - Rougel: 24.9831 - Rougelsum: 36.4002 - Gen Len: 142.0 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len | |:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:--------:| | 1.932 | 1.0 | 4000 | 1.8110 | 38.1151 | 15.2255 | 23.4286 | 34.2521 | 141.8905 | | 1.7001 | 2.0 | 8000 | 1.7790 | 39.8217 | 16.3042 | 24.649 | 35.831 | 142.0 | | 1.5 | 3.0 | 12000 | 1.7971 | 40.6108 | 17.0446 | 25.1977 | 36.5556 | 141.9865 | | 1.3316 | 4.0 | 16000 | 1.8106 | 40.0466 | 16.4851 | 24.7094 | 36.0998 | 141.9335 | | 1.1996 | 5.0 | 20000 | 1.8416 | 40.4866 | 16.7472 | 24.9831 | 36.4002 | 142.0 | ### Framework versions - Transformers 4.16.2 - Pytorch 1.10.0+cu111 - Datasets 1.18.3 - Tokenizers 0.11.6

Ayoola/cdial-yoruba-test

automatic-speech-recognition

--- license: apache-2.0 tags: - generated_from_trainer datasets: - pub_med_summarization_dataset metrics: - rouge model-index: - name: bart-large-finetuned-pubmed results: - task: name: Sequence-to-sequence Language Modeling type: text2text-generation dataset: name: pub_med_summarization_dataset type: pub_med_summarization_dataset args: document metrics: - name: Rouge1 type: rouge value: 10.946 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bart-large-finetuned-pubmed This model is a fine-tuned version of [facebook/bart-large](https://huggingface.co/facebook/bart-large) on the pub_med_summarization_dataset dataset. It achieves the following results on the evaluation set: - Loss: 1.8135 - Rouge1: 10.946 - Rouge2: 5.0933 - Rougel: 9.5608 - Rougelsum: 10.4259 - Gen Len: 19.0495 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len | |:-------------:|:-----:|:-----:|:---------------:|:-------:|:------:|:------:|:---------:|:-------:| | 2.0861 | 1.0 | 4000 | 1.8909 | 8.7344 | 3.6919 | 7.8804 | 8.3305 | 20.0 | | 1.8996 | 2.0 | 8000 | 1.8261 | 10.2124 | 4.6212 | 8.9842 | 9.7417 | 17.632 | | 1.7459 | 3.0 | 12000 | 1.8160 | 9.4933 | 4.4117 | 8.3977 | 9.0758 | 16.4775 | | 1.6258 | 4.0 | 16000 | 1.8136 | 10.8248 | 5.0335 | 9.4286 | 10.3123 | 18.724 | | 1.5214 | 5.0 | 20000 | 1.8135 | 10.946 | 5.0933 | 9.5608 | 10.4259 | 19.0495 | ### Framework versions - Transformers 4.16.2 - Pytorch 1.10.0+cu111 - Datasets 1.18.3 - Tokenizers 0.11.6

Ayran/DialoGPT-small-gandalf

conversational

--- license: apache-2.0 tags: - generated_from_trainer model-index: - name: distilbert-base-multilingual-cased-finetuned-viquad results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-multilingual-cased-finetuned-viquad This model is a fine-tuned version of [distilbert-base-multilingual-cased](https://huggingface.co/distilbert-base-multilingual-cased) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.4241 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | No log | 1.0 | 65 | 4.0975 | | No log | 2.0 | 130 | 3.9315 | | No log | 3.0 | 195 | 3.6742 | | No log | 4.0 | 260 | 3.4878 | | No log | 5.0 | 325 | 3.4241 | ### Framework versions - Transformers 4.15.0 - Pytorch 1.10.0+cu111 - Datasets 1.17.0 - Tokenizers 0.10.3

Ayran/DialoGPT-small-harry-potter-1-through-3

conversational

--- license: mit tags: - generated_from_trainer model-index: - name: xlm-roberta-base-finetuned-squad results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # xlm-roberta-base-finetuned-squad This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.5539 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | 0.7665 | 1.0 | 2295 | 0.5231 | | 0.5236 | 2.0 | 4590 | 0.5539 | ### Framework versions - Transformers 4.15.0 - Pytorch 1.10.0+cu111 - Datasets 1.17.0 - Tokenizers 0.10.3

Ayta/Haha

--- license: mit tags: - generated_from_trainer model-index: - name: xlm-roberta-base-finetuned-viquad results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # xlm-roberta-base-finetuned-viquad This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset. It achieves the following results on the evaluation set: - Loss: 2.3761 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | No log | 1.0 | 259 | 2.9945 | | 3.3665 | 2.0 | 518 | 2.3761 | ### Framework versions - Transformers 4.15.0 - Pytorch 1.10.0+cu111 - Datasets 1.17.0 - Tokenizers 0.10.3

Azura/data

--- license: apache-2.0 tags: - generated_from_trainer datasets: - emotion metrics: - accuracy - f1 model-index: - name: distilbert-base-uncased-finetuned-emotion results: - task: name: Text Classification type: text-classification dataset: name: emotion type: emotion args: default metrics: - name: Accuracy type: accuracy value: 0.9225 - name: F1 type: f1 value: 0.9227765339978083 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset. It achieves the following results on the evaluation set: - Loss: 0.2224 - Accuracy: 0.9225 - F1: 0.9228 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:| | 0.84 | 1.0 | 250 | 0.3133 | 0.909 | 0.9070 | | 0.2459 | 2.0 | 500 | 0.2224 | 0.9225 | 0.9228 | ### Framework versions - Transformers 4.11.3 - Pytorch 1.9.1 - Datasets 1.16.1 - Tokenizers 0.10.3

Azuris/DialoGPT-medium-senorita

conversational

--- language: - rus tags: - mbart inference: parameters: no_repeat_ngram_size: 4, num_beams : 5 datasets: - IlyaGusev/gazeta - samsum - samsum_(translated_into_Russian) widget: - text: | Джефф: Могу ли я обучить модель 🤗 Transformers на Amazon SageMaker? Филипп: Конечно, вы можете использовать новый контейнер для глубокого обучения HuggingFace. Джефф: Хорошо. Джефф: и как я могу начать? Джефф: где я могу найти документацию? Филипп: ок, ок, здесь можно найти все: https://huggingface.co/blog/the-partnership-amazon-sagemaker-and-hugging-face model-index: - name: "mbart_ruDialogSum" results: - task: name: Abstractive Dialogue Summarization type: abstractive-text-summarization dataset: name: "SAMSum Corpus (translated to Russian)" type: samsum metrics: - name: Validation ROGUE-1 type: rogue-1 value: 34.5 - name: Validation ROGUE-L type: rogue-l value: 33 - name: Test ROGUE-1 type: rogue-1 value: 31 - name: Test ROGUE-L type: rogue-l value: 28 --- ### 📝 Description MBart for Russian summarization fine-tuned for **dialogues** summarization. This model was firstly fine-tuned by [Ilya Gusev](https://hf.co/IlyaGusev) on [Gazeta dataset](https://huggingface.co/datasets/IlyaGusev/gazeta). We have **fine tuned** that model on [SamSum dataset]() **translated to Russian** using GoogleTranslateAPI 🤗 Moreover! We have implemented a **! telegram bot [@summarization_bot](https://t.me/summarization_bot) !** with the inference of this model. Add it to the chat and get summaries instead of dozens spam messages!  🤗 ### ❓ How to use with code ```python from transformers import MBartTokenizer, MBartForConditionalGeneration # Download model and tokenizer model_name = "Kirili4ik/mbart_ruDialogSum" tokenizer = AutoTokenizer.from_pretrained(model_name) model = MBartForConditionalGeneration.from_pretrained(model_name) model.eval() article_text = "..." input_ids = tokenizer( [article_text], max_length=600, padding="max_length", truncation=True, return_tensors="pt", )["input_ids"] output_ids = model.generate( input_ids=input_ids, top_k=0, num_beams=3, no_repeat_ngram_size=3 )[0] summary = tokenizer.decode(output_ids, skip_special_tokens=True) print(summary) ```

BAHIJA/distilbert-base-uncased-finetuned-cola

text-classification

--- language: - ru - ru-RU tags: - conversational --- ### 📝 Description DialoGPT trained on Russian language and fine tuned on my telegram chat. This model was created by [sberbank-ai](https://hf.co/sberbank-ai) and trained on Russian forums (see [Grossmend's model](https://hf.co/Grossmend/rudialogpt3_medium_based_on_gpt2)). You can find info about how it has been trained on [habr](https://habr.com/ru/company/icl_services/blog/548244/) (in Russian). I have created a **simple pipeline** and **fine tuned** that model on my own **exported telegram chat** (~30mb json). It is in fact very easy to get the data from telegram and fine tune a model. Therefore, I made a **colab tutorial** for it: https://colab.research.google.com/drive/1fnAVURjyZRK9VQg1Co_-SKUQnRES8l9R?usp=sharing ⚠️ Due to specifics of the data Hosted inference API may not work properly ⚠️ 🤗To try it use my [Spaces demo](https://huggingface.co/spaces/Kirili4ik/chat-with-Kirill)🤗 ### ❓ How to use with code ```python # Download model and tokenizer checkpoint = "Kirili4ik/ruDialoGpt3-medium-finetuned-telegram" tokenizer = AutoTokenizer.from_pretrained(checkpoint) model = AutoModelForCausalLM.from_pretrained(checkpoint) model.eval() # util function to get expected len after tokenizing def get_length_param(text: str, tokenizer) -> str: tokens_count = len(tokenizer.encode(text)) if tokens_count <= 15: len_param = '1' elif tokens_count <= 50: len_param = '2' elif tokens_count <= 256: len_param = '3' else: len_param = '-' return len_param # util function to get next person number (1/0) for Machine or Human in the dialogue def get_user_param(text: dict, machine_name_in_chat: str) -> str: if text['from'] == machine_name_in_chat: return '1' # machine else: return '0' # human chat_history_ids = torch.zeros((1, 0), dtype=torch.int) while True: next_who = input("Who's phrase?\t") #input("H / G?") # Human or GPT # In case Human if next_who == "H" or next_who == "Human": input_user = input("===> Human: ") # encode the new user input, add parameters and return a tensor in Pytorch new_user_input_ids = tokenizer.encode(f"|0|{get_length_param(input_user, tokenizer)}|" \ + input_user + tokenizer.eos_token, return_tensors="pt") # append the new user input tokens to the chat history chat_history_ids = torch.cat([chat_history_ids, new_user_input_ids], dim=-1) if next_who == "G" or next_who == "GPT": next_len = input("Phrase len? 1/2/3/-\t") #input("Exp. len?(-/1/2/3): ") # encode the new user input, add parameters and return a tensor in Pytorch new_user_input_ids = tokenizer.encode(f"|1|{next_len}|", return_tensors="pt") # append the new user input tokens to the chat history chat_history_ids = torch.cat([chat_history_ids, new_user_input_ids], dim=-1) # print(tokenizer.decode(chat_history_ids[-1])) # uncomment to see full gpt input # save previous len input_len = chat_history_ids.shape[-1] # generated a response; PS you can read about the parameters at hf.co/blog/how-to-generate chat_history_ids = model.generate( chat_history_ids, num_return_sequences=1, # use for more variants, but have to print [i] max_length=512, no_repeat_ngram_size=3, do_sample=True, top_k=50, top_p=0.9, temperature = 0.6, # 0 for greedy mask_token_id=tokenizer.mask_token_id, eos_token_id=tokenizer.eos_token_id, unk_token_id=tokenizer.unk_token_id, pad_token_id=tokenizer.pad_token_id, device='cpu' ) # pretty print last ouput tokens from bot print(f"===> GPT-3: {tokenizer.decode(chat_history_ids[:, input_len:][0], skip_special_tokens=True)}") ```

BSen/wav2vec2-large-xls-r-300m-turkish-colab

automatic-speech-recognition

--- language: en license: mit --- # GPT-Neo 2.7B - Shinen ## Model Description GPT-Neo 2.7B-Shinen is a finetune created using EleutherAI's GPT-Neo 2.7B model. Compared to GPT-Neo-2.7-Horni, this model is much heavier on the sexual content. **Warning: THIS model is NOT suitable for use by minors. The model will output X-rated content.** ## Training data The training data contains user-generated stories from sexstories.com. All stories are tagged using the following way: ``` [Theme: <theme1>, <theme2> ,<theme3>] <Story goes here> ``` ### How to use You can use this model directly with a pipeline for text generation. This example generates a different sequence each time it's run: ```py >>> from transformers import pipeline >>> generator = pipeline('text-generation', model='KoboldAI/GPT-Neo-2.7B-Shinen') >>> generator("She was staring at me", do_sample=True, min_length=50) [{'generated_text': 'She was staring at me with a look that said it all. She wanted me so badly tonight that I wanted'}] ``` ### Limitations and Biases GPT-Neo was trained as an autoregressive language model. This means that its core functionality is taking a string of text and predicting the next token. While language models are widely used for tasks other than this, there are a lot of unknowns with this work. GPT-Neo-Shinen was trained on a dataset known to contain profanity, lewd, and otherwise abrasive language. GPT-Neo-Shinen *WILL* produce socially unacceptable text without warning. GPT-Neo-Shinen will respond to particular prompts and offensive content may occur without warning. We recommend having a human curate or filter the outputs before releasing them, both to censor undesirable content and to improve the quality of the results. ### BibTeX entry and citation info The model is made using the following software: ```bibtex @software{gpt-neo, author = {Black, Sid and Leo, Gao and Wang, Phil and Leahy, Connor and Biderman, Stella}, title = {{GPT-Neo: Large Scale Autoregressive Language Modeling with Mesh-Tensorflow}}, month = mar, year = 2021, note = {{If you use this software, please cite it using these metadata.}}, publisher = {Zenodo}, version = {1.0}, doi = {10.5281/zenodo.5297715}, url = {https://doi.org/10.5281/zenodo.5297715} } ```

Bagus/wav2vec2-xlsr-greek-speech-emotion-recognition

audio-classification

--- language: - "ja" tags: - "japanese" - "masked-lm" - "wikipedia" license: "cc-by-sa-4.0" pipeline_tag: "fill-mask" mask_token: "[MASK]" widget: - text: "酸素ボンベを充[MASK]する。" --- # bert-base-japanese-char-extended ## Model Description This is a BERT model pre-trained on Japanese Wikipedia texts, derived from [bert-base-japanese-char-v2](https://huggingface.co/cl-tohoku/bert-base-japanese-char-v2). Character-embeddings are enhanced to include all 常用漢字/人名用漢字 characters using BertTokenizerFast. You can fine-tune `bert-base-japanese-char-extended` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/bert-base-japanese-upos), [dependency-parsing](https://huggingface.co/KoichiYasuoka/bert-base-japanese-wikipedia-ud-head), and so on. ## How to Use ```py from transformers import AutoTokenizer,AutoModelForMaskedLM tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/bert-base-japanese-char-extended") model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/bert-base-japanese-char-extended") ```

Bakkes/BakkesModWiki

--- language: - "ja" tags: - "japanese" - "token-classification" - "pos" - "wikipedia" - "dependency-parsing" datasets: - "universal_dependencies" license: "cc-by-sa-4.0" pipeline_tag: "token-classification" widget: - text: "国境の長いトンネルを抜けると雪国であった。" --- # bert-base-japanese-unidic-luw-upos ## Model Description This is a BERT model pre-trained on Japanese Wikipedia texts for POS-tagging and dependency-parsing, derived from [bert-base-japanese-v2](https://huggingface.co/cl-tohoku/bert-base-japanese-v2). Every long-unit-word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech). ## How to Use ```py import torch from transformers import AutoTokenizer,AutoModelForTokenClassification tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/bert-base-japanese-unidic-luw-upos") model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/bert-base-japanese-unidic-luw-upos") s="国境の長いトンネルを抜けると雪国であった。" t=tokenizer.tokenize(s) p=[model.config.id2label[q] for q in torch.argmax(model(tokenizer.encode(s,return_tensors="pt"))["logits"],dim=2)[0].tolist()[1:-1]] print(list(zip(t,p))) ``` or ```py import esupar nlp=esupar.load("KoichiYasuoka/bert-base-japanese-unidic-luw-upos") print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` [fugashi](https://pypi.org/project/fugashi), [unidic-lite](https://pypi.org/project/unidic-lite) and [pytokenizations](https://pypi.org/project/pytokenizations) are required. ## Reference 安岡孝一: [Transformersと国語研長単位による日本語係り受け解析モデルの製作](http://id.nii.ac.jp/1001/00216223/), 情報処理学会研究報告, Vol.2022-CH-128, No.7 (2022年2月), pp.1-8. ## See Also [esupar](https://github.com/KoichiYasuoka/esupar): Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa models

Bala/model_name

--- language: - "ja" tags: - "japanese" - "token-classification" - "pos" - "wikipedia" - "dependency-parsing" datasets: - "universal_dependencies" license: "cc-by-sa-4.0" pipeline_tag: "token-classification" widget: - text: "国境の長いトンネルを抜けると雪国であった。" --- # bert-base-japanese-upos ## Model Description This is a BERT model pre-trained on Japanese Wikipedia texts for POS-tagging and dependency-parsing, derived from [bert-base-japanese-char-extended](https://huggingface.co/KoichiYasuoka/bert-base-japanese-char-extended). Every short-unit-word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech). ## How to Use ```py import torch from transformers import AutoTokenizer,AutoModelForTokenClassification tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/bert-base-japanese-upos") model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/bert-base-japanese-upos") s="国境の長いトンネルを抜けると雪国であった。" p=[model.config.id2label[q] for q in torch.argmax(model(tokenizer.encode(s,return_tensors="pt"))["logits"],dim=2)[0].tolist()[1:-1]] print(list(zip(s,p))) ``` or ```py import esupar nlp=esupar.load("KoichiYasuoka/bert-base-japanese-upos") print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` ## See Also [esupar](https://github.com/KoichiYasuoka/esupar): Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa models

BalajiSathesh/DialoGPT-small-harrypotter

conversational

--- language: - "th" tags: - "thai" - "token-classification" - "pos" - "wikipedia" - "dependency-parsing" datasets: - "universal_dependencies" license: "apache-2.0" pipeline_tag: "token-classification" widget: - text: "หลายหัวดีกว่าหัวเดียว" --- # bert-base-thai-upos ## Model Description This is a BERT model pre-trained on Thai Wikipedia texts for POS-tagging and dependency-parsing, derived from [bert-base-th-cased](https://huggingface.co/Geotrend/bert-base-th-cased). Every word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech). ## How to Use ```py from transformers import AutoTokenizer,AutoModelForTokenClassification tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/bert-base-thai-upos") model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/bert-base-thai-upos") ``` or ```py import esupar nlp=esupar.load("KoichiYasuoka/bert-base-thai-upos") ``` ## See Also [esupar](https://github.com/KoichiYasuoka/esupar): Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa models

Balgow/prod_desc

--- language: - "ja" tags: - "japanese" - "masked-lm" - "wikipedia" license: "cc-by-sa-4.0" pipeline_tag: "fill-mask" mask_token: "[MASK]" widget: - text: "酸素ボンベを充[MASK]する。" --- # bert-large-japanese-char-extended ## Model Description This is a BERT model pre-trained on Japanese Wikipedia texts, derived from [bert-large-japanese-char](https://huggingface.co/cl-tohoku/bert-large-japanese-char). Character-embeddings are enhanced to include all 常用漢字/人名用漢字 characters using BertTokenizerFast. You can fine-tune `bert-large-japanese-char-extended` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/bert-large-japanese-upos), [dependency-parsing](https://huggingface.co/KoichiYasuoka/bert-large-japanese-wikipedia-ud-head), and so on. ## How to Use ```py from transformers import AutoTokenizer,AutoModelForMaskedLM tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/bert-large-japanese-char-extended") model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/bert-large-japanese-char-extended") ```

Banshee/LukeSkywalker

--- language: - "ja" tags: - "japanese" - "token-classification" - "pos" - "wikipedia" - "dependency-parsing" datasets: - "universal_dependencies" license: "cc-by-sa-4.0" pipeline_tag: "token-classification" widget: - text: "国境の長いトンネルを抜けると雪国であった。" --- # bert-large-japanese-luw-upos ## Model Description This is a BERT model pre-trained on Japanese Wikipedia texts for POS-tagging and dependency-parsing, derived from [bert-large-japanese-char-extended](https://huggingface.co/KoichiYasuoka/bert-large-japanese-char-extended). Every long-unit-word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech) and [FEATS](https://universaldependencies.org/u/feat/). ## How to Use ```py import torch from transformers import AutoTokenizer,AutoModelForTokenClassification tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/bert-large-japanese-luw-upos") model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/bert-large-japanese-luw-upos") s="国境の長いトンネルを抜けると雪国であった。" p=[model.config.id2label[q] for q in torch.argmax(model(tokenizer.encode(s,return_tensors="pt"))["logits"],dim=2)[0].tolist()[1:-1]] print(list(zip(s,p))) ``` or ```py import esupar nlp=esupar.load("KoichiYasuoka/bert-large-japanese-luw-upos") print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` ## Reference 安岡孝一: [Transformersと国語研長単位による日本語係り受け解析モデルの製作](http://id.nii.ac.jp/1001/00216223/), 情報処理学会研究報告, Vol.2022-CH-128, No.7 (2022年2月), pp.1-8. ## See Also [esupar](https://github.com/KoichiYasuoka/esupar): Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa models

Banshee/dialoGPT-small-luke

--- language: - "ja" tags: - "japanese" - "token-classification" - "pos" - "wikipedia" - "dependency-parsing" datasets: - "universal_dependencies" license: "cc-by-sa-4.0" pipeline_tag: "token-classification" widget: - text: "国境の長いトンネルを抜けると雪国であった。" --- # bert-large-japanese-upos ## Model Description This is a BERT model pre-trained on Japanese Wikipedia texts for POS-tagging and dependency-parsing, derived from [bert-large-japanese-char-extended](https://huggingface.co/KoichiYasuoka/bert-large-japanese-char-extended). Every short-unit-word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech). ## How to Use ```py import torch from transformers import AutoTokenizer,AutoModelForTokenClassification tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/bert-large-japanese-upos") model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/bert-large-japanese-upos") s="国境の長いトンネルを抜けると雪国であった。" p=[model.config.id2label[q] for q in torch.argmax(model(tokenizer.encode(s,return_tensors="pt"))["logits"],dim=2)[0].tolist()[1:-1]] print(list(zip(s,p))) ``` or ```py import esupar nlp=esupar.load("KoichiYasuoka/bert-large-japanese-upos") print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` ## See Also [esupar](https://github.com/KoichiYasuoka/esupar): Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa models

BatuhanYilmaz/bert-finetuned-ner

--- language: - "th" tags: - "thai" - "masked-lm" - "wikipedia" license: "apache-2.0" pipeline_tag: "fill-mask" mask_token: "[MASK]" --- # roberta-base-thai-spm ## Model Description This is a RoBERTa model pre-trained on Thai Wikipedia texts. You can fine-tune `roberta-base-thai-spm` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/roberta-base-thai-spm-upos), [dependency-parsing](https://huggingface.co/KoichiYasuoka/roberta-base-thai-spm-ud-head), and so on. ## How to Use ```py from transformers import AutoTokenizer,AutoModelForMaskedLM tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-base-thai-spm") model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/roberta-base-thai-spm") ```

BatuhanYilmaz/code-search-net-tokenizer1

--- language: - "th" tags: - "thai" - "masked-lm" - "wikipedia" license: "apache-2.0" pipeline_tag: "fill-mask" mask_token: "<mask>" widget: - text: "แผนกนี้กำลัง<mask>กับความท้าทายใหม่" --- # roberta-base-thai-syllable ## Model Description This is a RoBERTa model pre-trained on Thai Wikipedia texts, derived from [wangchanberta-base-wiki-syllable](https://huggingface.co/airesearch/wangchanberta-base-wiki-syllable). Character-embeddings are modified to use BertTokenizerFast. You can fine-tune `roberta-base-thai-syllable` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/roberta-base-thai-syllable-upos), [dependency-parsing](https://huggingface.co/KoichiYasuoka/roberta-base-thai-syllable-ud-goeswith), and so on. ## How to Use ```py from transformers import AutoTokenizer,AutoModelForMaskedLM tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-base-thai-syllable") model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/roberta-base-thai-syllable") ```

BatuhanYilmaz/distilbert-base-uncased-finetuned-squad-d5716d28

question-answering

--- language: - "lzh" tags: - "classical chinese" - "literary chinese" - "ancient chinese" - "masked-lm" license: "apache-2.0" pipeline_tag: "fill-mask" mask_token: "[MASK]" widget: - text: "孟子[MASK]梁惠王" --- # roberta-classical-chinese-base-char ## Model Description This is a RoBERTa model pre-trained on Classical Chinese texts, derived from [GuwenBERT-base](https://huggingface.co/ethanyt/guwenbert-base). Character-embeddings are enhanced into traditional/simplified characters. You can fine-tune `roberta-classical-chinese-base-char` for downstream tasks, such as [sentence-segmentation](https://huggingface.co/KoichiYasuoka/roberta-classical-chinese-base-sentence-segmentation), [POS-tagging](https://huggingface.co/KoichiYasuoka/roberta-classical-chinese-base-upos), [dependency-parsing](https://huggingface.co/KoichiYasuoka/roberta-classical-chinese-base-ud-goeswith), and so on. ## How to Use ```py from transformers import AutoTokenizer,AutoModelForMaskedLM tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-classical-chinese-base-char") model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/roberta-classical-chinese-base-char") ``` ## See Also [SuPar-Kanbun](https://github.com/KoichiYasuoka/SuPar-Kanbun): Tokenizer POS-tagger and Dependency-parser for Classical Chinese

BatuhanYilmaz/dummy-model

fill-mask

--- language: - "lzh" tags: - "classical chinese" - "literary chinese" - "ancient chinese" - "sentence segmentation" - "token-classification" license: "apache-2.0" pipeline_tag: "token-classification" widget: - text: "子曰學而時習之不亦説乎有朋自遠方來不亦樂乎人不知而不慍不亦君子乎" --- # roberta-classical-chinese-base-sentence-segmentation ## Model Description This is a RoBERTa model pre-trained on Classical Chinese texts for sentence segmentation, derived from [roberta-classical-chinese-base-char](https://huggingface.co/KoichiYasuoka/roberta-classical-chinese-base-char). Every segmented sentence begins with token-class "B" and ends with token-class "E" (except for single-character sentence with token-class "S"). ## How to Use ```py import torch from transformers import AutoTokenizer,AutoModelForTokenClassification tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-classical-chinese-base-sentence-segmentation") model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/roberta-classical-chinese-base-sentence-segmentation") s="子曰學而時習之不亦説乎有朋自遠方來不亦樂乎人不知而不慍不亦君子乎" p=[model.config.id2label[q] for q in torch.argmax(model(tokenizer.encode(s,return_tensors="pt"))["logits"],dim=2)[0].tolist()[1:-1]] print("".join(c+"。" if q=="E" or q=="S" else c for c,q in zip(s,p))) ``` ## Reference Koichi Yasuoka: [Sentence Segmentation of Classical Chinese Texts Using Transformers and BERT/RoBERTa Models](http://hdl.handle.net/2433/266539), IPSJ Symposium Series, Vol.2021, No.1 (December 2021), pp.104-109.

BeIR/query-gen-msmarco-t5-large-v1

text2text-generation

--- language: - "ja" tags: - "japanese" - "token-classification" - "pos" - "dependency-parsing" datasets: - "universal_dependencies" license: "cc-by-sa-4.0" pipeline_tag: "token-classification" widget: - text: "国境の長いトンネルを抜けると雪国であった。" --- # roberta-large-japanese-luw-upos ## Model Description This is a RoBERTa model pre-trained on 青空文庫 texts for POS-tagging and dependency-parsing, derived from [roberta-large-japanese-aozora](https://huggingface.co/KoichiYasuoka/roberta-large-japanese-aozora). Every long-unit-word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech). ## How to Use ```py from transformers import AutoTokenizer,AutoModelForTokenClassification,TokenClassificationPipeline tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-large-japanese-luw-upos") model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/roberta-large-japanese-luw-upos") pipeline=TokenClassificationPipeline(tokenizer=tokenizer,model=model,aggregation_strategy="simple") nlp=lambda x:[(x[t["start"]:t["end"]],t["entity_group"]) for t in pipeline(x)] print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` or ```py import esupar nlp=esupar.load("KoichiYasuoka/roberta-large-japanese-luw-upos") print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` ## Reference 安岡孝一: [Transformersと国語研長単位による日本語係り受け解析モデルの製作](http://id.nii.ac.jp/1001/00216223/), 情報処理学会研究報告, Vol.2022-CH-128, No.7 (2022年2月), pp.1-8. ## See Also [esupar](https://github.com/KoichiYasuoka/esupar): Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa models

BeIR/sparta-msmarco-distilbert-base-v1

feature-extraction

--- language: - "ja" tags: - "japanese" - "masked-lm" license: "cc-by-sa-4.0" pipeline_tag: "fill-mask" mask_token: "[MASK]" widget: - text: "日本に着いたら[MASK]を訪ねなさい。" --- # roberta-small-japanese-aozora-char ## Model Description This is a RoBERTa model pre-trained on 青空文庫 texts with character tokenizer. You can fine-tune `roberta-small-japanese-aozora-char` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/roberta-small-japanese-char-luw-upos), dependency-parsing, and so on. ## How to Use ```py from transformers import AutoTokenizer,AutoModelForMaskedLM tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-small-japanese-aozora-char") model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/roberta-small-japanese-aozora-char") ```

BearThreat/distilbert-base-uncased-finetuned-cola

text-classification

--- language: - "ja" tags: - "japanese" - "masked-lm" license: "cc-by-sa-4.0" pipeline_tag: "fill-mask" mask_token: "[MASK]" widget: - text: "日本に着いたら[MASK]を訪ねなさい。" --- # roberta-small-japanese-aozora ## Model Description This is a RoBERTa model pre-trained on 青空文庫 texts with [Japanese-LUW-Tokenizer](https://github.com/KoichiYasuoka/Japanese-LUW-Tokenizer). You can fine-tune `roberta-small-japanese-aozora` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/roberta-small-japanese-luw-upos), dependency-parsing, and so on. ## How to Use ```py from transformers import AutoTokenizer,AutoModelForMaskedLM tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-small-japanese-aozora") model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/roberta-small-japanese-aozora") ```

Beatriz/model_name

--- language: - "ja" tags: - "japanese" - "token-classification" - "pos" - "dependency-parsing" datasets: - "universal_dependencies" license: "cc-by-sa-4.0" pipeline_tag: "token-classification" widget: - text: "国境の長いトンネルを抜けると雪国であった。" --- # roberta-small-japanese-char-luw-upos ## Model Description This is a RoBERTa model pre-trained on 青空文庫 texts for POS-tagging and dependency-parsing, derived from [roberta-small-japanese-aozora-char](https://huggingface.co/KoichiYasuoka/roberta-small-japanese-aozora-char). Every long-unit-word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech). ## How to Use ```py from transformers import AutoTokenizer,AutoModelForTokenClassification,TokenClassificationPipeline tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-small-japanese-char-luw-upos") model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/roberta-small-japanese-char-luw-upos") pipeline=TokenClassificationPipeline(tokenizer=tokenizer,model=model,aggregation_strategy="simple") nlp=lambda x:[(x[t["start"]:t["end"]],t["entity_group"]) for t in pipeline(x)] print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` or ```py import esupar nlp=esupar.load("KoichiYasuoka/roberta-small-japanese-char-luw-upos") print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` ## See Also [esupar](https://github.com/KoichiYasuoka/esupar): Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa models

Bee-Garbs/DialoGPT-cartman-small

--- language: - "ja" tags: - "japanese" - "token-classification" - "pos" - "dependency-parsing" datasets: - "universal_dependencies" license: "cc-by-sa-4.0" pipeline_tag: "token-classification" widget: - text: "国境の長いトンネルを抜けると雪国であった。" --- # roberta-small-japanese-luw-upos ## Model Description This is a RoBERTa model pre-trained on 青空文庫 texts for POS-tagging and dependency-parsing, derived from [roberta-small-japanese-aozora](https://huggingface.co/KoichiYasuoka/roberta-small-japanese-aozora). Every long-unit-word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech). ## How to Use ```py from transformers import AutoTokenizer,AutoModelForTokenClassification,TokenClassificationPipeline tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-small-japanese-luw-upos") model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/roberta-small-japanese-luw-upos") pipeline=TokenClassificationPipeline(tokenizer=tokenizer,model=model,aggregation_strategy="simple") nlp=lambda x:[(x[t["start"]:t["end"]],t["entity_group"]) for t in pipeline(x)] print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` or ```py import esupar nlp=esupar.load("KoichiYasuoka/roberta-small-japanese-luw-upos") print(nlp("国境の長いトンネルを抜けると雪国であった。")) ``` ## See Also [esupar](https://github.com/KoichiYasuoka/esupar): Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa models

Beelow/model

--- tags: - conversational --- #Harry Potter DialoGPT Model

Beelow/wav2vec2-ukrainian-model-large

--- language: el widget: - text: "μπαινω στο <mask> και τι να δω." --- # Α lite RoBERTa fill mask model trained mostly in greek tweets The training dataset of this model consists of 23 million tweets in Greek, of approximately 5000 users in total, spanning from 2008 to 2018. The model has been trained to support the work for the paper [Multimodal Hate Speech Detection in Greek Social Media](https://www.mdpi.com/2414-4088/5/7/34) ## Load the pretrained model ```python from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("Konstantinos/BERTaTweetGR") model = AutoModel.from_pretrained("Konstantinos/BERTaTweetGR") ```

Begimay/Task

from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("r3dhummingbird/DialoGPT-medium-joshua") model = AutoModelForCausalLM.from_pretrained("r3dhummingbird/DialoGPT-medium-joshua")

BigSalmon/BlankSlots

text2text-generation

--- tags: - conversational --- # Rick Sanchez DialoGPT Model

BigSalmon/GPTNeo350MInformalToFormalLincoln5

text-generation

This is **KOREAN** Bert Masked LM pretrained model adapted in **BEAUTY** domain. (BertForMaskedLM) About 60,000 reviews were used. It was fine-tuned based on _beomi/kcbert-base_ model weights. Enjoy!

BigSalmon/MrLincoln7

--- tags: - translation - torch==1.8.0 widget: - text: "Inference Unavailable" --- ### marianmt-th-zh_cn * source languages: th * target languages: zh_cn * dataset: * model: transformer-align * pre-processing: normalization + SentencePiece * test set scores: 15.53 ## Training Training scripts from [LalitaDeelert/NLP-ZH_TH-Project](https://github.com/LalitaDeelert/NLP-ZH_TH-Project). Experiments tracked at [cstorm125/marianmt-th-zh_cn](https://wandb.ai/cstorm125/marianmt-th-zh_cn). ``` export WANDB_PROJECT=marianmt-th-zh_cn python train_model.py --input_fname ../data/v1/Train.csv \\\\\\\\ \\\\t--output_dir ../models/marianmt-th-zh_cn \\\\\\\\ \\\\t--source_lang th --target_lang zh \\\\\\\\ \\\\t--metric_tokenize zh --fp16 ``` ## Usage ``` from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("Lalita/marianmt-zh_cn-th") model = AutoModelForSeq2SeqLM.from_pretrained("Lalita/marianmt-zh_cn-th").cpu() src_text = [ 'ฉันรักคุณ', 'ฉันอยากกินข้าว', ] translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True)) print([tokenizer.decode(t, skip_special_tokens=True) for t in translated]) > ['我爱你', '我想吃饭。'] ``` ## Requirements ``` transformers==4.6.0 torch==1.8.0 ```

BigSalmon/MrLincoln8

text-generation

--- tags: - translation - torch==1.8.0 widget: - text: "Inference Unavailable" --- ### marianmt-zh_cn-th * source languages: zh_cn * target languages: th * dataset: * model: transformer-align * pre-processing: normalization + SentencePiece * test set scores: syllable: 15.95, word: 8.43 ## Training Training scripts from [LalitaDeelert/NLP-ZH_TH-Project](https://github.com/LalitaDeelert/NLP-ZH_TH-Project). Experiments tracked at [cstorm125/marianmt-zh_cn-th](https://wandb.ai/cstorm125/marianmt-zh_cn-th). ``` export WANDB_PROJECT=marianmt-zh_cn-th python train_model.py --input_fname ../data/v1/Train.csv \\\\\\\\\\\\\\\\ \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\t--output_dir ../models/marianmt-zh_cn-th \\\\\\\\\\\\\\\\ \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\t--source_lang zh --target_lang th \\\\\\\\\\\\\\\\ \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\t--metric_tokenize th_syllable --fp16 ``` ## Usage ``` from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("Lalita/marianmt-zh_cn-th") model = AutoModelForSeq2SeqLM.from_pretrained("Lalita/marianmt-zh_cn-th").cpu() src_text = [ '我爱你', '我想吃米饭', ] translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True)) print([tokenizer.decode(t, skip_special_tokens=True) for t in translated]) > ['ผมรักคุณนะ', 'ฉันอยากกินข้าว'] ``` ## Requirements ``` transformers==4.6.0 torch==1.8.0 ```

BigSalmon/Robertsy

fill-mask

--- license: apache-2.0 tags: - generated_from_trainer datasets: - Langame/starter model-index: - name: distilgpt2-starter results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilgpt2-starter This model is a fine-tuned version of [distilgpt2](https://huggingface.co/distilgpt2) on the Langame/starter dataset. It achieves the following results on the evaluation set: - Loss: 6.0234 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 4 - eval_batch_size: 8 - seed: 42 - distributed_type: multi-GPU - gradient_accumulation_steps: 2 - total_train_batch_size: 8 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 500.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:------:|:----:|:---------------:| | No log | 66.67 | 200 | 3.6445 | | No log | 133.33 | 400 | 4.5703 | | 1.0101 | 200.0 | 600 | 5.2109 | | 1.0101 | 266.67 | 800 | 5.5430 | | 0.0681 | 333.33 | 1000 | 5.7227 | | 0.0681 | 400.0 | 1200 | 5.8672 | | 0.0681 | 466.67 | 1400 | 5.9961 | ### Framework versions - Transformers 4.17.0.dev0 - Pytorch 1.10.0+cu111 - Datasets 1.18.1 - Tokenizers 0.11.0

BigSalmon/T5F

text2text-generation

--- language: - zh license: apache-2.0 --- # Mengzi-BERT base fin model (Chinese) Continue trained mengzi-bert-base with 20G financial news and research reports. Masked language modeling(MLM), part-of-speech(POS) tagging and sentence order prediction(SOP) are used as training task. [Mengzi: Towards Lightweight yet Ingenious Pre-trained Models for Chinese](https://arxiv.org/abs/2110.06696) ## Usage ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained("Langboat/mengzi-bert-base-fin") model = BertModel.from_pretrained("Langboat/mengzi-bert-base-fin") ``` ## Citation If you find the technical report or resource is useful, please cite the following technical report in your paper. ``` @misc{zhang2021mengzi, title={Mengzi: Towards Lightweight yet Ingenious Pre-trained Models for Chinese}, author={Zhuosheng Zhang and Hanqing Zhang and Keming Chen and Yuhang Guo and Jingyun Hua and Yulong Wang and Ming Zhou}, year={2021}, eprint={2110.06696}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

BigSalmon/T5Salmon

text2text-generation

--- language: - zh license: apache-2.0 widget: - text: "生活的真谛是[MASK]。" --- # Mengzi-BERT base model (Chinese) Pretrained model on 300G Chinese corpus. Masked language modeling(MLM), part-of-speech(POS) tagging and sentence order prediction(SOP) are used as training task. [Mengzi: A lightweight yet Powerful Chinese Pre-trained Language Model](https://arxiv.org/abs/2110.06696) ## Usage ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained("Langboat/mengzi-bert-base") model = BertModel.from_pretrained("Langboat/mengzi-bert-base") ``` ## Scores on nine chinese tasks (without any data augmentation) | Model | AFQMC | TNEWS | IFLYTEK | CMNLI | WSC | CSL | CMRC2018 | C3 | CHID | |-|-|-|-|-|-|-|-|-|-| |RoBERTa-wwm-ext| 74.30 | 57.51 | 60.80 | 80.70 | 67.20 | 80.67 | 77.59 | 67.06 | 83.78 | |Mengzi-BERT-base| 74.58 | 57.97 | 60.68 | 82.12 | 87.50 | 85.40 | 78.54 | 71.70 | 84.16 | RoBERTa-wwm-ext scores are from CLUE baseline ## Citation If you find the technical report or resource is useful, please cite the following technical report in your paper. ``` @misc{zhang2021mengzi, title={Mengzi: Towards Lightweight yet Ingenious Pre-trained Models for Chinese}, author={Zhuosheng Zhang and Hanqing Zhang and Keming Chen and Yuhang Guo and Jingyun Hua and Yulong Wang and Ming Zhou}, year={2021}, eprint={2110.06696}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

BinksSachary/ShaxxBot2

conversational

--- license: apache-2.0 tags: - generated_from_trainer datasets: - amazon_reviews_multi metrics: - accuracy model-index: - name: results results: - task: name: Text Classification type: text-classification dataset: name: amazon_reviews_multi type: amazon_reviews_multi args: es metrics: - name: Accuracy type: accuracy value: 0.8404 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # results This model is a fine-tuned version of [BSC-TeMU/roberta-base-bne](https://huggingface.co/BSC-TeMU/roberta-base-bne) on the amazon_reviews_multi dataset. It achieves the following results on the evaluation set: - Loss: 0.3793 - Accuracy: 0.8404 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 0.3542 | 1.0 | 125 | 0.3611 | 0.839 | | 0.2255 | 2.0 | 250 | 0.3793 | 0.8404 | ### Framework versions - Transformers 4.11.3 - Pytorch 1.9.0+cu111 - Datasets 1.12.1 - Tokenizers 0.10.3

Blaine-Mason/hackMIT-finetuned-sst2

text-classification

--- language: "fr" thumbnail: tags: - wav2vec2 license: "apache-2.0" --- # LeBenchmark: wav2vec2 large model trained on 1K hours of French speech LeBenchmark provides an ensemble of pretrained wav2vec2 models on different French datasets containing spontaneous, read, and broadcasted speech. For more information on the different benchmarks that can be used to evaluate the wav2vec2 models, please refer to our paper at: [Task Agnostic and Task Specific Self-Supervised Learning from Speech with LeBenchmark](https://openreview.net/pdf?id=TSvj5dmuSd) ## Model and data descriptions We release four different models that can be found under our HuggingFace organization. Two different wav2vec2 architectures *Base* and *Large* are coupled with our small (1K), medium (3K), and large (7K) corpus. A larger one should come later. In short: - [wav2vec2-FR-7K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-7K-large): Large wav2vec2 trained on 7.6K hours of French speech (1.8K Males / 1.0K Females / 4.8K unknown). - [wav2vec2-FR-7K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-7K-base): Base wav2vec2 trained on 7.6K hours of French speech (1.8K Males / 1.0K Females / 4.8K unknown). - [wav2vec2-FR-3K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-3K-large): Large wav2vec2 trained on 2.9K hours of French speech (1.8K Males / 1.0K Females / 0.1K unknown). - [wav2vec2-FR-3K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-3K-base): Base wav2vec2 trained on 2.9K hours of French speech (1.8K Males / 1.0K Females / 0.1K unknown). - [wav2vec2-FR-2.6K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-2.6K-base): Base wav2vec2 trained on 2.6K hours of French speech (**no spontaneous speech**). - [wav2vec2-FR-1K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-1K-large): Large wav2vec2 trained on 1K hours of French speech (0.5K Males / 0.5K Females). - [wav2vec2-FR-1K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-1K-base): Base wav2vec2 trained on 1K hours of French speech (0.5K Males / 0.5K Females). ## Intended uses & limitations Pretrained wav2vec2 models are distributed under the Apache-2.0 license. Hence, they can be reused extensively without strict limitations. However, benchmarks and data may be linked to corpora that are not completely open-sourced. ## Fine-tune with Fairseq for ASR with CTC As our wav2vec2 models were trained with Fairseq, then can be used in the different tools that they provide to fine-tune the model for ASR with CTC. The full procedure has been nicely summarized in [this blogpost](https://huggingface.co/blog/fine-tune-wav2vec2-english). Please note that due to the nature of CTC, speech-to-text results aren't expected to be state-of-the-art. Moreover, future features might appear depending on the involvement of Fairseq and HuggingFace on this part. ## Integrate to SpeechBrain for ASR, Speaker, Source Separation ... Pretrained wav2vec models recently gained in popularity. At the same time, [SpeechBrain toolkit](https://speechbrain.github.io) came out, proposing a new and simpler way of dealing with state-of-the-art speech & deep-learning technologies. While it currently is in beta, SpeechBrain offers two different ways of nicely integrating wav2vec2 models that were trained with Fairseq i.e our LeBenchmark models! 1. Extract wav2vec2 features on-the-fly (with a frozen wav2vec2 encoder) to be combined with any speech-related architecture. Examples are: E2E ASR with CTC+Att+Language Models; Speaker Recognition or Verification, Source Separation ... 2. *Experimental:* To fully benefit from wav2vec2, the best solution remains to fine-tune the model while you train your downstream task. This is very simply allowed within SpeechBrain as just a flag needs to be turned on. Thus, our wav2vec2 models can be fine-tuned while training your favorite ASR pipeline or Speaker Recognizer. **If interested, simply follow this [tutorial](https://colab.research.google.com/drive/17Hu1pxqhfMisjkSgmM2CnZxfqDyn2hSY?usp=sharing)** ## Referencing LeBenchmark ``` @article{Evain2021LeBenchmarkAR, title={LeBenchmark: A Reproducible Framework for Assessing Self-Supervised Representation Learning from Speech}, author={Sol{\`e}ne Evain and Ha Nguyen and Hang Le and Marcely Zanon Boito and Salima Mdhaffar and Sina Alisamir and Ziyi Tong and N. Tomashenko and Marco Dinarelli and Titouan Parcollet and A. Allauzen and Y. Est{\`e}ve and B. Lecouteux and F. Portet and S. Rossato and F. Ringeval and D. Schwab and L. Besacier}, journal={ArXiv}, year={2021}, volume={abs/2104.11462} } ```

Blazeolmo/Scrabunzi

--- language: "fr" thumbnail: tags: - wav2vec2 license: "apache-2.0" --- # LeBenchmark: wav2vec2 large model trained on 2.6K hours of French speech (no spontaneous speech) LeBenchmark provides an ensemble of pretrained wav2vec2 models on different French datasets containing spontaneous, read, and broadcasted speech. For more information on the different benchmarks that can be used to evaluate the wav2vec2 models, please refer to our paper at: [Task Agnostic and Task Specific Self-Supervised Learning from Speech with LeBenchmark](https://openreview.net/pdf?id=TSvj5dmuSd) ## Model and data descriptions We release four different models that can be found under our HuggingFace organization. Two different wav2vec2 architectures *Base* and *Large* are coupled with our small (1K), medium (3K), and large (7K) corpus. A larger one should come later. In short: - [wav2vec2-FR-7K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-7K-large): Large wav2vec2 trained on 7.6K hours of French speech (1.8K Males / 1.0K Females / 4.8K unknown). - [wav2vec2-FR-7K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-7K-base): Base wav2vec2 trained on 7.6K hours of French speech (1.8K Males / 1.0K Females / 4.8K unknown). - [wav2vec2-FR-3K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-3K-large): Large wav2vec2 trained on 2.9K hours of French speech (1.8K Males / 1.0K Females / 0.1K unknown). - [wav2vec2-FR-3K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-3K-base): Base wav2vec2 trained on 2.9K hours of French speech (1.8K Males / 1.0K Females / 0.1K unknown). - [wav2vec2-FR-2.6K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-2.6K-base): Base wav2vec2 trained on 2.6K hours of French speech (**no spontaneous speech**). - [wav2vec2-FR-1K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-1K-large): Large wav2vec2 trained on 1K hours of French speech (0.5K Males / 0.5K Females). - [wav2vec2-FR-1K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-1K-base): Base wav2vec2 trained on 1K hours of French speech (0.5K Males / 0.5K Females). ## Intended uses & limitations Pretrained wav2vec2 models are distributed under the Apache-2.0 license. Hence, they can be reused extensively without strict limitations. However, benchmarks and data may be linked to corpora that are not completely open-sourced. ## Fine-tune with Fairseq for ASR with CTC As our wav2vec2 models were trained with Fairseq, then can be used in the different tools that they provide to fine-tune the model for ASR with CTC. The full procedure has been nicely summarized in [this blogpost](https://huggingface.co/blog/fine-tune-wav2vec2-english). Please note that due to the nature of CTC, speech-to-text results aren't expected to be state-of-the-art. Moreover, future features might appear depending on the involvement of Fairseq and HuggingFace on this part. ## Integrate to SpeechBrain for ASR, Speaker, Source Separation ... Pretrained wav2vec models recently gained in popularity. At the same time, [SpeechBrain toolkit](https://speechbrain.github.io) came out, proposing a new and simpler way of dealing with state-of-the-art speech & deep-learning technologies. While it currently is in beta, SpeechBrain offers two different ways of nicely integrating wav2vec2 models that were trained with Fairseq i.e our LeBenchmark models! 1. Extract wav2vec2 features on-the-fly (with a frozen wav2vec2 encoder) to be combined with any speech-related architecture. Examples are: E2E ASR with CTC+Att+Language Models; Speaker Recognition or Verification, Source Separation ... 2. *Experimental:* To fully benefit from wav2vec2, the best solution remains to fine-tune the model while you train your downstream task. This is very simply allowed within SpeechBrain as just a flag needs to be turned on. Thus, our wav2vec2 models can be fine-tuned while training your favorite ASR pipeline or Speaker Recognizer. **If interested, simply follow this [tutorial](https://colab.research.google.com/drive/17Hu1pxqhfMisjkSgmM2CnZxfqDyn2hSY?usp=sharing)** ## Referencing LeBenchmark ``` @article{Evain2021LeBenchmarkAR, title={LeBenchmark: A Reproducible Framework for Assessing Self-Supervised Representation Learning from Speech}, author={Sol{\`e}ne Evain and Ha Nguyen and Hang Le and Marcely Zanon Boito and Salima Mdhaffar and Sina Alisamir and Ziyi Tong and N. Tomashenko and Marco Dinarelli and Titouan Parcollet and A. Allauzen and Y. Est{\`e}ve and B. Lecouteux and F. Portet and S. Rossato and F. Ringeval and D. Schwab and L. Besacier}, journal={ArXiv}, year={2021}, volume={abs/2104.11462} } ```

BlightZz/DialoGPT-medium-Kurisu

conversational

--- language: "fr" thumbnail: tags: - wav2vec2 license: "apache-2.0" --- # LeBenchmark: wav2vec2 large model trained on 3K hours of French speech LeBenchmark provides an ensemble of pretrained wav2vec2 models on different French dataset containing spontaneous, read and broadcasted speech. For more information on the different benchmark that can be used to evaluate the wav2vec2 models, please refer to our paper at: [Task Agnostic and Task Specific Self-Supervised Learning from Speech with LeBenchmark](https://openreview.net/pdf?id=TSvj5dmuSd) ## Model and data descriptions We release four different models that can be found under our HuggingFace organisation. Two different wav2vec2 architectures *Base* and *Large* are coupled with our small (*S*) and medium (*M*) corpus. A larger one shoud come later. In short: - [wav2vec2-FR-M-Large](#): Large wav2vec2 trained on 2.9K hours of French speech (1.8K Males / 1.0K Females / 0.1K unknown). - [wav2vec2-FR-M-Base](https://huggingface.co/LeBenchmark/wav2vec2-FR-M-base): Base wav2vec2 trained on 2.9K hours of French speech (1.8K Males / 1.0K Females / 0.1K unknown). - [wav2vec2-FR-S-Large](https://huggingface.co/LeBenchmark/wav2vec2-FR-S-large): Large wav2vec2 trained on 1K hours of French speech (0.5K Males / 0.5K Females). - [wav2vec2-FR-S-Base](https://huggingface.co/LeBenchmark/wav2vec2-FR-S-base): Base wav2vec2 trained on 1K hours of French speech (0.5K Males / 0.5K Females). ## Intended uses & limitations Pretrained wav2vec2 models are distributed under the apache-2.0 licence. Hence, they can be reused extensively without strict limitations. However, benchmarks and data may be linked to corpus that are not completely open-sourced. ## Fine-tune with Fairseq for ASR with CTC As our wav2vec2 models were trained with Fairseq, then can be used in the different tools that they provide to fine-tune the model for ASR with CTC. The full procedure has been nicely summarized in [this blogpost](https://huggingface.co/blog/fine-tune-wav2vec2-english). Please note that due to the nature of CTC, speech-to-text results aren't expected to be state-of-the-art. Moreover, future features might appear depending on the involvement of Fairseq and HuggingFace on this part. ## Integrate to SpeechBrain for ASR, Speaker, Source Separation ... Pretrained wav2vec models recently gained in popularity. At the same time [SpeechBrain toolkit](https://speechbrain.github.io) came out, proposing a new and simpler way of dealing with state-of-the-art speech & deep-learning technologies. While it currently is in beta, SpeechBrain offers two different ways of nicely integrating wav2vec2 models that were trained with Fairseq i.e our LeBenchmark models! 1. Extract wav2vec2 features on-the-fly (with a frozen wav2vec2 encoder) to be combined with any speech related architecture. Examples are: E2E ASR with CTC+Att+Language Models; Speaker Recognition or Verification, Source Separation ... 2. *Experimental:* To fully benefit from wav2vec2, the best solution remains to fine-tune the model while you train your downstream task. This is very simply allowed within SpeechBrain as just a flag needs to be turned on. Thus, our wav2vec2 models can be fine-tuned while training your favorite ASR pipeline or Speaker Recognizer. **If interested, simply follow this [tutorial](https://colab.research.google.com/drive/17Hu1pxqhfMisjkSgmM2CnZxfqDyn2hSY?usp=sharing)** ## Referencing LeBenchmark ``` Reference to come ```

BlindMan820/Sarcastic-News-Headlines

text-classification

--- language: "fr" thumbnail: tags: - wav2vec2 license: "apache-2.0" --- # LeBenchmark: wav2vec2 large model trained on 7K hours of French speech LeBenchmark provides an ensemble of pretrained wav2vec2 models on different French datasets containing spontaneous, read, and broadcasted speech. For more information on the different benchmarks that can be used to evaluate the wav2vec2 models, please refer to our paper at: [Task Agnostic and Task Specific Self-Supervised Learning from Speech with LeBenchmark](https://openreview.net/pdf?id=TSvj5dmuSd) ## Model and data descriptions We release four different models that can be found under our HuggingFace organization. Two different wav2vec2 architectures *Base* and *Large* are coupled with our small (1K), medium (3K), and large (7K) corpus. A larger one should come later. In short: - [wav2vec2-FR-7K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-7K-large): Large wav2vec2 trained on 7.6K hours of French speech (1.8K Males / 1.0K Females / 4.8K unknown). - [wav2vec2-FR-7K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-7K-base): Base wav2vec2 trained on 7.6K hours of French speech (1.8K Males / 1.0K Females / 4.8K unknown). - [wav2vec2-FR-3K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-3K-large): Large wav2vec2 trained on 2.9K hours of French speech (1.8K Males / 1.0K Females / 0.1K unknown). - [wav2vec2-FR-3K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-3K-base): Base wav2vec2 trained on 2.9K hours of French speech (1.8K Males / 1.0K Females / 0.1K unknown). - [wav2vec2-FR-2.6K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-2.6K-base): Base wav2vec2 trained on 2.6K hours of French speech (**no spontaneous speech**). - [wav2vec2-FR-1K-large](https://huggingface.co/LeBenchmark/wav2vec2-FR-1K-large): Large wav2vec2 trained on 1K hours of French speech (0.5K Males / 0.5K Females). - [wav2vec2-FR-1K-base](https://huggingface.co/LeBenchmark/wav2vec2-FR-1K-base): Base wav2vec2 trained on 1K hours of French speech (0.5K Males / 0.5K Females). ## Intended uses & limitations Pretrained wav2vec2 models are distributed under the Apache-2.0 license. Hence, they can be reused extensively without strict limitations. However, benchmarks and data may be linked to corpora that are not completely open-sourced. ## Fine-tune with Fairseq for ASR with CTC As our wav2vec2 models were trained with Fairseq, then can be used in the different tools that they provide to fine-tune the model for ASR with CTC. The full procedure has been nicely summarized in [this blogpost](https://huggingface.co/blog/fine-tune-wav2vec2-english). Please note that due to the nature of CTC, speech-to-text results aren't expected to be state-of-the-art. Moreover, future features might appear depending on the involvement of Fairseq and HuggingFace on this part. ## Integrate to SpeechBrain for ASR, Speaker, Source Separation ... Pretrained wav2vec models recently gained in popularity. At the same time, [SpeechBrain toolkit](https://speechbrain.github.io) came out, proposing a new and simpler way of dealing with state-of-the-art speech & deep-learning technologies. While it currently is in beta, SpeechBrain offers two different ways of nicely integrating wav2vec2 models that were trained with Fairseq i.e our LeBenchmark models! 1. Extract wav2vec2 features on-the-fly (with a frozen wav2vec2 encoder) to be combined with any speech-related architecture. Examples are: E2E ASR with CTC+Att+Language Models; Speaker Recognition or Verification, Source Separation ... 2. *Experimental:* To fully benefit from wav2vec2, the best solution remains to fine-tune the model while you train your downstream task. This is very simply allowed within SpeechBrain as just a flag needs to be turned on. Thus, our wav2vec2 models can be fine-tuned while training your favorite ASR pipeline or Speaker Recognizer. **If interested, simply follow this [tutorial](https://colab.research.google.com/drive/17Hu1pxqhfMisjkSgmM2CnZxfqDyn2hSY?usp=sharing)** ## Referencing LeBenchmark ``` @article{Evain2021LeBenchmarkAR, title={LeBenchmark: A Reproducible Framework for Assessing Self-Supervised Representation Learning from Speech}, author={Sol{\`e}ne Evain and Ha Nguyen and Hang Le and Marcely Zanon Boito and Salima Mdhaffar and Sina Alisamir and Ziyi Tong and N. Tomashenko and Marco Dinarelli and Titouan Parcollet and A. Allauzen and Y. Est{\`e}ve and B. Lecouteux and F. Portet and S. Rossato and F. Ringeval and D. Schwab and L. Besacier}, journal={ArXiv}, year={2021}, volume={abs/2104.11462} } ```

Botjallu/DialoGPT-small-harrypotter

--- license: apache-2.0 tags: - generated_from_trainer datasets: - common_voice model-index: - name: Wav2Vec2_xls_r_300m_hi_cv7 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Wav2Vec2_xls_r_300m_hi_cv7 This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice dataset. It achieves the following results on the evaluation set: - Loss: 0.6567 - Wer: 0.6273 - Cer: 0.2093 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 16 - eval_batch_size: 32 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 100 - num_epochs: 35 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Wer | Cer | |:-------------:|:-----:|:----:|:---------------:|:------:|:------:| | 5.6969 | 9.52 | 400 | 3.3092 | 1.0 | 0.9800 | | 1.7721 | 19.05 | 800 | 0.7769 | 0.7045 | 0.2367 | | 0.6384 | 28.57 | 1200 | 0.6567 | 0.6273 | 0.2093 | ### Framework versions - Transformers 4.17.0.dev0 - Pytorch 1.10.2+cu102 - Datasets 1.18.3 - Tokenizers 0.11.0

BotterHax/DialoGPT-small-harrypotter

conversational

--- language: - hi license: apache-2.0 tags: - automatic-speech-recognition - Openslr Multilingual - mozilla-foundation/common_voice_7_0 - generated_from_trainer model-index: - name: Wav2Vec2_xls_r_300m_hi_final results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Wav2Vec2_xls_r_300m_hi_final This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the ['Openslr Multilingual and code-switching ASR challenge'](http://www.openslr.org/103/) dataset and ['mozilla-foundation/common_voice_7_0'](https://huggingface.co/datasets/mozilla-foundation/common_voice_7_0) dataset. It achieves the following results on the evaluation set: - Loss: 0.3035 - Wer: 0.3137 - Cer: 0.0972 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 16 - eval_batch_size: 32 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 100 - num_epochs: 8 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Wer | Cer | |:-------------:|:-----:|:----:|:---------------:|:------:|:------:| | 0.9821 | 0.64 | 400 | 0.5059 | 0.4783 | 0.1573 | | 0.6861 | 1.28 | 800 | 0.4201 | 0.4247 | 0.1356 | | 0.585 | 1.92 | 1200 | 0.3797 | 0.3811 | 0.1210 | | 0.5193 | 2.56 | 1600 | 0.3577 | 0.3652 | 0.1152 | | 0.4583 | 3.21 | 2000 | 0.3422 | 0.3519 | 0.1111 | | 0.4282 | 3.85 | 2400 | 0.3261 | 0.3450 | 0.1071 | | 0.3951 | 4.49 | 2800 | 0.3201 | 0.3325 | 0.1048 | | 0.3619 | 5.13 | 3200 | 0.3167 | 0.3296 | 0.1030 | | 0.345 | 5.77 | 3600 | 0.3157 | 0.3210 | 0.1013 | | 0.338 | 6.41 | 4000 | 0.3051 | 0.3143 | 0.0982 | | 0.3155 | 7.05 | 4400 | 0.3059 | 0.3154 | 0.0986 | | 0.3057 | 7.69 | 4800 | 0.3035 | 0.3137 | 0.0972 | ### Framework versions - Transformers 4.17.0.dev0 - Pytorch 1.10.2+cu102 - Datasets 1.18.3 - Tokenizers 0.11.0

Branex/gpt-neo-2.7B

--- language: - hi license: apache-2.0 tags: - Openslr Multilingual - automatic-speech-recognition - generated_from_trainer - hf-asr-leaderboard - mozilla-foundation/common_voice_7_0 - robust-speech-event datasets: - mozilla-foundation/common_voice_7_0 model-index: - name: Wav2Vec2_xls_r_300m_hi_final results: - task: name: Automatic Speech Recognition type: automatic-speech-recognition dataset: name: Common Voice 7.0 type: mozilla-foundation/common_voice_7_0 args: hi metrics: - name: Test WER type: wer value: 34.21 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Wav2Vec2_xls_r_300m_hi_final This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the ['Openslr Multilingual and code-switching ASR challenge'](http://www.openslr.org/103/) dataset and ['mozilla-foundation/common_voice_7_0'](https://huggingface.co/datasets/mozilla-foundation/common_voice_7_0) dataset. It achieves the following results on the evaluation set: - Loss: 0.3035 - Wer: 0.3137 - Cer: 0.0972 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 16 - eval_batch_size: 32 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 100 - num_epochs: 8 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Wer | Cer | |:-------------:|:-----:|:----:|:---------------:|:------:|:------:| | 0.9821 | 0.64 | 400 | 0.5059 | 0.4783 | 0.1573 | | 0.6861 | 1.28 | 800 | 0.4201 | 0.4247 | 0.1356 | | 0.585 | 1.92 | 1200 | 0.3797 | 0.3811 | 0.1210 | | 0.5193 | 2.56 | 1600 | 0.3577 | 0.3652 | 0.1152 | | 0.4583 | 3.21 | 2000 | 0.3422 | 0.3519 | 0.1111 | | 0.4282 | 3.85 | 2400 | 0.3261 | 0.3450 | 0.1071 | | 0.3951 | 4.49 | 2800 | 0.3201 | 0.3325 | 0.1048 | | 0.3619 | 5.13 | 3200 | 0.3167 | 0.3296 | 0.1030 | | 0.345 | 5.77 | 3600 | 0.3157 | 0.3210 | 0.1013 | | 0.338 | 6.41 | 4000 | 0.3051 | 0.3143 | 0.0982 | | 0.3155 | 7.05 | 4400 | 0.3059 | 0.3154 | 0.0986 | | 0.3057 | 7.69 | 4800 | 0.3035 | 0.3137 | 0.0972 | ### Framework versions - Transformers 4.17.0.dev0 - Pytorch 1.10.2+cu102 - Datasets 1.18.3 - Tokenizers 0.11.0

Brayan/CNN_Brain_Tumor

--- license: apache-2.0 language: - hi tags: - automatic-speech-recognition - Harveenchadha/indic-voice - generated_from_trainer model-index: - name: Wav2Vec2_xls_r_openslr_Hi_V2 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Wav2Vec2_xls_r_openslr_Hi_V2 This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the [Harveenchadha/indic-voice](https://huggingface.co/datasets/Harveenchadha/indic-voice) dataset. It achieves the following results on the evaluation set: - Loss: 0.3184 - Wer: 0.3104 - Cer: 0.0958 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 16 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 200 - num_epochs: 12 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Cer | Validation Loss | Wer | |:-------------:|:-----:|:----:|:------:|:---------------:|:------:| | 7.1097 | 0.48 | 300 | 0.9965 | 3.3989 | 1.0 | | 3.0235 | 0.96 | 600 | 0.3163 | 1.3183 | 0.7977 | | 1.1419 | 1.44 | 900 | 0.1913 | 0.6416 | 0.5543 | | 0.8242 | 1.92 | 1200 | 0.1608 | 0.5063 | 0.4804 | | 0.6876 | 2.56 | 1600 | 0.1387 | 0.4401 | 0.4280 | | 0.5868 | 3.21 | 2000 | 0.1249 | 0.3940 | 0.3907 | | 0.5285 | 3.85 | 2400 | 0.1200 | 0.3661 | 0.3763 | | 0.5 | 4.49 | 2800 | 0.3528 | 0.3610 | 0.1136 | | 0.4538 | 5.13 | 3200 | 0.3403 | 0.3485 | 0.1086 | | 0.4165 | 5.77 | 3600 | 0.3335 | 0.3439 | 0.1062 | | 0.3989 | 6.41 | 4000 | 0.3264 | 0.3340 | 0.1036 | | 0.3679 | 7.05 | 4400 | 0.3256 | 0.3287 | 0.1013 | | 0.3517 | 7.69 | 4800 | 0.3212 | 0.3223 | 0.1002 | | 0.3357 | 8.33 | 5200 | 0.3173 | 0.3196 | 0.0986 | | 0.3225 | 8.97 | 5600 | 0.3142 | 0.3177 | 0.0985 | | 0.3057 | 9.62 | 6000 | 0.3199 | 0.3156 | 0.0975 | | 0.2972 | 10.26 | 6400 | 0.3139 | 0.3128 | 0.0967 | | 0.2881 | 10.9 | 6800 | 0.3184 | 0.3107 | 0.0957 | | 0.2791 | 11.54 | 7200 | 0.3184 | 0.3104 | 0.0958 | ### Framework versions - Transformers 4.17.0.dev0 - Pytorch 1.10.2+cu102 - Datasets 1.18.2.dev0 - Tokenizers 0.11.0

Broadus20/DialoGPT-small-harrypotter

text-generation

--- license: apache-2.0 tags: - generated_from_trainer datasets: - imdb model-index: - name: distilbert-base-uncased-finetuned-imdb results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-imdb This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 2.3114 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3.0 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | 2.5561 | 1.0 | 782 | 2.3738 | | 2.4474 | 2.0 | 1564 | 2.3108 | | 2.4037 | 3.0 | 2346 | 2.3017 | ### Framework versions - Transformers 4.12.5 - Pytorch 1.10.0 - Datasets 1.15.1 - Tokenizers 0.10.3

Brona/model1

--- license: apache-2.0 tags: - translation - generated_from_trainer datasets: - kde4 metrics: - bleu model-index: - name: marian-finetuned-kde4-en-to-fr results: - task: name: Sequence-to-sequence Language Modeling type: text2text-generation dataset: name: kde4 type: kde4 args: en-fr metrics: - name: Bleu type: bleu value: 52.94538305859332 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # marian-finetuned-kde4-en-to-fr This model is a fine-tuned version of [Helsinki-NLP/opus-mt-en-fr](https://huggingface.co/Helsinki-NLP/opus-mt-en-fr) on the kde4 dataset. It achieves the following results on the evaluation set: - Loss: 0.8558 - Bleu: 52.9454 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 32 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 - mixed_precision_training: Native AMP ### Training results ### Framework versions - Transformers 4.12.5 - Pytorch 1.10.0 - Datasets 1.15.1 - Tokenizers 0.10.3

BrunoNogueira/DialoGPT-kungfupanda

conversational

--- tags: - automatic-speech-recognition - generated_from_trainer license: mit language: - lb metrics: - wer pipeline_tag: automatic-speech-recognition model-index: - name: Lemswasabi/wav2vec2-large-xlsr-53-842h-luxembourgish-4h results: - task: type: automatic-speech-recognition # Required. Example: automatic-speech-recognition name: Speech Recognition # Optional. Example: Speech Recognition metrics: - type: wer value: 19.44 name: Dev WER - type: wer value: 18.77 name: Test WER - type: cer value: 7.16 name: Dev CER - type: cer value: 6.43 name: Test CER --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # ## Model description We fine-tuned a wav2vec 2.0 large XLSR-53 checkpoint with 842h of unlabelled Luxembourgish speech collected from [RTL.lu](https://www.rtl.lu/). Then the model was fine-tuned on 4h of labelled Luxembourgish speech from the same domain. ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 7.5e-05 - train_batch_size: 3 - eval_batch_size: 3 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 12 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 2000 - num_epochs: 50.0 - mixed_precision_training: Native AMP ### Framework versions - Transformers 4.20.0.dev0 - Pytorch 1.11.0+cu113 - Datasets 2.2.1 - Tokenizers 0.12.1 ## Citation This model is a result of our paper `IMPROVING LUXEMBOURGISH SPEECH RECOGNITION WITH CROSS-LINGUAL SPEECH REPRESENTATIONS` submitted to the [IEEE SLT 2022 workshop](https://slt2022.org/) ``` @misc{lb-wav2vec2, author = {Nguyen, Le Minh and Nayak, Shekhar and Coler, Matt.}, keywords = {Luxembourgish, multilingual speech recognition, language modelling, wav2vec 2.0 XLSR-53, under-resourced language}, title = {IMPROVING LUXEMBOURGISH SPEECH RECOGNITION WITH CROSS-LINGUAL SPEECH REPRESENTATIONS}, year = {2022}, copyright = {2023 IEEE} } ```

Bryan190/Aguy190

--- license: apache-2.0 tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-squad-Endpoint_with_impossible.csv results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-squad-Endpoint_with_impossible.csv This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.7950 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | 1.25 | 1.0 | 1273 | 0.8052 | | 1.1199 | 2.0 | 2546 | 0.7950 | ### Framework versions - Transformers 4.16.2 - Pytorch 1.10.0+cu111 - Datasets 1.18.3 - Tokenizers 0.11.0

Bryanwong/wangchanberta-ner

--- license: apache-2.0 tags: - generated_from_trainer model-index: - name: distilbert-base-uncased-finetuned-squad results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-squad This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 1.7713 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | 2.0325 | 1.0 | 585 | 1.7520 | | 1.609 | 2.0 | 1170 | 1.7713 | ### Framework versions - Transformers 4.16.2 - Pytorch 1.10.0+cu111 - Datasets 1.18.3 - Tokenizers 0.11.0

Brykee/BrykeeBot

--- license: apache-2.0 tags: - generated_from_trainer model-index: - name: distilgpt2-finetuned-wikitext2 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilgpt2-finetuned-wikitext2 This model is a fine-tuned version of [distilgpt2](https://huggingface.co/distilgpt2) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 3.6424 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | 3.7608 | 1.0 | 2334 | 3.6655 | | 3.6335 | 2.0 | 4668 | 3.6455 | | 3.6066 | 3.0 | 7002 | 3.6424 | ### Framework versions - Transformers 4.11.2 - Pytorch 1.9.0+cu102 - Tokenizers 0.10.3

Brykee/DialoGPT-medium-Morty

conversational

--- tags: - generated_from_trainer model-index: - name: first results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # first This model is a fine-tuned version of [longformer-gottbert-base-8192-aw512-](https://huggingface.co/longformer-8192-aw512-gottbert-base) on the a 500 million token subset of the german parts of the OSCAR dataset. It achieves the following results on the custom evaluation set: - Loss: 1.4981 ## Model description The weights of the model are initialized from the german version of Roberta [gottbert-base](https://huggingface.co/uklfr/gottbert-base). The local attention windows have a fixed size of 512 tokens across all layers. The maximum sequence length is 8192. ## Intended uses & limitations Longformer models enable processing long texts using a mixture of local attention on each subword token and task specific global attention on a subset of the tokens. ## Training and evaluation data The [OSCAR](https://oscar-corpus.com) dataset is freely avaible corpus of filtered web texts from the Common Crawl in various languages. We used the 2017 version of the dataset. ## Training procedure The model was trained with masked language modeling for 3 epochs on a customly created 500 million tokens subset of the german proportion of the [OSCAR](https://oscar-corpus.com) dataset. It was validated using 5% of the original subset. ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 2 - eval_batch_size: 4 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 500 - num_epochs: 3.0 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:-----:|:---------------:| | 2.5636 | 0.1 | 500 | 2.2399 | | 2.0426 | 0.2 | 1000 | 1.8841 | | 1.9653 | 0.3 | 1500 | 1.7807 | | 1.9422 | 0.4 | 2000 | 1.7206 | | 1.9323 | 0.49 | 2500 | 1.6800 | | 1.7587 | 0.59 | 3000 | 1.6507 | | 1.7239 | 0.69 | 3500 | 1.6316 | | 1.7452 | 0.79 | 4000 | 1.6137 | | 1.7415 | 0.89 | 4500 | 1.5983 | | 1.7733 | 0.99 | 5000 | 1.5830 | | 1.7656 | 1.09 | 5500 | 1.5735 | | 1.6543 | 1.19 | 6000 | 1.5643 | | 1.7131 | 1.28 | 6500 | 1.5546 | | 1.6456 | 1.38 | 7000 | 1.5503 | | 1.716 | 1.48 | 7500 | 1.5422 | | 1.806 | 1.58 | 8000 | 1.5377 | | 1.8407 | 1.68 | 8500 | 1.5327 | | 1.6371 | 1.78 | 9000 | 1.5278 | | 1.6453 | 1.88 | 9500 | 1.5231 | | 1.7754 | 1.98 | 10000 | 1.5214 | | 1.7695 | 2.08 | 10500 | 1.5165 | | 1.7109 | 2.17 | 11000 | 1.5138 | | 1.6992 | 2.27 | 11500 | 1.5107 | | 1.6707 | 2.37 | 12000 | 1.5097 | | 1.6835 | 2.47 | 12500 | 1.5040 | | 1.7171 | 2.57 | 13000 | 1.5041 | | 1.7257 | 2.67 | 13500 | 1.4990 | | 1.6287 | 2.77 | 14000 | 1.5017 | | 1.7737 | 2.87 | 14500 | 1.4983 | | 1.4002 | 2.96 | 15000 | 1.4992 | ### Framework versions - Transformers 4.15.0 - Pytorch 1.10.1+cu113 - Datasets 1.17.0 - Tokenizers 0.10.3