First version of bert2bert-fa-news-headline

README.md

@@ -0,0 +1,24 @@
+---
+language: fa
+license: apache-2.0
+tags:
+- summarization
+---
+
+A Bert2Bert model on VoA Persian Corpus (a medium-sized corpus of 7.9 million words, 2003-2008) generates headlines. The model achieved a 25.30 ROUGE-2 score. 
+ 
+For more detail, please follow the [News Headline Generation](https://github.com/m3hrdadfi/news-headline-generation) repo. 
+
+
+## Eval results
+The following table summarizes the ROUGE scores obtained by the Bert2Bert model.
+
+|    %    | Precision | Recall | FMeasure |
+|:-------:|:---------:|:------:|:--------:|
+| ROUGE-1 |   43.78   |  45.52 |   43.54  |
+| ROUGE-2 |   24.50   | 25.30* |   24.24  |
+| ROUGE-L |   41.20   |  42.22 |   40.76  |
+
+
+## Questions?
+Post a Github issue on the [News Headline Generation](https://github.com/hooshvare/news-headline-generation/issues) repo.

config.json

@@ -0,0 +1,145 @@
+{
+  "architectures": [
+    "EncoderDecoderModel"
+  ],
+  "decoder": {
+    "_name_or_path": "HooshvareLab/bert-fa-base-uncased",
+    "add_cross_attention": true,
+    "architectures": [
+      "BertForMaskedLM"
+    ],
+    "attention_probs_dropout_prob": 0.1,
+    "bad_words_ids": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "decoder_start_token_id": null,
+    "do_sample": false,
+    "early_stopping": false,
+    "eos_token_id": null,
+    "finetuning_task": null,
+    "gradient_checkpointing": false,
+    "hidden_act": "gelu",
+    "hidden_dropout_prob": 0.1,
+    "hidden_size": 768,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "intermediate_size": 3072,
+    "is_decoder": true,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-12,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "max_position_embeddings": 512,
+    "min_length": 0,
+    "model_type": "bert",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 12,
+    "num_beams": 1,
+    "num_hidden_layers": 12,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "pad_token_id": 0,
+    "prefix": null,
+    "pruned_heads": {},
+    "repetition_penalty": 1.0,
+    "return_dict": false,
+    "sep_token_id": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torchscript": false,
+    "type_vocab_size": 2,
+    "use_bfloat16": false,
+    "use_cache": true,
+    "vocab_size": 100000,
+    "xla_device": null
+  },
+  "decoder_start_token_id": 2,
+  "early_stopping": true,
+  "encoder": {
+    "_name_or_path": "HooshvareLab/bert-fa-base-uncased",
+    "add_cross_attention": false,
+    "architectures": [
+      "BertForMaskedLM"
+    ],
+    "attention_probs_dropout_prob": 0.1,
+    "bad_words_ids": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "decoder_start_token_id": null,
+    "do_sample": false,
+    "early_stopping": false,
+    "eos_token_id": null,
+    "finetuning_task": null,
+    "gradient_checkpointing": false,
+    "hidden_act": "gelu",
+    "hidden_dropout_prob": 0.1,
+    "hidden_size": 768,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "intermediate_size": 3072,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-12,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "max_position_embeddings": 512,
+    "min_length": 0,
+    "model_type": "bert",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 12,
+    "num_beams": 1,
+    "num_hidden_layers": 12,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "pad_token_id": 0,
+    "prefix": null,
+    "pruned_heads": {},
+    "repetition_penalty": 1.0,
+    "return_dict": false,
+    "sep_token_id": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torchscript": false,
+    "type_vocab_size": 2,
+    "use_bfloat16": false,
+    "use_cache": true,
+    "vocab_size": 100000,
+    "xla_device": null
+  },
+  "eos_token_id": 4,
+  "is_encoder_decoder": true,
+  "length_penalty": 2.0,
+  "max_length": 32,
+  "min_length": 5,
+  "model_type": "encoder-decoder",
+  "no_repeat_ngram_size": 3,
+  "num_beams": 3,
+  "pad_token_id": 0,
+  "vocab_size": 100000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76c4ad0427a424efb8b3c8df8499697c61d9bb2be263149150aadc41a12bad2d
+size 1416836998