Lalita commited on
Commit
420c402
1 Parent(s): 1ad0e87

first commit

Browse files
README.md ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - translation
4
+ - torch==1.8.0
5
+ widget:
6
+ - text: "Inference Unavailable"
7
+ ---
8
+ ### marianmt-zh_cn-th
9
+ * source languages: zh_cn
10
+ * target languages: th
11
+ * dataset:
12
+ * model: transformer-align
13
+ * pre-processing: normalization + SentencePiece
14
+ * test set translations:
15
+ * test set scores:
16
+
17
+ ## Training
18
+
19
+ 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).
20
+
21
+ ```
22
+ export WANDB_PROJECT=marianmt-zh_cn-th
23
+ python train_model.py --input_fname ../data/v1/Train.csv \
24
+ \\t--output_dir ../models/marianmt-zh_cn-th \
25
+ \\t--source_lang zh --target_lang th \
26
+ \\t--metric_tokenize th_syllable --fp16
27
+ ```
28
+
29
+ ## Usage
30
+
31
+ ```
32
+ from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
33
+
34
+ tokenizer = AutoTokenizer.from_pretrained("cstorm125/marianmt-zh_cn-th")
35
+ model = AutoModelForSeq2SeqLM.from_pretrained("cstorm125/marianmt-zh_cn-th").cpu()
36
+
37
+ src_text = [
38
+ '我爱你',
39
+ '我想吃米饭',
40
+ ]
41
+ translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
42
+ print([tokenizer.decode(t, skip_special_tokens=True) for t in translated])
43
+
44
+ > ['ผมรักคุณนะ', 'ฉันอยากกินข้าว']
45
+ ```
46
+
47
+ ## Requirements
48
+ ```
49
+ transformers==4.6.0
50
+ torch==1.8.0
51
+ ```
config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "modelx",
3
+ "_num_labels": 3,
4
+ "activation_dropout": 0.0,
5
+ "activation_function": "swish",
6
+ "add_bias_logits": false,
7
+ "add_final_layer_norm": false,
8
+ "architectures": [
9
+ "MarianMTModel"
10
+ ],
11
+ "attention_dropout": 0.0,
12
+ "bad_words_ids": [
13
+ [
14
+ 3
15
+ ]
16
+ ],
17
+ "bos_token_id": 1,
18
+ "classif_dropout": 0.0,
19
+ "classifier_dropout": 0.0,
20
+ "d_model": 512,
21
+ "decoder_attention_heads": 8,
22
+ "decoder_ffn_dim": 2048,
23
+ "decoder_layerdrop": 0.0,
24
+ "decoder_layers": 6,
25
+ "decoder_start_token_id": 3,
26
+ "dropout": 0.1,
27
+ "encoder_attention_heads": 8,
28
+ "encoder_ffn_dim": 2048,
29
+ "encoder_layerdrop": 0.0,
30
+ "encoder_layers": 6,
31
+ "eos_token_id": 2,
32
+ "forced_eos_token_id": 0,
33
+ "gradient_checkpointing": false,
34
+ "id2label": {
35
+ "0": "LABEL_0",
36
+ "1": "LABEL_1",
37
+ "2": "LABEL_2"
38
+ },
39
+ "init_std": 0.02,
40
+ "is_encoder_decoder": true,
41
+ "label2id": {
42
+ "LABEL_0": 0,
43
+ "LABEL_1": 1,
44
+ "LABEL_2": 2
45
+ },
46
+ "max_length": 512,
47
+ "max_position_embeddings": 512,
48
+ "model_type": "marian",
49
+ "normalize_before": false,
50
+ "normalize_embedding": false,
51
+ "num_beams": 4,
52
+ "num_hidden_layers": 6,
53
+ "pad_token_id": 3,
54
+ "scale_embedding": true,
55
+ "static_position_embeddings": true,
56
+ "transformers_version": "4.6.0",
57
+ "unk_token_id": 0,
58
+ "use_cache": true,
59
+ "vocab_size": 32001
60
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:43f3ca7183ed67b4a5bbc5adfb416460e837f9576b1c33fc04c29febe3a65240
3
+ size 242307293
source.spm ADDED
Binary file (993 kB). View file
 
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}
target.spm ADDED
Binary file (993 kB). View file
 
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_lang": null, "target_lang": null, "unk_token": "<unk>", "eos_token": "</s>", "pad_token": "<pad>", "model_max_length": 512, "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "modelx"}
vocab.json ADDED
The diff for this file is too large to render. See raw diff