Guocheng commited on
Commit
967e848
1 Parent(s): 24ea331
README.md ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ model_index:
4
+ - name: HelsinkiNLP-FineTuned-Legal-es-zh
5
+ results:
6
+ - task:
7
+ name: Sequence-to-sequence Language Modeling
8
+ type: text2text-generation
9
+ ---
10
+
11
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
12
+ should probably proofread and complete it, then remove this comment. -->
13
+
14
+ # HelsinkiNLP-FineTuned-Legal-es-zh
15
+
16
+ This model is a fine-tuned version of [Helsinki-NLP/opus-tatoeba-es-zh](https://huggingface.co/Helsinki-NLP/opus-tatoeba-es-zh) on an unkown dataset.
17
+ It achieves the following results on the evaluation set:
18
+ - Loss: 2.1078
19
+
20
+ ## Model description
21
+
22
+ More information needed
23
+
24
+ ## Intended uses & limitations
25
+
26
+ More information needed
27
+
28
+ ## Training and evaluation data
29
+
30
+ More information needed
31
+
32
+ ## Training procedure
33
+
34
+ ### Training hyperparameters
35
+
36
+ The following hyperparameters were used during training:
37
+ - learning_rate: 2e-05
38
+ - train_batch_size: 8
39
+ - eval_batch_size: 8
40
+ - seed: 42
41
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
42
+ - lr_scheduler_type: linear
43
+ - lr_scheduler_warmup_steps: 2000
44
+ - num_epochs: 10
45
+ - mixed_precision_training: Native AMP
46
+
47
+ ### Training results
48
+
49
+ | Training Loss | Epoch | Step | Validation Loss |
50
+ |:-------------:|:-----:|:----:|:---------------:|
51
+ | 2.9584 | 0.36 | 400 | 2.6800 |
52
+ | 2.6402 | 0.71 | 800 | 2.5017 |
53
+ | 2.5038 | 1.07 | 1200 | 2.3907 |
54
+ | 2.3279 | 1.43 | 1600 | 2.2999 |
55
+ | 2.2258 | 1.78 | 2000 | 2.2343 |
56
+ | 2.1061 | 2.14 | 2400 | 2.1961 |
57
+ | 1.9279 | 2.5 | 2800 | 2.1569 |
58
+ | 1.9059 | 2.85 | 3200 | 2.1245 |
59
+ | 1.7491 | 3.21 | 3600 | 2.1227 |
60
+ | 1.6301 | 3.57 | 4000 | 2.1169 |
61
+ | 1.6871 | 3.92 | 4400 | 2.0979 |
62
+ | 1.5203 | 4.28 | 4800 | 2.1074 |
63
+ | 1.4646 | 4.63 | 5200 | 2.1024 |
64
+ | 1.4739 | 4.99 | 5600 | 2.0905 |
65
+ | 1.338 | 5.35 | 6000 | 2.0946 |
66
+ | 1.3152 | 5.7 | 6400 | 2.0974 |
67
+ | 1.306 | 6.06 | 6800 | 2.0985 |
68
+ | 1.1991 | 6.42 | 7200 | 2.0962 |
69
+ | 1.2113 | 6.77 | 7600 | 2.1092 |
70
+ | 1.1983 | 7.13 | 8000 | 2.1060 |
71
+ | 1.1238 | 7.49 | 8400 | 2.1102 |
72
+ | 1.1417 | 7.84 | 8800 | 2.1078 |
73
+
74
+
75
+ ### Framework versions
76
+
77
+ - Transformers 4.7.0
78
+ - Pytorch 1.8.1+cu101
79
+ - Datasets 1.8.0
80
+ - Tokenizers 0.10.3
config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Helsinki-NLP/opus-tatoeba-es-zh",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "swish",
5
+ "add_bias_logits": false,
6
+ "add_final_layer_norm": false,
7
+ "architectures": [
8
+ "MarianMTModel"
9
+ ],
10
+ "attention_dropout": 0.0,
11
+ "bad_words_ids": [
12
+ [
13
+ 65000
14
+ ]
15
+ ],
16
+ "bos_token_id": 0,
17
+ "classifier_dropout": 0.0,
18
+ "d_model": 512,
19
+ "decoder_attention_heads": 8,
20
+ "decoder_ffn_dim": 2048,
21
+ "decoder_layerdrop": 0.0,
22
+ "decoder_layers": 6,
23
+ "decoder_start_token_id": 65000,
24
+ "do_blenderbot_90_layernorm": false,
25
+ "dropout": 0.1,
26
+ "encoder_attention_heads": 8,
27
+ "encoder_ffn_dim": 2048,
28
+ "encoder_layerdrop": 0.0,
29
+ "encoder_layers": 6,
30
+ "eos_token_id": 0,
31
+ "extra_pos_embeddings": 2,
32
+ "force_bos_token_to_be_generated": false,
33
+ "forced_eos_token_id": 0,
34
+ "gradient_checkpointing": false,
35
+ "id2label": {
36
+ "0": "LABEL_0",
37
+ "1": "LABEL_1",
38
+ "2": "LABEL_2"
39
+ },
40
+ "init_std": 0.02,
41
+ "is_encoder_decoder": true,
42
+ "label2id": {
43
+ "LABEL_0": 0,
44
+ "LABEL_1": 1,
45
+ "LABEL_2": 2
46
+ },
47
+ "max_length": 512,
48
+ "max_position_embeddings": 512,
49
+ "model_type": "marian",
50
+ "normalize_before": false,
51
+ "normalize_embedding": false,
52
+ "num_beams": 6,
53
+ "num_hidden_layers": 6,
54
+ "pad_token_id": 65000,
55
+ "scale_embedding": true,
56
+ "static_position_embeddings": true,
57
+ "transformers_version": "4.7.0",
58
+ "use_cache": true,
59
+ "vocab_size": 65001
60
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f8be54f504ebe69bc7d077e6ce5d47220f6f36b51618a2f61920253c031fd2d
3
+ size 310026845
source.spm ADDED
Binary file (835 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 (809 kB). View file
 
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_lang": "spa", "target_lang": "zho", "unk_token": "<unk>", "eos_token": "</s>", "pad_token": "<pad>", "model_max_length": 512, "sp_model_kwargs": {}, "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "Helsinki-NLP/opus-tatoeba-es-zh", "use_fast": false}
vocab.json ADDED
The diff for this file is too large to render. See raw diff