Ethan Sim
commited on
Commit
•
d7da4db
1
Parent(s):
a9a2c0d
stage best big fine unsampled tfidf wce model
Browse files- checkpoint-32000/config.json +46 -0
- checkpoint-32000/generation_config.json +16 -0
- checkpoint-32000/optimizer.pt +3 -0
- checkpoint-32000/pytorch_model.bin +3 -0
- checkpoint-32000/rng_state.pth +3 -0
- checkpoint-32000/scaler.pt +3 -0
- checkpoint-32000/scheduler.pt +3 -0
- checkpoint-32000/source.spm +0 -0
- checkpoint-32000/special_tokens_map.json +5 -0
- checkpoint-32000/target.spm +0 -0
- checkpoint-32000/tokenizer_config.json +12 -0
- checkpoint-32000/trainer_state.json +80 -0
- checkpoint-32000/training_args.bin +3 -0
- checkpoint-32000/vocab.json +0 -0
- config.json +46 -0
- generation_config.json +16 -0
- pytorch_model.bin +3 -0
- source.spm +0 -0
- special_tokens_map.json +5 -0
- target.spm +0 -0
- tokenizer_config.json +12 -0
- training_args.bin +3 -0
- vocab.json +0 -0
checkpoint-32000/config.json
ADDED
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "Helsinki-NLP/opus-mt-tc-big-en-fr",
|
3 |
+
"activation_dropout": 0.0,
|
4 |
+
"activation_function": "relu",
|
5 |
+
"architectures": [
|
6 |
+
"MarianMTModel"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"bad_words_ids": [
|
10 |
+
[
|
11 |
+
53016
|
12 |
+
]
|
13 |
+
],
|
14 |
+
"bos_token_id": 0,
|
15 |
+
"classifier_dropout": 0.0,
|
16 |
+
"d_model": 1024,
|
17 |
+
"decoder_attention_heads": 16,
|
18 |
+
"decoder_ffn_dim": 4096,
|
19 |
+
"decoder_layerdrop": 0.0,
|
20 |
+
"decoder_layers": 6,
|
21 |
+
"decoder_start_token_id": 53016,
|
22 |
+
"decoder_vocab_size": 53017,
|
23 |
+
"dropout": 0.1,
|
24 |
+
"encoder_attention_heads": 16,
|
25 |
+
"encoder_ffn_dim": 4096,
|
26 |
+
"encoder_layerdrop": 0.0,
|
27 |
+
"encoder_layers": 6,
|
28 |
+
"eos_token_id": 43311,
|
29 |
+
"forced_eos_token_id": 43311,
|
30 |
+
"init_std": 0.02,
|
31 |
+
"is_encoder_decoder": true,
|
32 |
+
"max_length": 512,
|
33 |
+
"max_position_embeddings": 1024,
|
34 |
+
"model_type": "marian",
|
35 |
+
"normalize_embedding": false,
|
36 |
+
"num_beams": 4,
|
37 |
+
"num_hidden_layers": 6,
|
38 |
+
"pad_token_id": 53016,
|
39 |
+
"scale_embedding": true,
|
40 |
+
"share_encoder_decoder_embeddings": true,
|
41 |
+
"static_position_embeddings": true,
|
42 |
+
"torch_dtype": "float32",
|
43 |
+
"transformers_version": "4.28.0",
|
44 |
+
"use_cache": true,
|
45 |
+
"vocab_size": 53017
|
46 |
+
}
|
checkpoint-32000/generation_config.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bad_words_ids": [
|
3 |
+
[
|
4 |
+
53016
|
5 |
+
]
|
6 |
+
],
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"decoder_start_token_id": 53016,
|
9 |
+
"eos_token_id": 43311,
|
10 |
+
"forced_eos_token_id": 43311,
|
11 |
+
"max_length": 512,
|
12 |
+
"num_beams": 4,
|
13 |
+
"pad_token_id": 53016,
|
14 |
+
"renormalize_logits": true,
|
15 |
+
"transformers_version": "4.28.0"
|
16 |
+
}
|
checkpoint-32000/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0b7d908550705ba53670f7691c48f9170f990bcf3907a0b1492d5a9d3515d8f0
|
3 |
+
size 1845322757
|
checkpoint-32000/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6fe7f29734646ea64745e1fb9c697ec186d8002ecd41d780c0b05796245f8118
|
3 |
+
size 922885701
|
checkpoint-32000/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:130a4fc57c723d7800c5eac87bf28c38ce67d7d1bdf71c179873cbaabfcea0b1
|
3 |
+
size 14639
|
checkpoint-32000/scaler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:97b776940b0c5587dab32e7e970f8a08b91b7b270a9b905b5c2df91aec615d84
|
3 |
+
size 557
|
checkpoint-32000/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:448b4a0ac8ff84d011d3c0924a8ddf1b053ab7eed882c9fbea385d7a1e97de77
|
3 |
+
size 627
|
checkpoint-32000/source.spm
ADDED
Binary file (802 kB). View file
|
|
checkpoint-32000/special_tokens_map.json
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"eos_token": "</s>",
|
3 |
+
"pad_token": "<pad>",
|
4 |
+
"unk_token": "<unk>"
|
5 |
+
}
|
checkpoint-32000/target.spm
ADDED
Binary file (820 kB). View file
|
|
checkpoint-32000/tokenizer_config.json
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"clean_up_tokenization_spaces": true,
|
3 |
+
"eos_token": "</s>",
|
4 |
+
"model_max_length": 512,
|
5 |
+
"pad_token": "<pad>",
|
6 |
+
"separate_vocabs": false,
|
7 |
+
"source_lang": "en",
|
8 |
+
"sp_model_kwargs": {},
|
9 |
+
"target_lang": "fr",
|
10 |
+
"tokenizer_class": "MarianTokenizer",
|
11 |
+
"unk_token": "<unk>"
|
12 |
+
}
|
checkpoint-32000/trainer_state.json
ADDED
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 42.9908,
|
3 |
+
"best_model_checkpoint": "opus_big_fine_tfidf_wce_unsampled/checkpoint-32000",
|
4 |
+
"epoch": 0.7886629698089956,
|
5 |
+
"global_step": 32000,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.2,
|
12 |
+
"learning_rate": 1.9753635243376465e-05,
|
13 |
+
"loss": 0.2036,
|
14 |
+
"step": 8000
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.2,
|
18 |
+
"eval_bleu": 41.9332,
|
19 |
+
"eval_gen_len": 33.7565,
|
20 |
+
"eval_loss": 0.09845346212387085,
|
21 |
+
"eval_runtime": 98.5797,
|
22 |
+
"eval_samples_per_second": 10.58,
|
23 |
+
"eval_steps_per_second": 0.67,
|
24 |
+
"step": 8000
|
25 |
+
},
|
26 |
+
{
|
27 |
+
"epoch": 0.39,
|
28 |
+
"learning_rate": 1.950723967960567e-05,
|
29 |
+
"loss": 0.1717,
|
30 |
+
"step": 16000
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.39,
|
34 |
+
"eval_bleu": 42.466,
|
35 |
+
"eval_gen_len": 34.3557,
|
36 |
+
"eval_loss": 0.09779991954565048,
|
37 |
+
"eval_runtime": 101.8795,
|
38 |
+
"eval_samples_per_second": 10.238,
|
39 |
+
"eval_steps_per_second": 0.648,
|
40 |
+
"step": 16000
|
41 |
+
},
|
42 |
+
{
|
43 |
+
"epoch": 0.59,
|
44 |
+
"learning_rate": 1.926090573012939e-05,
|
45 |
+
"loss": 0.1668,
|
46 |
+
"step": 24000
|
47 |
+
},
|
48 |
+
{
|
49 |
+
"epoch": 0.59,
|
50 |
+
"eval_bleu": 42.674,
|
51 |
+
"eval_gen_len": 33.7651,
|
52 |
+
"eval_loss": 0.09813275933265686,
|
53 |
+
"eval_runtime": 147.0398,
|
54 |
+
"eval_samples_per_second": 7.093,
|
55 |
+
"eval_steps_per_second": 0.449,
|
56 |
+
"step": 24000
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 0.79,
|
60 |
+
"learning_rate": 1.9014540973505854e-05,
|
61 |
+
"loss": 0.1623,
|
62 |
+
"step": 32000
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 0.79,
|
66 |
+
"eval_bleu": 42.9908,
|
67 |
+
"eval_gen_len": 33.8389,
|
68 |
+
"eval_loss": 0.09677895158529282,
|
69 |
+
"eval_runtime": 81.8819,
|
70 |
+
"eval_samples_per_second": 12.738,
|
71 |
+
"eval_steps_per_second": 0.806,
|
72 |
+
"step": 32000
|
73 |
+
}
|
74 |
+
],
|
75 |
+
"max_steps": 649200,
|
76 |
+
"num_train_epochs": 16,
|
77 |
+
"total_flos": 1.0342418777269862e+17,
|
78 |
+
"trial_name": null,
|
79 |
+
"trial_params": null
|
80 |
+
}
|
checkpoint-32000/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:318f091bf0bbc5a7ace3758915c85825376e2bde0e6c103735ea1e3a5831a158
|
3 |
+
size 3771
|
checkpoint-32000/vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
config.json
ADDED
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "Helsinki-NLP/opus-mt-tc-big-en-fr",
|
3 |
+
"activation_dropout": 0.0,
|
4 |
+
"activation_function": "relu",
|
5 |
+
"architectures": [
|
6 |
+
"MarianMTModel"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"bad_words_ids": [
|
10 |
+
[
|
11 |
+
53016
|
12 |
+
]
|
13 |
+
],
|
14 |
+
"bos_token_id": 0,
|
15 |
+
"classifier_dropout": 0.0,
|
16 |
+
"d_model": 1024,
|
17 |
+
"decoder_attention_heads": 16,
|
18 |
+
"decoder_ffn_dim": 4096,
|
19 |
+
"decoder_layerdrop": 0.0,
|
20 |
+
"decoder_layers": 6,
|
21 |
+
"decoder_start_token_id": 53016,
|
22 |
+
"decoder_vocab_size": 53017,
|
23 |
+
"dropout": 0.1,
|
24 |
+
"encoder_attention_heads": 16,
|
25 |
+
"encoder_ffn_dim": 4096,
|
26 |
+
"encoder_layerdrop": 0.0,
|
27 |
+
"encoder_layers": 6,
|
28 |
+
"eos_token_id": 43311,
|
29 |
+
"forced_eos_token_id": 43311,
|
30 |
+
"init_std": 0.02,
|
31 |
+
"is_encoder_decoder": true,
|
32 |
+
"max_length": 512,
|
33 |
+
"max_position_embeddings": 1024,
|
34 |
+
"model_type": "marian",
|
35 |
+
"normalize_embedding": false,
|
36 |
+
"num_beams": 4,
|
37 |
+
"num_hidden_layers": 6,
|
38 |
+
"pad_token_id": 53016,
|
39 |
+
"scale_embedding": true,
|
40 |
+
"share_encoder_decoder_embeddings": true,
|
41 |
+
"static_position_embeddings": true,
|
42 |
+
"torch_dtype": "float32",
|
43 |
+
"transformers_version": "4.28.0",
|
44 |
+
"use_cache": true,
|
45 |
+
"vocab_size": 53017
|
46 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bad_words_ids": [
|
3 |
+
[
|
4 |
+
53016
|
5 |
+
]
|
6 |
+
],
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"decoder_start_token_id": 53016,
|
9 |
+
"eos_token_id": 43311,
|
10 |
+
"forced_eos_token_id": 43311,
|
11 |
+
"max_length": 512,
|
12 |
+
"num_beams": 4,
|
13 |
+
"pad_token_id": 53016,
|
14 |
+
"renormalize_logits": true,
|
15 |
+
"transformers_version": "4.28.0"
|
16 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6fe7f29734646ea64745e1fb9c697ec186d8002ecd41d780c0b05796245f8118
|
3 |
+
size 922885701
|
source.spm
ADDED
Binary file (802 kB). View file
|
|
special_tokens_map.json
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"eos_token": "</s>",
|
3 |
+
"pad_token": "<pad>",
|
4 |
+
"unk_token": "<unk>"
|
5 |
+
}
|
target.spm
ADDED
Binary file (820 kB). View file
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"clean_up_tokenization_spaces": true,
|
3 |
+
"eos_token": "</s>",
|
4 |
+
"model_max_length": 512,
|
5 |
+
"pad_token": "<pad>",
|
6 |
+
"separate_vocabs": false,
|
7 |
+
"source_lang": "en",
|
8 |
+
"sp_model_kwargs": {},
|
9 |
+
"target_lang": "fr",
|
10 |
+
"tokenizer_class": "MarianTokenizer",
|
11 |
+
"unk_token": "<unk>"
|
12 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:318f091bf0bbc5a7ace3758915c85825376e2bde0e6c103735ea1e3a5831a158
|
3 |
+
size 3771
|
vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|