Hugging Face's logo

BioMedTok
/
SentencePieceBPE-PubMed-FR-Morphemes

Fill-Mask
Transformers
PyTorch
TensorBoard
camembert
Inference Endpoints
Model card Files Metrics Community
qanastek commited on
Commit
1a0b1a9
•
1 Parent(s): d2b571f

Upload 43 files

Browse files
Files changed (43) hide show
  1. config.json +26 -0
  2. optimizer.pt +3 -0
  3. pytorch_model.bin +3 -0
  4. rng_state_0.pth +3 -0
  5. rng_state_1.pth +3 -0
  6. rng_state_10.pth +3 -0
  7. rng_state_11.pth +3 -0
  8. rng_state_12.pth +3 -0
  9. rng_state_13.pth +3 -0
  10. rng_state_14.pth +3 -0
  11. rng_state_15.pth +3 -0
  12. rng_state_16.pth +3 -0
  13. rng_state_17.pth +3 -0
  14. rng_state_18.pth +3 -0
  15. rng_state_19.pth +3 -0
  16. rng_state_2.pth +3 -0
  17. rng_state_20.pth +3 -0
  18. rng_state_21.pth +3 -0
  19. rng_state_22.pth +3 -0
  20. rng_state_23.pth +3 -0
  21. rng_state_24.pth +3 -0
  22. rng_state_25.pth +3 -0
  23. rng_state_26.pth +3 -0
  24. rng_state_27.pth +3 -0
  25. rng_state_28.pth +3 -0
  26. rng_state_29.pth +3 -0
  27. rng_state_3.pth +3 -0
  28. rng_state_30.pth +3 -0
  29. rng_state_31.pth +3 -0
  30. rng_state_4.pth +3 -0
  31. rng_state_5.pth +3 -0
  32. rng_state_6.pth +3 -0
  33. rng_state_7.pth +3 -0
  34. rng_state_8.pth +3 -0
  35. rng_state_9.pth +3 -0
  36. scaler.pt +3 -0
  37. scheduler.pt +3 -0
  38. sentencepiece.bpe.model +3 -0
  39. special_tokens_map.json +19 -0
  40. tokenizer.json +0 -0
  41. tokenizer_config.json +27 -0
  42. trainer_state.json +1060 -0
  43. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "CamembertForMaskedLM"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 5,
7
+ "classifier_dropout": null,
8
+ "eos_token_id": 6,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 3072,
14
+ "layer_norm_eps": 1e-12,
15
+ "max_position_embeddings": 514,
16
+ "model_type": "camembert",
17
+ "num_attention_heads": 12,
18
+ "num_hidden_layers": 12,
19
+ "pad_token_id": 1,
20
+ "position_embedding_type": "absolute",
21
+ "torch_dtype": "float32",
22
+ "transformers_version": "4.22.2",
23
+ "type_vocab_size": 1,
24
+ "use_cache": true,
25
+ "vocab_size": 32005
26
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0b3b5b7b394a6486fefb3f82730d8ea35b70ce4a0f7795180ddbe19bd93f909
3
+ size 885361433
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c626eb07cbfb74bd09c86a1aaab4e4f8a9728188fc2eaca9b87e9ce4b35a2c04
3
+ size 442693931
rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_10.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_11.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_12.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_13.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_14.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_15.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_16.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_17.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_18.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_19.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_20.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_21.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_22.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_23.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_24.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_25.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_26.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_27.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_28.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_29.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_30.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_31.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_8.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
rng_state_9.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:484ec9dc72d04dabb7b8de7e6a70b6160ac782115a43d2b99de1e0fa6a50be92
3
+ size 14503
scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:517b8e9913beb30b2ae33d6974fd6fee1e6f87a69ee2028115af6866d0917bf4
3
+ size 559
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf1b14dc67ef2bc1bee53bdacbb2399f7d41740a2186d6a858d661b8bd3dfcfc
3
+ size 623
sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:396ac84a865b2fb297be51d0623dd7e30ef41231edd5b83a8419c0a271ac31ae
3
+ size 794706
special_tokens_map.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<s>NOTUSED",
4
+ "</s>NOTUSED"
5
+ ],
6
+ "bos_token": "<s>",
7
+ "cls_token": "<s>",
8
+ "eos_token": "</s>",
9
+ "mask_token": {
10
+ "content": "<mask>",
11
+ "lstrip": true,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<pad>",
17
+ "sep_token": "</s>",
18
+ "unk_token": "<unk>"
19
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<s>NOTUSED",
4
+ "</s>NOTUSED"
5
+ ],
6
+ "bos_token": "<s>",
7
+ "clean_up_tokenization_spaces": true,
8
+ "cls_token": "<s>",
9
+ "eos_token": "</s>",
10
+ "mask_token": {
11
+ "__type": "AddedToken",
12
+ "content": "<mask>",
13
+ "lstrip": true,
14
+ "normalized": true,
15
+ "rstrip": false,
16
+ "single_word": false
17
+ },
18
+ "max_len": 512,
19
+ "model_max_length": 512,
20
+ "name_or_path": "./sentencepiece_tokenizers_morphemes_v4/SentencePieceTokenizerMorphemesExcluded_PubMed_Abstracts_CHARS_lowercased_fixed_utf8_V4/",
21
+ "pad_token": "<pad>",
22
+ "sep_token": "</s>",
23
+ "sp_model_kwargs": {},
24
+ "special_tokens_map_file": "./sentencepiece_tokenizers_morphemes_v4/SentencePieceTokenizerMorphemesExcluded_PubMed_Abstracts_CHARS_lowercased_fixed_utf8_V4/special_tokens_map.json",
25
+ "tokenizer_class": "CamembertTokenizer",
26
+ "unk_token": "<unk>"
27
+ }
trainer_state.json ADDED
@@ -0,0 +1,1060 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 17.0,
5
+ "global_step": 86615,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 5e-09,
13
+ "loss": 10.525,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.1,
18
+ "learning_rate": 2.5e-06,
19
+ "loss": 9.5219,
20
+ "step": 500
21
+ },
22
+ {
23
+ "epoch": 0.2,
24
+ "learning_rate": 5e-06,
25
+ "loss": 7.6948,
26
+ "step": 1000
27
+ },
28
+ {
29
+ "epoch": 0.29,
30
+ "learning_rate": 7.5e-06,
31
+ "loss": 6.4159,
32
+ "step": 1500
33
+ },
34
+ {
35
+ "epoch": 0.39,
36
+ "learning_rate": 1e-05,
37
+ "loss": 6.1298,
38
+ "step": 2000
39
+ },
40
+ {
41
+ "epoch": 0.49,
42
+ "learning_rate": 1.25e-05,
43
+ "loss": 6.0017,
44
+ "step": 2500
45
+ },
46
+ {
47
+ "epoch": 0.59,
48
+ "learning_rate": 1.5e-05,
49
+ "loss": 5.9201,
50
+ "step": 3000
51
+ },
52
+ {
53
+ "epoch": 0.69,
54
+ "learning_rate": 1.75e-05,
55
+ "loss": 5.8528,
56
+ "step": 3500
57
+ },
58
+ {
59
+ "epoch": 0.79,
60
+ "learning_rate": 2e-05,
61
+ "loss": 5.8017,
62
+ "step": 4000
63
+ },
64
+ {
65
+ "epoch": 0.88,
66
+ "learning_rate": 2.25e-05,
67
+ "loss": 5.7573,
68
+ "step": 4500
69
+ },
70
+ {
71
+ "epoch": 0.98,
72
+ "learning_rate": 2.5e-05,
73
+ "loss": 5.7162,
74
+ "step": 5000
75
+ },
76
+ {
77
+ "epoch": 1.08,
78
+ "learning_rate": 2.7495000000000004e-05,
79
+ "loss": 5.6827,
80
+ "step": 5500
81
+ },
82
+ {
83
+ "epoch": 1.18,
84
+ "learning_rate": 2.9995e-05,
85
+ "loss": 5.6557,
86
+ "step": 6000
87
+ },
88
+ {
89
+ "epoch": 1.28,
90
+ "learning_rate": 3.2495000000000007e-05,
91
+ "loss": 5.6229,
92
+ "step": 6500
93
+ },
94
+ {
95
+ "epoch": 1.37,
96
+ "learning_rate": 3.4995e-05,
97
+ "loss": 5.599,
98
+ "step": 7000
99
+ },
100
+ {
101
+ "epoch": 1.47,
102
+ "learning_rate": 3.7495e-05,
103
+ "loss": 5.5785,
104
+ "step": 7500
105
+ },
106
+ {
107
+ "epoch": 1.57,
108
+ "learning_rate": 3.9995000000000006e-05,
109
+ "loss": 5.5579,
110
+ "step": 8000
111
+ },
112
+ {
113
+ "epoch": 1.67,
114
+ "learning_rate": 4.2495e-05,
115
+ "loss": 5.5374,
116
+ "step": 8500
117
+ },
118
+ {
119
+ "epoch": 1.77,
120
+ "learning_rate": 4.4995000000000005e-05,
121
+ "loss": 5.5223,
122
+ "step": 9000
123
+ },
124
+ {
125
+ "epoch": 1.86,
126
+ "learning_rate": 4.7495e-05,
127
+ "loss": 5.5075,
128
+ "step": 9500
129
+ },
130
+ {
131
+ "epoch": 1.96,
132
+ "learning_rate": 4.9995000000000005e-05,
133
+ "loss": 5.4966,
134
+ "step": 10000
135
+ },
136
+ {
137
+ "epoch": 2.06,
138
+ "learning_rate": 4.9987721893491125e-05,
139
+ "loss": 5.4806,
140
+ "step": 10500
141
+ },
142
+ {
143
+ "epoch": 2.16,
144
+ "learning_rate": 4.997539447731755e-05,
145
+ "loss": 5.4673,
146
+ "step": 11000
147
+ },
148
+ {
149
+ "epoch": 2.26,
150
+ "learning_rate": 4.996306706114399e-05,
151
+ "loss": 5.4577,
152
+ "step": 11500
153
+ },
154
+ {
155
+ "epoch": 2.36,
156
+ "learning_rate": 4.9950739644970415e-05,
157
+ "loss": 5.4477,
158
+ "step": 12000
159
+ },
160
+ {
161
+ "epoch": 2.45,
162
+ "learning_rate": 4.993841222879685e-05,
163
+ "loss": 5.4356,
164
+ "step": 12500
165
+ },
166
+ {
167
+ "epoch": 2.55,
168
+ "learning_rate": 4.9926084812623276e-05,
169
+ "loss": 5.4225,
170
+ "step": 13000
171
+ },
172
+ {
173
+ "epoch": 2.65,
174
+ "learning_rate": 4.9913757396449704e-05,
175
+ "loss": 5.4196,
176
+ "step": 13500
177
+ },
178
+ {
179
+ "epoch": 2.75,
180
+ "learning_rate": 4.990142998027614e-05,
181
+ "loss": 5.4087,
182
+ "step": 14000
183
+ },
184
+ {
185
+ "epoch": 2.85,
186
+ "learning_rate": 4.9889127218934914e-05,
187
+ "loss": 5.4021,
188
+ "step": 14500
189
+ },
190
+ {
191
+ "epoch": 2.94,
192
+ "learning_rate": 4.987679980276134e-05,
193
+ "loss": 5.3944,
194
+ "step": 15000
195
+ },
196
+ {
197
+ "epoch": 3.04,
198
+ "learning_rate": 4.986449704142012e-05,
199
+ "loss": 5.3828,
200
+ "step": 15500
201
+ },
202
+ {
203
+ "epoch": 3.14,
204
+ "learning_rate": 4.985216962524655e-05,
205
+ "loss": 5.3786,
206
+ "step": 16000
207
+ },
208
+ {
209
+ "epoch": 3.24,
210
+ "learning_rate": 4.983984220907298e-05,
211
+ "loss": 5.3733,
212
+ "step": 16500
213
+ },
214
+ {
215
+ "epoch": 3.34,
216
+ "learning_rate": 4.982751479289941e-05,
217
+ "loss": 5.3666,
218
+ "step": 17000
219
+ },
220
+ {
221
+ "epoch": 3.43,
222
+ "learning_rate": 4.981518737672584e-05,
223
+ "loss": 5.3626,
224
+ "step": 17500
225
+ },
226
+ {
227
+ "epoch": 3.53,
228
+ "learning_rate": 4.9802859960552274e-05,
229
+ "loss": 5.353,
230
+ "step": 18000
231
+ },
232
+ {
233
+ "epoch": 3.63,
234
+ "learning_rate": 4.97905325443787e-05,
235
+ "loss": 5.3496,
236
+ "step": 18500
237
+ },
238
+ {
239
+ "epoch": 3.73,
240
+ "learning_rate": 4.977820512820513e-05,
241
+ "loss": 5.3454,
242
+ "step": 19000
243
+ },
244
+ {
245
+ "epoch": 3.83,
246
+ "learning_rate": 4.976590236686391e-05,
247
+ "loss": 5.3396,
248
+ "step": 19500
249
+ },
250
+ {
251
+ "epoch": 3.93,
252
+ "learning_rate": 4.975359960552268e-05,
253
+ "loss": 5.3352,
254
+ "step": 20000
255
+ },
256
+ {
257
+ "epoch": 4.02,
258
+ "learning_rate": 4.974129684418146e-05,
259
+ "loss": 5.3312,
260
+ "step": 20500
261
+ },
262
+ {
263
+ "epoch": 4.12,
264
+ "learning_rate": 4.9728969428007896e-05,
265
+ "loss": 5.3245,
266
+ "step": 21000
267
+ },
268
+ {
269
+ "epoch": 4.22,
270
+ "learning_rate": 4.9716642011834323e-05,
271
+ "loss": 5.3126,
272
+ "step": 21500
273
+ },
274
+ {
275
+ "epoch": 4.32,
276
+ "learning_rate": 4.970431459566075e-05,
277
+ "loss": 5.0663,
278
+ "step": 22000
279
+ },
280
+ {
281
+ "epoch": 4.42,
282
+ "learning_rate": 4.969198717948718e-05,
283
+ "loss": 4.7802,
284
+ "step": 22500
285
+ },
286
+ {
287
+ "epoch": 4.51,
288
+ "learning_rate": 4.967965976331361e-05,
289
+ "loss": 4.5089,
290
+ "step": 23000
291
+ },
292
+ {
293
+ "epoch": 4.61,
294
+ "learning_rate": 4.966733234714005e-05,
295
+ "loss": 4.2703,
296
+ "step": 23500
297
+ },
298
+ {
299
+ "epoch": 4.71,
300
+ "learning_rate": 4.965500493096647e-05,
301
+ "loss": 4.0607,
302
+ "step": 24000
303
+ },
304
+ {
305
+ "epoch": 4.81,
306
+ "learning_rate": 4.96426775147929e-05,
307
+ "loss": 3.8757,
308
+ "step": 24500
309
+ },
310
+ {
311
+ "epoch": 4.91,
312
+ "learning_rate": 4.963037475345168e-05,
313
+ "loss": 3.5749,
314
+ "step": 25000
315
+ },
316
+ {
317
+ "epoch": 5.0,
318
+ "learning_rate": 4.961807199211045e-05,
319
+ "loss": 3.0471,
320
+ "step": 25500
321
+ },
322
+ {
323
+ "epoch": 5.1,
324
+ "learning_rate": 4.960574457593689e-05,
325
+ "loss": 2.5977,
326
+ "step": 26000
327
+ },
328
+ {
329
+ "epoch": 5.2,
330
+ "learning_rate": 4.9593417159763314e-05,
331
+ "loss": 2.1306,
332
+ "step": 26500
333
+ },
334
+ {
335
+ "epoch": 5.3,
336
+ "learning_rate": 4.958108974358975e-05,
337
+ "loss": 1.9047,
338
+ "step": 27000
339
+ },
340
+ {
341
+ "epoch": 5.4,
342
+ "learning_rate": 4.9568762327416176e-05,
343
+ "loss": 1.7619,
344
+ "step": 27500
345
+ },
346
+ {
347
+ "epoch": 5.5,
348
+ "learning_rate": 4.9556434911242603e-05,
349
+ "loss": 1.6574,
350
+ "step": 28000
351
+ },
352
+ {
353
+ "epoch": 5.59,
354
+ "learning_rate": 4.954410749506904e-05,
355
+ "loss": 1.5785,
356
+ "step": 28500
357
+ },
358
+ {
359
+ "epoch": 5.69,
360
+ "learning_rate": 4.9531780078895465e-05,
361
+ "loss": 1.5168,
362
+ "step": 29000
363
+ },
364
+ {
365
+ "epoch": 5.79,
366
+ "learning_rate": 4.951945266272189e-05,
367
+ "loss": 1.467,
368
+ "step": 29500
369
+ },
370
+ {
371
+ "epoch": 5.89,
372
+ "learning_rate": 4.950712524654833e-05,
373
+ "loss": 1.4234,
374
+ "step": 30000
375
+ },
376
+ {
377
+ "epoch": 5.99,
378
+ "learning_rate": 4.94948224852071e-05,
379
+ "loss": 1.3853,
380
+ "step": 30500
381
+ },
382
+ {
383
+ "epoch": 6.08,
384
+ "learning_rate": 4.948251972386588e-05,
385
+ "loss": 1.3438,
386
+ "step": 31000
387
+ },
388
+ {
389
+ "epoch": 6.18,
390
+ "learning_rate": 4.947019230769231e-05,
391
+ "loss": 1.2958,
392
+ "step": 31500
393
+ },
394
+ {
395
+ "epoch": 6.28,
396
+ "learning_rate": 4.945786489151874e-05,
397
+ "loss": 1.241,
398
+ "step": 32000
399
+ },
400
+ {
401
+ "epoch": 6.38,
402
+ "learning_rate": 4.9445537475345174e-05,
403
+ "loss": 1.1933,
404
+ "step": 32500
405
+ },
406
+ {
407
+ "epoch": 6.48,
408
+ "learning_rate": 4.9433210059171594e-05,
409
+ "loss": 1.1561,
410
+ "step": 33000
411
+ },
412
+ {
413
+ "epoch": 6.58,
414
+ "learning_rate": 4.942088264299803e-05,
415
+ "loss": 1.1211,
416
+ "step": 33500
417
+ },
418
+ {
419
+ "epoch": 6.67,
420
+ "learning_rate": 4.940855522682446e-05,
421
+ "loss": 1.097,
422
+ "step": 34000
423
+ },
424
+ {
425
+ "epoch": 6.77,
426
+ "learning_rate": 4.939622781065089e-05,
427
+ "loss": 1.0713,
428
+ "step": 34500
429
+ },
430
+ {
431
+ "epoch": 6.87,
432
+ "learning_rate": 4.9383925049309666e-05,
433
+ "loss": 1.0513,
434
+ "step": 35000
435
+ },
436
+ {
437
+ "epoch": 6.97,
438
+ "learning_rate": 4.937159763313609e-05,
439
+ "loss": 1.0334,
440
+ "step": 35500
441
+ },
442
+ {
443
+ "epoch": 7.07,
444
+ "learning_rate": 4.9359294871794875e-05,
445
+ "loss": 1.012,
446
+ "step": 36000
447
+ },
448
+ {
449
+ "epoch": 7.16,
450
+ "learning_rate": 4.934696745562131e-05,
451
+ "loss": 0.9953,
452
+ "step": 36500
453
+ },
454
+ {
455
+ "epoch": 7.26,
456
+ "learning_rate": 4.933464003944773e-05,
457
+ "loss": 0.9811,
458
+ "step": 37000
459
+ },
460
+ {
461
+ "epoch": 7.36,
462
+ "learning_rate": 4.9322312623274164e-05,
463
+ "loss": 0.9679,
464
+ "step": 37500
465
+ },
466
+ {
467
+ "epoch": 7.46,
468
+ "learning_rate": 4.930998520710059e-05,
469
+ "loss": 0.9546,
470
+ "step": 38000
471
+ },
472
+ {
473
+ "epoch": 7.56,
474
+ "learning_rate": 4.9297657790927026e-05,
475
+ "loss": 0.9435,
476
+ "step": 38500
477
+ },
478
+ {
479
+ "epoch": 7.65,
480
+ "learning_rate": 4.9285330374753454e-05,
481
+ "loss": 0.9284,
482
+ "step": 39000
483
+ },
484
+ {
485
+ "epoch": 7.75,
486
+ "learning_rate": 4.927300295857988e-05,
487
+ "loss": 0.9222,
488
+ "step": 39500
489
+ },
490
+ {
491
+ "epoch": 7.85,
492
+ "learning_rate": 4.926070019723866e-05,
493
+ "loss": 0.913,
494
+ "step": 40000
495
+ },
496
+ {
497
+ "epoch": 7.95,
498
+ "learning_rate": 4.924837278106509e-05,
499
+ "loss": 0.9039,
500
+ "step": 40500
501
+ },
502
+ {
503
+ "epoch": 8.05,
504
+ "learning_rate": 4.9236070019723866e-05,
505
+ "loss": 0.8952,
506
+ "step": 41000
507
+ },
508
+ {
509
+ "epoch": 8.15,
510
+ "learning_rate": 4.92237426035503e-05,
511
+ "loss": 0.8863,
512
+ "step": 41500
513
+ },
514
+ {
515
+ "epoch": 8.24,
516
+ "learning_rate": 4.921141518737673e-05,
517
+ "loss": 0.8774,
518
+ "step": 42000
519
+ },
520
+ {
521
+ "epoch": 8.34,
522
+ "learning_rate": 4.9199087771203155e-05,
523
+ "loss": 0.8715,
524
+ "step": 42500
525
+ },
526
+ {
527
+ "epoch": 8.44,
528
+ "learning_rate": 4.918676035502959e-05,
529
+ "loss": 0.8636,
530
+ "step": 43000
531
+ },
532
+ {
533
+ "epoch": 8.54,
534
+ "learning_rate": 4.917443293885602e-05,
535
+ "loss": 0.8558,
536
+ "step": 43500
537
+ },
538
+ {
539
+ "epoch": 8.64,
540
+ "learning_rate": 4.916210552268245e-05,
541
+ "loss": 0.8472,
542
+ "step": 44000
543
+ },
544
+ {
545
+ "epoch": 8.73,
546
+ "learning_rate": 4.914977810650888e-05,
547
+ "loss": 0.8439,
548
+ "step": 44500
549
+ },
550
+ {
551
+ "epoch": 8.83,
552
+ "learning_rate": 4.9137475345167654e-05,
553
+ "loss": 0.8341,
554
+ "step": 45000
555
+ },
556
+ {
557
+ "epoch": 8.93,
558
+ "learning_rate": 4.912514792899409e-05,
559
+ "loss": 0.8298,
560
+ "step": 45500
561
+ },
562
+ {
563
+ "epoch": 9.03,
564
+ "learning_rate": 4.9112820512820516e-05,
565
+ "loss": 0.8235,
566
+ "step": 46000
567
+ },
568
+ {
569
+ "epoch": 9.13,
570
+ "learning_rate": 4.910049309664694e-05,
571
+ "loss": 0.8192,
572
+ "step": 46500
573
+ },
574
+ {
575
+ "epoch": 9.22,
576
+ "learning_rate": 4.908816568047338e-05,
577
+ "loss": 0.8103,
578
+ "step": 47000
579
+ },
580
+ {
581
+ "epoch": 9.32,
582
+ "learning_rate": 4.907586291913215e-05,
583
+ "loss": 0.8077,
584
+ "step": 47500
585
+ },
586
+ {
587
+ "epoch": 9.42,
588
+ "learning_rate": 4.906353550295858e-05,
589
+ "loss": 0.8009,
590
+ "step": 48000
591
+ },
592
+ {
593
+ "epoch": 9.52,
594
+ "learning_rate": 4.905120808678501e-05,
595
+ "loss": 0.7984,
596
+ "step": 48500
597
+ },
598
+ {
599
+ "epoch": 9.62,
600
+ "learning_rate": 4.903888067061144e-05,
601
+ "loss": 0.7924,
602
+ "step": 49000
603
+ },
604
+ {
605
+ "epoch": 9.72,
606
+ "learning_rate": 4.9026553254437876e-05,
607
+ "loss": 0.7877,
608
+ "step": 49500
609
+ },
610
+ {
611
+ "epoch": 9.81,
612
+ "learning_rate": 4.9014225838264304e-05,
613
+ "loss": 0.7824,
614
+ "step": 50000
615
+ },
616
+ {
617
+ "epoch": 9.91,
618
+ "learning_rate": 4.900192307692308e-05,
619
+ "loss": 0.7794,
620
+ "step": 50500
621
+ },
622
+ {
623
+ "epoch": 10.01,
624
+ "learning_rate": 4.898962031558186e-05,
625
+ "loss": 0.7749,
626
+ "step": 51000
627
+ },
628
+ {
629
+ "epoch": 10.11,
630
+ "learning_rate": 4.897729289940829e-05,
631
+ "loss": 0.7697,
632
+ "step": 51500
633
+ },
634
+ {
635
+ "epoch": 10.21,
636
+ "learning_rate": 4.8964965483234716e-05,
637
+ "loss": 0.7651,
638
+ "step": 52000
639
+ },
640
+ {
641
+ "epoch": 10.3,
642
+ "learning_rate": 4.8952638067061144e-05,
643
+ "loss": 0.7603,
644
+ "step": 52500
645
+ },
646
+ {
647
+ "epoch": 10.4,
648
+ "learning_rate": 4.894031065088758e-05,
649
+ "loss": 0.7583,
650
+ "step": 53000
651
+ },
652
+ {
653
+ "epoch": 10.5,
654
+ "learning_rate": 4.892798323471401e-05,
655
+ "loss": 0.755,
656
+ "step": 53500
657
+ },
658
+ {
659
+ "epoch": 10.6,
660
+ "learning_rate": 4.891565581854043e-05,
661
+ "loss": 0.7497,
662
+ "step": 54000
663
+ },
664
+ {
665
+ "epoch": 10.7,
666
+ "learning_rate": 4.890332840236687e-05,
667
+ "loss": 0.7457,
668
+ "step": 54500
669
+ },
670
+ {
671
+ "epoch": 10.79,
672
+ "learning_rate": 4.8891000986193295e-05,
673
+ "loss": 0.7429,
674
+ "step": 55000
675
+ },
676
+ {
677
+ "epoch": 10.89,
678
+ "learning_rate": 4.887869822485207e-05,
679
+ "loss": 0.7381,
680
+ "step": 55500
681
+ },
682
+ {
683
+ "epoch": 10.99,
684
+ "learning_rate": 4.8866370808678504e-05,
685
+ "loss": 0.7369,
686
+ "step": 56000
687
+ },
688
+ {
689
+ "epoch": 11.09,
690
+ "learning_rate": 4.885406804733728e-05,
691
+ "loss": 0.7321,
692
+ "step": 56500
693
+ },
694
+ {
695
+ "epoch": 11.19,
696
+ "learning_rate": 4.8841740631163714e-05,
697
+ "loss": 0.7293,
698
+ "step": 57000
699
+ },
700
+ {
701
+ "epoch": 11.29,
702
+ "learning_rate": 4.882941321499014e-05,
703
+ "loss": 0.7258,
704
+ "step": 57500
705
+ },
706
+ {
707
+ "epoch": 11.38,
708
+ "learning_rate": 4.881708579881657e-05,
709
+ "loss": 0.7243,
710
+ "step": 58000
711
+ },
712
+ {
713
+ "epoch": 11.48,
714
+ "learning_rate": 4.8804758382643e-05,
715
+ "loss": 0.7207,
716
+ "step": 58500
717
+ },
718
+ {
719
+ "epoch": 11.58,
720
+ "learning_rate": 4.879243096646943e-05,
721
+ "loss": 0.716,
722
+ "step": 59000
723
+ },
724
+ {
725
+ "epoch": 11.68,
726
+ "learning_rate": 4.878010355029586e-05,
727
+ "loss": 0.7123,
728
+ "step": 59500
729
+ },
730
+ {
731
+ "epoch": 11.78,
732
+ "learning_rate": 4.876777613412229e-05,
733
+ "loss": 0.7106,
734
+ "step": 60000
735
+ },
736
+ {
737
+ "epoch": 11.87,
738
+ "learning_rate": 4.875544871794872e-05,
739
+ "loss": 0.7086,
740
+ "step": 60500
741
+ },
742
+ {
743
+ "epoch": 11.97,
744
+ "learning_rate": 4.8743121301775154e-05,
745
+ "loss": 0.7058,
746
+ "step": 61000
747
+ },
748
+ {
749
+ "epoch": 12.07,
750
+ "learning_rate": 4.873084319526628e-05,
751
+ "loss": 0.7038,
752
+ "step": 61500
753
+ },
754
+ {
755
+ "epoch": 12.17,
756
+ "learning_rate": 4.8718515779092704e-05,
757
+ "loss": 0.7001,
758
+ "step": 62000
759
+ },
760
+ {
761
+ "epoch": 12.27,
762
+ "learning_rate": 4.870618836291914e-05,
763
+ "loss": 0.6989,
764
+ "step": 62500
765
+ },
766
+ {
767
+ "epoch": 12.37,
768
+ "learning_rate": 4.869386094674556e-05,
769
+ "loss": 0.6932,
770
+ "step": 63000
771
+ },
772
+ {
773
+ "epoch": 12.46,
774
+ "learning_rate": 4.8681533530571994e-05,
775
+ "loss": 0.6922,
776
+ "step": 63500
777
+ },
778
+ {
779
+ "epoch": 12.56,
780
+ "learning_rate": 4.866920611439843e-05,
781
+ "loss": 0.6894,
782
+ "step": 64000
783
+ },
784
+ {
785
+ "epoch": 12.66,
786
+ "learning_rate": 4.8656878698224855e-05,
787
+ "loss": 0.6862,
788
+ "step": 64500
789
+ },
790
+ {
791
+ "epoch": 12.76,
792
+ "learning_rate": 4.864455128205128e-05,
793
+ "loss": 0.6846,
794
+ "step": 65000
795
+ },
796
+ {
797
+ "epoch": 12.86,
798
+ "learning_rate": 4.863224852071006e-05,
799
+ "loss": 0.6832,
800
+ "step": 65500
801
+ },
802
+ {
803
+ "epoch": 12.95,
804
+ "learning_rate": 4.861994575936884e-05,
805
+ "loss": 0.6807,
806
+ "step": 66000
807
+ },
808
+ {
809
+ "epoch": 13.05,
810
+ "learning_rate": 4.8607642998027616e-05,
811
+ "loss": 0.678,
812
+ "step": 66500
813
+ },
814
+ {
815
+ "epoch": 13.15,
816
+ "learning_rate": 4.859531558185404e-05,
817
+ "loss": 0.6761,
818
+ "step": 67000
819
+ },
820
+ {
821
+ "epoch": 13.25,
822
+ "learning_rate": 4.858298816568048e-05,
823
+ "loss": 0.6738,
824
+ "step": 67500
825
+ },
826
+ {
827
+ "epoch": 13.35,
828
+ "learning_rate": 4.8570660749506905e-05,
829
+ "loss": 0.6713,
830
+ "step": 68000
831
+ },
832
+ {
833
+ "epoch": 13.44,
834
+ "learning_rate": 4.855833333333333e-05,
835
+ "loss": 0.6697,
836
+ "step": 68500
837
+ },
838
+ {
839
+ "epoch": 13.54,
840
+ "learning_rate": 4.8546005917159767e-05,
841
+ "loss": 0.6667,
842
+ "step": 69000
843
+ },
844
+ {
845
+ "epoch": 13.64,
846
+ "learning_rate": 4.8533678500986194e-05,
847
+ "loss": 0.6653,
848
+ "step": 69500
849
+ },
850
+ {
851
+ "epoch": 13.74,
852
+ "learning_rate": 4.852135108481263e-05,
853
+ "loss": 0.6634,
854
+ "step": 70000
855
+ },
856
+ {
857
+ "epoch": 13.84,
858
+ "learning_rate": 4.8509023668639056e-05,
859
+ "loss": 0.6603,
860
+ "step": 70500
861
+ },
862
+ {
863
+ "epoch": 13.94,
864
+ "learning_rate": 4.849669625246548e-05,
865
+ "loss": 0.6591,
866
+ "step": 71000
867
+ },
868
+ {
869
+ "epoch": 14.03,
870
+ "learning_rate": 4.8484393491124265e-05,
871
+ "loss": 0.6573,
872
+ "step": 71500
873
+ },
874
+ {
875
+ "epoch": 14.13,
876
+ "learning_rate": 4.847206607495069e-05,
877
+ "loss": 0.6543,
878
+ "step": 72000
879
+ },
880
+ {
881
+ "epoch": 14.23,
882
+ "learning_rate": 4.845973865877712e-05,
883
+ "loss": 0.6525,
884
+ "step": 72500
885
+ },
886
+ {
887
+ "epoch": 14.33,
888
+ "learning_rate": 4.8447411242603555e-05,
889
+ "loss": 0.6518,
890
+ "step": 73000
891
+ },
892
+ {
893
+ "epoch": 14.43,
894
+ "learning_rate": 4.843508382642998e-05,
895
+ "loss": 0.6494,
896
+ "step": 73500
897
+ },
898
+ {
899
+ "epoch": 14.52,
900
+ "learning_rate": 4.8422756410256416e-05,
901
+ "loss": 0.6478,
902
+ "step": 74000
903
+ },
904
+ {
905
+ "epoch": 14.62,
906
+ "learning_rate": 4.841042899408284e-05,
907
+ "loss": 0.646,
908
+ "step": 74500
909
+ },
910
+ {
911
+ "epoch": 14.72,
912
+ "learning_rate": 4.839810157790927e-05,
913
+ "loss": 0.6438,
914
+ "step": 75000
915
+ },
916
+ {
917
+ "epoch": 14.82,
918
+ "learning_rate": 4.8385774161735706e-05,
919
+ "loss": 0.6421,
920
+ "step": 75500
921
+ },
922
+ {
923
+ "epoch": 14.92,
924
+ "learning_rate": 4.837349605522682e-05,
925
+ "loss": 0.6417,
926
+ "step": 76000
927
+ },
928
+ {
929
+ "epoch": 15.01,
930
+ "learning_rate": 4.8361168639053256e-05,
931
+ "loss": 0.6416,
932
+ "step": 76500
933
+ },
934
+ {
935
+ "epoch": 15.11,
936
+ "learning_rate": 4.834884122287969e-05,
937
+ "loss": 0.637,
938
+ "step": 77000
939
+ },
940
+ {
941
+ "epoch": 15.21,
942
+ "learning_rate": 4.833651380670612e-05,
943
+ "loss": 0.6362,
944
+ "step": 77500
945
+ },
946
+ {
947
+ "epoch": 15.31,
948
+ "learning_rate": 4.8324186390532545e-05,
949
+ "loss": 0.6355,
950
+ "step": 78000
951
+ },
952
+ {
953
+ "epoch": 15.41,
954
+ "learning_rate": 4.831185897435897e-05,
955
+ "loss": 0.6336,
956
+ "step": 78500
957
+ },
958
+ {
959
+ "epoch": 15.51,
960
+ "learning_rate": 4.829953155818541e-05,
961
+ "loss": 0.6304,
962
+ "step": 79000
963
+ },
964
+ {
965
+ "epoch": 15.6,
966
+ "learning_rate": 4.828720414201184e-05,
967
+ "loss": 0.6306,
968
+ "step": 79500
969
+ },
970
+ {
971
+ "epoch": 15.7,
972
+ "learning_rate": 4.827487672583826e-05,
973
+ "loss": 0.6271,
974
+ "step": 80000
975
+ },
976
+ {
977
+ "epoch": 15.8,
978
+ "learning_rate": 4.8262573964497044e-05,
979
+ "loss": 0.6278,
980
+ "step": 80500
981
+ },
982
+ {
983
+ "epoch": 15.9,
984
+ "learning_rate": 4.825024654832347e-05,
985
+ "loss": 0.6245,
986
+ "step": 81000
987
+ },
988
+ {
989
+ "epoch": 16.0,
990
+ "learning_rate": 4.8237919132149906e-05,
991
+ "loss": 0.6236,
992
+ "step": 81500
993
+ },
994
+ {
995
+ "epoch": 16.09,
996
+ "learning_rate": 4.822561637080868e-05,
997
+ "loss": 0.622,
998
+ "step": 82000
999
+ },
1000
+ {
1001
+ "epoch": 16.19,
1002
+ "learning_rate": 4.821328895463511e-05,
1003
+ "loss": 0.6209,
1004
+ "step": 82500
1005
+ },
1006
+ {
1007
+ "epoch": 16.29,
1008
+ "learning_rate": 4.820096153846154e-05,
1009
+ "loss": 0.6187,
1010
+ "step": 83000
1011
+ },
1012
+ {
1013
+ "epoch": 16.39,
1014
+ "learning_rate": 4.818863412228797e-05,
1015
+ "loss": 0.6189,
1016
+ "step": 83500
1017
+ },
1018
+ {
1019
+ "epoch": 16.49,
1020
+ "learning_rate": 4.8176331360946746e-05,
1021
+ "loss": 0.6159,
1022
+ "step": 84000
1023
+ },
1024
+ {
1025
+ "epoch": 16.58,
1026
+ "learning_rate": 4.816400394477318e-05,
1027
+ "loss": 0.615,
1028
+ "step": 84500
1029
+ },
1030
+ {
1031
+ "epoch": 16.68,
1032
+ "learning_rate": 4.815167652859961e-05,
1033
+ "loss": 0.6155,
1034
+ "step": 85000
1035
+ },
1036
+ {
1037
+ "epoch": 16.78,
1038
+ "learning_rate": 4.8139349112426035e-05,
1039
+ "loss": 0.6121,
1040
+ "step": 85500
1041
+ },
1042
+ {
1043
+ "epoch": 16.88,
1044
+ "learning_rate": 4.812702169625247e-05,
1045
+ "loss": 0.6115,
1046
+ "step": 86000
1047
+ },
1048
+ {
1049
+ "epoch": 16.98,
1050
+ "learning_rate": 4.8114718934911245e-05,
1051
+ "loss": 0.6113,
1052
+ "step": 86500
1053
+ }
1054
+ ],
1055
+ "max_steps": 2038000,
1056
+ "num_train_epochs": 400,
1057
+ "total_flos": 2.3340589932538233e+19,
1058
+ "trial_name": null,
1059
+ "trial_params": null
1060
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5aef58edfbd2cb823c70074076feca97bc7dc16cebbeaab97e70f84adf08606b
3
+ size 3567