Upload 43 files
Browse files- config.json +26 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state_0.pth +3 -0
- rng_state_1.pth +3 -0
- rng_state_10.pth +3 -0
- rng_state_11.pth +3 -0
- rng_state_12.pth +3 -0
- rng_state_13.pth +3 -0
- rng_state_14.pth +3 -0
- rng_state_15.pth +3 -0
- rng_state_16.pth +3 -0
- rng_state_17.pth +3 -0
- rng_state_18.pth +3 -0
- rng_state_19.pth +3 -0
- rng_state_2.pth +3 -0
- rng_state_20.pth +3 -0
- rng_state_21.pth +3 -0
- rng_state_22.pth +3 -0
- rng_state_23.pth +3 -0
- rng_state_24.pth +3 -0
- rng_state_25.pth +3 -0
- rng_state_26.pth +3 -0
- rng_state_27.pth +3 -0
- rng_state_28.pth +3 -0
- rng_state_29.pth +3 -0
- rng_state_3.pth +3 -0
- rng_state_30.pth +3 -0
- rng_state_31.pth +3 -0
- rng_state_4.pth +3 -0
- rng_state_5.pth +3 -0
- rng_state_6.pth +3 -0
- rng_state_7.pth +3 -0
- rng_state_8.pth +3 -0
- rng_state_9.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- sentencepiece.bpe.model +3 -0
- special_tokens_map.json +19 -0
- tokenizer.json +0 -0
- tokenizer_config.json +27 -0
- trainer_state.json +1162 -0
- 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 sha256:756188d67318fa1a055950cf2eb569f5f323cdee0daa9b1649116d4903844808
|
| 3 |
+
size 885361433
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ca0e1b2e353e7cfc346c30405619fe1dce3e9e24f0efd3f80a8f1e0909ee5905
|
| 3 |
+
size 442693931
|
rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_10.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_11.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_12.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_13.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_14.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_15.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_16.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_17.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_18.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_19.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_20.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_21.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_22.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_23.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_24.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_25.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_26.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_27.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_28.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_29.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_30.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_31.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_4.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_5.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_6.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_7.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_8.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
rng_state_9.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47cc786f8ac6a6a33451664b6ab99e34fca764667acda64fdc68ac144ab42ccd
|
| 3 |
+
size 14503
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4458dc90a89464e6e59718d96232a722bd5b4b4cb80b60d0ce7ac85658b08f78
|
| 3 |
+
size 559
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a33ac90482b9734bb5a38700ef3f99b6f70aa024f4c9b6b7b654f131d290332
|
| 3 |
+
size 623
|
sentencepiece.bpe.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:647626f4859e5684180febd87b9bd6a39ec527b1cc4c60d6269d77c971830b71
|
| 3 |
+
size 791002
|
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_Wikipedia_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_Wikipedia_CHARS_lowercased_fixed_utf8_V4/special_tokens_map.json",
|
| 25 |
+
"tokenizer_class": "CamembertTokenizer",
|
| 26 |
+
"unk_token": "<unk>"
|
| 27 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1162 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 25.0,
|
| 5 |
+
"global_step": 95050,
|
| 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": 10.5513,
|
| 14 |
+
"step": 1
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.13,
|
| 18 |
+
"learning_rate": 2.5e-06,
|
| 19 |
+
"loss": 9.5944,
|
| 20 |
+
"step": 500
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.26,
|
| 24 |
+
"learning_rate": 5e-06,
|
| 25 |
+
"loss": 8.0091,
|
| 26 |
+
"step": 1000
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.39,
|
| 30 |
+
"learning_rate": 7.5e-06,
|
| 31 |
+
"loss": 6.96,
|
| 32 |
+
"step": 1500
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.53,
|
| 36 |
+
"learning_rate": 1e-05,
|
| 37 |
+
"loss": 6.6546,
|
| 38 |
+
"step": 2000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.66,
|
| 42 |
+
"learning_rate": 1.25e-05,
|
| 43 |
+
"loss": 6.4887,
|
| 44 |
+
"step": 2500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.79,
|
| 48 |
+
"learning_rate": 1.5e-05,
|
| 49 |
+
"loss": 6.3737,
|
| 50 |
+
"step": 3000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.92,
|
| 54 |
+
"learning_rate": 1.75e-05,
|
| 55 |
+
"loss": 6.2809,
|
| 56 |
+
"step": 3500
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.05,
|
| 60 |
+
"learning_rate": 2e-05,
|
| 61 |
+
"loss": 6.211,
|
| 62 |
+
"step": 4000
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 1.18,
|
| 66 |
+
"learning_rate": 2.25e-05,
|
| 67 |
+
"loss": 6.1475,
|
| 68 |
+
"step": 4500
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 1.32,
|
| 72 |
+
"learning_rate": 2.5e-05,
|
| 73 |
+
"loss": 6.0923,
|
| 74 |
+
"step": 5000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.45,
|
| 78 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 79 |
+
"loss": 6.0436,
|
| 80 |
+
"step": 5500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.58,
|
| 84 |
+
"learning_rate": 3e-05,
|
| 85 |
+
"loss": 6.0011,
|
| 86 |
+
"step": 6000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 1.71,
|
| 90 |
+
"learning_rate": 3.2495000000000007e-05,
|
| 91 |
+
"loss": 5.9694,
|
| 92 |
+
"step": 6500
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 1.84,
|
| 96 |
+
"learning_rate": 3.4995e-05,
|
| 97 |
+
"loss": 5.9366,
|
| 98 |
+
"step": 7000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.97,
|
| 102 |
+
"learning_rate": 3.7495e-05,
|
| 103 |
+
"loss": 5.9044,
|
| 104 |
+
"step": 7500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 2.1,
|
| 108 |
+
"learning_rate": 3.9995000000000006e-05,
|
| 109 |
+
"loss": 5.8788,
|
| 110 |
+
"step": 8000
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 2.24,
|
| 114 |
+
"learning_rate": 4.2495e-05,
|
| 115 |
+
"loss": 5.8576,
|
| 116 |
+
"step": 8500
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.37,
|
| 120 |
+
"learning_rate": 4.4995000000000005e-05,
|
| 121 |
+
"loss": 5.8413,
|
| 122 |
+
"step": 9000
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 2.5,
|
| 126 |
+
"learning_rate": 4.7495e-05,
|
| 127 |
+
"loss": 5.818,
|
| 128 |
+
"step": 9500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 2.63,
|
| 132 |
+
"learning_rate": 4.9995000000000005e-05,
|
| 133 |
+
"loss": 5.8055,
|
| 134 |
+
"step": 10000
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 2.76,
|
| 138 |
+
"learning_rate": 4.998348557055865e-05,
|
| 139 |
+
"loss": 5.7852,
|
| 140 |
+
"step": 10500
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 2.89,
|
| 144 |
+
"learning_rate": 4.9966971141117293e-05,
|
| 145 |
+
"loss": 5.768,
|
| 146 |
+
"step": 11000
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 3.02,
|
| 150 |
+
"learning_rate": 4.995042361662696e-05,
|
| 151 |
+
"loss": 5.7555,
|
| 152 |
+
"step": 11500
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 3.16,
|
| 156 |
+
"learning_rate": 4.993387609213662e-05,
|
| 157 |
+
"loss": 5.7434,
|
| 158 |
+
"step": 12000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 3.29,
|
| 162 |
+
"learning_rate": 4.991732856764628e-05,
|
| 163 |
+
"loss": 5.7306,
|
| 164 |
+
"step": 12500
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 3.42,
|
| 168 |
+
"learning_rate": 4.9900814138204925e-05,
|
| 169 |
+
"loss": 5.7245,
|
| 170 |
+
"step": 13000
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 3.55,
|
| 174 |
+
"learning_rate": 4.988426661371459e-05,
|
| 175 |
+
"loss": 5.7136,
|
| 176 |
+
"step": 13500
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.68,
|
| 180 |
+
"learning_rate": 4.986771908922425e-05,
|
| 181 |
+
"loss": 5.7022,
|
| 182 |
+
"step": 14000
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.81,
|
| 186 |
+
"learning_rate": 4.985117156473392e-05,
|
| 187 |
+
"loss": 5.6951,
|
| 188 |
+
"step": 14500
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.95,
|
| 192 |
+
"learning_rate": 4.983465713529256e-05,
|
| 193 |
+
"loss": 5.6872,
|
| 194 |
+
"step": 15000
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 4.08,
|
| 198 |
+
"learning_rate": 4.981810961080223e-05,
|
| 199 |
+
"loss": 5.6745,
|
| 200 |
+
"step": 15500
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 4.21,
|
| 204 |
+
"learning_rate": 4.9801562086311895e-05,
|
| 205 |
+
"loss": 5.6695,
|
| 206 |
+
"step": 16000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 4.34,
|
| 210 |
+
"learning_rate": 4.9785014561821555e-05,
|
| 211 |
+
"loss": 5.6663,
|
| 212 |
+
"step": 16500
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 4.47,
|
| 216 |
+
"learning_rate": 4.9768500132380194e-05,
|
| 217 |
+
"loss": 5.6611,
|
| 218 |
+
"step": 17000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 4.6,
|
| 222 |
+
"learning_rate": 4.975195260788986e-05,
|
| 223 |
+
"loss": 5.6569,
|
| 224 |
+
"step": 17500
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.73,
|
| 228 |
+
"learning_rate": 4.9735405083399527e-05,
|
| 229 |
+
"loss": 5.6487,
|
| 230 |
+
"step": 18000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.87,
|
| 234 |
+
"learning_rate": 4.9718857558909186e-05,
|
| 235 |
+
"loss": 5.6432,
|
| 236 |
+
"step": 18500
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 5.0,
|
| 240 |
+
"learning_rate": 4.970231003441885e-05,
|
| 241 |
+
"loss": 5.6399,
|
| 242 |
+
"step": 19000
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 5.13,
|
| 246 |
+
"learning_rate": 4.96857956049775e-05,
|
| 247 |
+
"loss": 5.6297,
|
| 248 |
+
"step": 19500
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 5.26,
|
| 252 |
+
"learning_rate": 4.9669248080487165e-05,
|
| 253 |
+
"loss": 5.625,
|
| 254 |
+
"step": 20000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 5.39,
|
| 258 |
+
"learning_rate": 4.9652700555996824e-05,
|
| 259 |
+
"loss": 5.6231,
|
| 260 |
+
"step": 20500
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 5.52,
|
| 264 |
+
"learning_rate": 4.963615303150649e-05,
|
| 265 |
+
"loss": 5.6183,
|
| 266 |
+
"step": 21000
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 5.65,
|
| 270 |
+
"learning_rate": 4.961963860206513e-05,
|
| 271 |
+
"loss": 5.6134,
|
| 272 |
+
"step": 21500
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 5.79,
|
| 276 |
+
"learning_rate": 4.9603091077574796e-05,
|
| 277 |
+
"loss": 5.6107,
|
| 278 |
+
"step": 22000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 5.92,
|
| 282 |
+
"learning_rate": 4.958654355308446e-05,
|
| 283 |
+
"loss": 5.6064,
|
| 284 |
+
"step": 22500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 6.05,
|
| 288 |
+
"learning_rate": 4.956999602859412e-05,
|
| 289 |
+
"loss": 5.6008,
|
| 290 |
+
"step": 23000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 6.18,
|
| 294 |
+
"learning_rate": 4.955348159915277e-05,
|
| 295 |
+
"loss": 5.5932,
|
| 296 |
+
"step": 23500
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 6.31,
|
| 300 |
+
"learning_rate": 4.9536934074662434e-05,
|
| 301 |
+
"loss": 5.5929,
|
| 302 |
+
"step": 24000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 6.44,
|
| 306 |
+
"learning_rate": 4.95203865501721e-05,
|
| 307 |
+
"loss": 5.5891,
|
| 308 |
+
"step": 24500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 6.58,
|
| 312 |
+
"learning_rate": 4.9503872120730746e-05,
|
| 313 |
+
"loss": 5.5828,
|
| 314 |
+
"step": 25000
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 6.71,
|
| 318 |
+
"learning_rate": 4.9487324596240405e-05,
|
| 319 |
+
"loss": 5.5846,
|
| 320 |
+
"step": 25500
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 6.84,
|
| 324 |
+
"learning_rate": 4.947077707175007e-05,
|
| 325 |
+
"loss": 5.581,
|
| 326 |
+
"step": 26000
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 6.97,
|
| 330 |
+
"learning_rate": 4.945422954725973e-05,
|
| 331 |
+
"loss": 5.5753,
|
| 332 |
+
"step": 26500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 7.1,
|
| 336 |
+
"learning_rate": 4.943768202276939e-05,
|
| 337 |
+
"loss": 5.5719,
|
| 338 |
+
"step": 27000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 7.23,
|
| 342 |
+
"learning_rate": 4.942113449827906e-05,
|
| 343 |
+
"loss": 5.5668,
|
| 344 |
+
"step": 27500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 7.36,
|
| 348 |
+
"learning_rate": 4.9404586973788724e-05,
|
| 349 |
+
"loss": 5.564,
|
| 350 |
+
"step": 28000
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 7.5,
|
| 354 |
+
"learning_rate": 4.938807254434737e-05,
|
| 355 |
+
"loss": 5.5632,
|
| 356 |
+
"step": 28500
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 7.63,
|
| 360 |
+
"learning_rate": 4.9371525019857036e-05,
|
| 361 |
+
"loss": 5.3943,
|
| 362 |
+
"step": 29000
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 7.76,
|
| 366 |
+
"learning_rate": 4.9354977495366695e-05,
|
| 367 |
+
"loss": 5.1613,
|
| 368 |
+
"step": 29500
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 7.89,
|
| 372 |
+
"learning_rate": 4.933842997087636e-05,
|
| 373 |
+
"loss": 4.9594,
|
| 374 |
+
"step": 30000
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 8.02,
|
| 378 |
+
"learning_rate": 4.932188244638603e-05,
|
| 379 |
+
"loss": 4.7739,
|
| 380 |
+
"step": 30500
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 8.15,
|
| 384 |
+
"learning_rate": 4.930533492189569e-05,
|
| 385 |
+
"loss": 4.6014,
|
| 386 |
+
"step": 31000
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 8.29,
|
| 390 |
+
"learning_rate": 4.928878739740535e-05,
|
| 391 |
+
"loss": 4.4413,
|
| 392 |
+
"step": 31500
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 8.42,
|
| 396 |
+
"learning_rate": 4.9272239872915014e-05,
|
| 397 |
+
"loss": 4.2832,
|
| 398 |
+
"step": 32000
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 8.55,
|
| 402 |
+
"learning_rate": 4.925572544347366e-05,
|
| 403 |
+
"loss": 4.1308,
|
| 404 |
+
"step": 32500
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 8.68,
|
| 408 |
+
"learning_rate": 4.923917791898332e-05,
|
| 409 |
+
"loss": 3.9864,
|
| 410 |
+
"step": 33000
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 8.81,
|
| 414 |
+
"learning_rate": 4.9222630394492985e-05,
|
| 415 |
+
"loss": 3.8488,
|
| 416 |
+
"step": 33500
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 8.94,
|
| 420 |
+
"learning_rate": 4.920608287000265e-05,
|
| 421 |
+
"loss": 3.699,
|
| 422 |
+
"step": 34000
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 9.07,
|
| 426 |
+
"learning_rate": 4.918953534551232e-05,
|
| 427 |
+
"loss": 3.5521,
|
| 428 |
+
"step": 34500
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 9.21,
|
| 432 |
+
"learning_rate": 4.917302091607096e-05,
|
| 433 |
+
"loss": 3.3783,
|
| 434 |
+
"step": 35000
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 9.34,
|
| 438 |
+
"learning_rate": 4.915647339158062e-05,
|
| 439 |
+
"loss": 2.8694,
|
| 440 |
+
"step": 35500
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 9.47,
|
| 444 |
+
"learning_rate": 4.913992586709028e-05,
|
| 445 |
+
"loss": 2.3898,
|
| 446 |
+
"step": 36000
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 9.6,
|
| 450 |
+
"learning_rate": 4.912337834259995e-05,
|
| 451 |
+
"loss": 2.1786,
|
| 452 |
+
"step": 36500
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 9.73,
|
| 456 |
+
"learning_rate": 4.9106863913158595e-05,
|
| 457 |
+
"loss": 2.0476,
|
| 458 |
+
"step": 37000
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 9.86,
|
| 462 |
+
"learning_rate": 4.9090316388668254e-05,
|
| 463 |
+
"loss": 1.9528,
|
| 464 |
+
"step": 37500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 9.99,
|
| 468 |
+
"learning_rate": 4.90738019592269e-05,
|
| 469 |
+
"loss": 1.8765,
|
| 470 |
+
"step": 38000
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 10.13,
|
| 474 |
+
"learning_rate": 4.9057254434736566e-05,
|
| 475 |
+
"loss": 1.8172,
|
| 476 |
+
"step": 38500
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 10.26,
|
| 480 |
+
"learning_rate": 4.904070691024623e-05,
|
| 481 |
+
"loss": 1.7658,
|
| 482 |
+
"step": 39000
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 10.39,
|
| 486 |
+
"learning_rate": 4.902415938575589e-05,
|
| 487 |
+
"loss": 1.7213,
|
| 488 |
+
"step": 39500
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 10.52,
|
| 492 |
+
"learning_rate": 4.900764495631454e-05,
|
| 493 |
+
"loss": 1.6845,
|
| 494 |
+
"step": 40000
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 10.65,
|
| 498 |
+
"learning_rate": 4.8991097431824204e-05,
|
| 499 |
+
"loss": 1.6475,
|
| 500 |
+
"step": 40500
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 10.78,
|
| 504 |
+
"learning_rate": 4.8974549907333864e-05,
|
| 505 |
+
"loss": 1.6163,
|
| 506 |
+
"step": 41000
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 10.92,
|
| 510 |
+
"learning_rate": 4.8958002382843524e-05,
|
| 511 |
+
"loss": 1.5924,
|
| 512 |
+
"step": 41500
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 11.05,
|
| 516 |
+
"learning_rate": 4.894148795340217e-05,
|
| 517 |
+
"loss": 1.5644,
|
| 518 |
+
"step": 42000
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 11.18,
|
| 522 |
+
"learning_rate": 4.8924940428911836e-05,
|
| 523 |
+
"loss": 1.5381,
|
| 524 |
+
"step": 42500
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 11.31,
|
| 528 |
+
"learning_rate": 4.89083929044215e-05,
|
| 529 |
+
"loss": 1.5176,
|
| 530 |
+
"step": 43000
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 11.44,
|
| 534 |
+
"learning_rate": 4.889184537993117e-05,
|
| 535 |
+
"loss": 1.4933,
|
| 536 |
+
"step": 43500
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 11.57,
|
| 540 |
+
"learning_rate": 4.887529785544083e-05,
|
| 541 |
+
"loss": 1.4755,
|
| 542 |
+
"step": 44000
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 11.7,
|
| 546 |
+
"learning_rate": 4.8858783425999473e-05,
|
| 547 |
+
"loss": 1.4564,
|
| 548 |
+
"step": 44500
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 11.84,
|
| 552 |
+
"learning_rate": 4.884223590150914e-05,
|
| 553 |
+
"loss": 1.4382,
|
| 554 |
+
"step": 45000
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 11.97,
|
| 558 |
+
"learning_rate": 4.88256883770188e-05,
|
| 559 |
+
"loss": 1.4251,
|
| 560 |
+
"step": 45500
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 12.1,
|
| 564 |
+
"learning_rate": 4.880914085252846e-05,
|
| 565 |
+
"loss": 1.4069,
|
| 566 |
+
"step": 46000
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 12.23,
|
| 570 |
+
"learning_rate": 4.8792593328038126e-05,
|
| 571 |
+
"loss": 1.3901,
|
| 572 |
+
"step": 46500
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 12.36,
|
| 576 |
+
"learning_rate": 4.877604580354779e-05,
|
| 577 |
+
"loss": 1.3754,
|
| 578 |
+
"step": 47000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 12.49,
|
| 582 |
+
"learning_rate": 4.875949827905745e-05,
|
| 583 |
+
"loss": 1.3633,
|
| 584 |
+
"step": 47500
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 12.62,
|
| 588 |
+
"learning_rate": 4.874295075456712e-05,
|
| 589 |
+
"loss": 1.3459,
|
| 590 |
+
"step": 48000
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 12.76,
|
| 594 |
+
"learning_rate": 4.8726436325125763e-05,
|
| 595 |
+
"loss": 1.3374,
|
| 596 |
+
"step": 48500
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 12.89,
|
| 600 |
+
"learning_rate": 4.870992189568441e-05,
|
| 601 |
+
"loss": 1.3237,
|
| 602 |
+
"step": 49000
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 13.02,
|
| 606 |
+
"learning_rate": 4.8693374371194075e-05,
|
| 607 |
+
"loss": 1.3117,
|
| 608 |
+
"step": 49500
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 13.15,
|
| 612 |
+
"learning_rate": 4.8676826846703735e-05,
|
| 613 |
+
"loss": 1.3009,
|
| 614 |
+
"step": 50000
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 13.28,
|
| 618 |
+
"learning_rate": 4.8660279322213395e-05,
|
| 619 |
+
"loss": 1.2906,
|
| 620 |
+
"step": 50500
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 13.41,
|
| 624 |
+
"learning_rate": 4.864373179772306e-05,
|
| 625 |
+
"loss": 1.2816,
|
| 626 |
+
"step": 51000
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 13.55,
|
| 630 |
+
"learning_rate": 4.8627217368281707e-05,
|
| 631 |
+
"loss": 1.2717,
|
| 632 |
+
"step": 51500
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 13.68,
|
| 636 |
+
"learning_rate": 4.861066984379137e-05,
|
| 637 |
+
"loss": 1.2648,
|
| 638 |
+
"step": 52000
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 13.81,
|
| 642 |
+
"learning_rate": 4.859412231930103e-05,
|
| 643 |
+
"loss": 1.2561,
|
| 644 |
+
"step": 52500
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 13.94,
|
| 648 |
+
"learning_rate": 4.85775747948107e-05,
|
| 649 |
+
"loss": 1.2473,
|
| 650 |
+
"step": 53000
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 14.07,
|
| 654 |
+
"learning_rate": 4.8561027270320365e-05,
|
| 655 |
+
"loss": 1.2351,
|
| 656 |
+
"step": 53500
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 14.2,
|
| 660 |
+
"learning_rate": 4.854451284087901e-05,
|
| 661 |
+
"loss": 1.2292,
|
| 662 |
+
"step": 54000
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 14.33,
|
| 666 |
+
"learning_rate": 4.852796531638867e-05,
|
| 667 |
+
"loss": 1.2221,
|
| 668 |
+
"step": 54500
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 14.47,
|
| 672 |
+
"learning_rate": 4.851141779189834e-05,
|
| 673 |
+
"loss": 1.2137,
|
| 674 |
+
"step": 55000
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 14.6,
|
| 678 |
+
"learning_rate": 4.8494870267407997e-05,
|
| 679 |
+
"loss": 1.2065,
|
| 680 |
+
"step": 55500
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 14.73,
|
| 684 |
+
"learning_rate": 4.847832274291766e-05,
|
| 685 |
+
"loss": 1.1981,
|
| 686 |
+
"step": 56000
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 14.86,
|
| 690 |
+
"learning_rate": 4.846180831347631e-05,
|
| 691 |
+
"loss": 1.1931,
|
| 692 |
+
"step": 56500
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 14.99,
|
| 696 |
+
"learning_rate": 4.844526078898597e-05,
|
| 697 |
+
"loss": 1.1876,
|
| 698 |
+
"step": 57000
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 15.12,
|
| 702 |
+
"learning_rate": 4.8428713264495635e-05,
|
| 703 |
+
"loss": 1.1777,
|
| 704 |
+
"step": 57500
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 15.26,
|
| 708 |
+
"learning_rate": 4.84121657400053e-05,
|
| 709 |
+
"loss": 1.1734,
|
| 710 |
+
"step": 58000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 15.39,
|
| 714 |
+
"learning_rate": 4.8395651310563947e-05,
|
| 715 |
+
"loss": 1.1639,
|
| 716 |
+
"step": 58500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 15.52,
|
| 720 |
+
"learning_rate": 4.8379103786073606e-05,
|
| 721 |
+
"loss": 1.1619,
|
| 722 |
+
"step": 59000
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 15.65,
|
| 726 |
+
"learning_rate": 4.836255626158327e-05,
|
| 727 |
+
"loss": 1.1534,
|
| 728 |
+
"step": 59500
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 15.78,
|
| 732 |
+
"learning_rate": 4.834600873709293e-05,
|
| 733 |
+
"loss": 1.1484,
|
| 734 |
+
"step": 60000
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 15.91,
|
| 738 |
+
"learning_rate": 4.832946121260259e-05,
|
| 739 |
+
"loss": 1.1453,
|
| 740 |
+
"step": 60500
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 16.04,
|
| 744 |
+
"learning_rate": 4.831291368811226e-05,
|
| 745 |
+
"loss": 1.1395,
|
| 746 |
+
"step": 61000
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 16.18,
|
| 750 |
+
"learning_rate": 4.8296366163621925e-05,
|
| 751 |
+
"loss": 1.1322,
|
| 752 |
+
"step": 61500
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 16.31,
|
| 756 |
+
"learning_rate": 4.827985173418057e-05,
|
| 757 |
+
"loss": 1.1269,
|
| 758 |
+
"step": 62000
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 16.44,
|
| 762 |
+
"learning_rate": 4.826330420969023e-05,
|
| 763 |
+
"loss": 1.1231,
|
| 764 |
+
"step": 62500
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 16.57,
|
| 768 |
+
"learning_rate": 4.8246756685199896e-05,
|
| 769 |
+
"loss": 1.1167,
|
| 770 |
+
"step": 63000
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 16.7,
|
| 774 |
+
"learning_rate": 4.823020916070956e-05,
|
| 775 |
+
"loss": 1.1136,
|
| 776 |
+
"step": 63500
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 16.83,
|
| 780 |
+
"learning_rate": 4.821366163621923e-05,
|
| 781 |
+
"loss": 1.107,
|
| 782 |
+
"step": 64000
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 16.96,
|
| 786 |
+
"learning_rate": 4.819714720677787e-05,
|
| 787 |
+
"loss": 1.1027,
|
| 788 |
+
"step": 64500
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 17.1,
|
| 792 |
+
"learning_rate": 4.818059968228753e-05,
|
| 793 |
+
"loss": 1.0964,
|
| 794 |
+
"step": 65000
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 17.23,
|
| 798 |
+
"learning_rate": 4.8164052157797194e-05,
|
| 799 |
+
"loss": 1.0933,
|
| 800 |
+
"step": 65500
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 17.36,
|
| 804 |
+
"learning_rate": 4.814750463330686e-05,
|
| 805 |
+
"loss": 1.088,
|
| 806 |
+
"step": 66000
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 17.49,
|
| 810 |
+
"learning_rate": 4.813095710881652e-05,
|
| 811 |
+
"loss": 1.0861,
|
| 812 |
+
"step": 66500
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 17.62,
|
| 816 |
+
"learning_rate": 4.8114442679375165e-05,
|
| 817 |
+
"loss": 1.0817,
|
| 818 |
+
"step": 67000
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 17.75,
|
| 822 |
+
"learning_rate": 4.809789515488483e-05,
|
| 823 |
+
"loss": 1.0783,
|
| 824 |
+
"step": 67500
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 17.89,
|
| 828 |
+
"learning_rate": 4.80813476303945e-05,
|
| 829 |
+
"loss": 1.075,
|
| 830 |
+
"step": 68000
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 18.02,
|
| 834 |
+
"learning_rate": 4.8064833200953144e-05,
|
| 835 |
+
"loss": 1.0672,
|
| 836 |
+
"step": 68500
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 18.15,
|
| 840 |
+
"learning_rate": 4.80482856764628e-05,
|
| 841 |
+
"loss": 1.0647,
|
| 842 |
+
"step": 69000
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 18.28,
|
| 846 |
+
"learning_rate": 4.803173815197247e-05,
|
| 847 |
+
"loss": 1.0618,
|
| 848 |
+
"step": 69500
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 18.41,
|
| 852 |
+
"learning_rate": 4.801519062748213e-05,
|
| 853 |
+
"loss": 1.057,
|
| 854 |
+
"step": 70000
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 18.54,
|
| 858 |
+
"learning_rate": 4.7998643102991796e-05,
|
| 859 |
+
"loss": 1.0528,
|
| 860 |
+
"step": 70500
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 18.67,
|
| 864 |
+
"learning_rate": 4.7982095578501455e-05,
|
| 865 |
+
"loss": 1.0503,
|
| 866 |
+
"step": 71000
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 18.81,
|
| 870 |
+
"learning_rate": 4.79655811490601e-05,
|
| 871 |
+
"loss": 1.0455,
|
| 872 |
+
"step": 71500
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 18.94,
|
| 876 |
+
"learning_rate": 4.794903362456977e-05,
|
| 877 |
+
"loss": 1.0429,
|
| 878 |
+
"step": 72000
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 19.07,
|
| 882 |
+
"learning_rate": 4.7932486100079434e-05,
|
| 883 |
+
"loss": 1.0371,
|
| 884 |
+
"step": 72500
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 19.2,
|
| 888 |
+
"learning_rate": 4.791593857558909e-05,
|
| 889 |
+
"loss": 1.0341,
|
| 890 |
+
"step": 73000
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 19.33,
|
| 894 |
+
"learning_rate": 4.789939105109876e-05,
|
| 895 |
+
"loss": 1.0291,
|
| 896 |
+
"step": 73500
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 19.46,
|
| 900 |
+
"learning_rate": 4.7882876621657405e-05,
|
| 901 |
+
"loss": 1.0261,
|
| 902 |
+
"step": 74000
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 19.59,
|
| 906 |
+
"learning_rate": 4.7866329097167065e-05,
|
| 907 |
+
"loss": 1.0241,
|
| 908 |
+
"step": 74500
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 19.73,
|
| 912 |
+
"learning_rate": 4.7849781572676724e-05,
|
| 913 |
+
"loss": 1.02,
|
| 914 |
+
"step": 75000
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 19.86,
|
| 918 |
+
"learning_rate": 4.783323404818639e-05,
|
| 919 |
+
"loss": 1.018,
|
| 920 |
+
"step": 75500
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 19.99,
|
| 924 |
+
"learning_rate": 4.781668652369606e-05,
|
| 925 |
+
"loss": 1.0154,
|
| 926 |
+
"step": 76000
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 20.12,
|
| 930 |
+
"learning_rate": 4.7800138999205724e-05,
|
| 931 |
+
"loss": 1.0116,
|
| 932 |
+
"step": 76500
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 20.25,
|
| 936 |
+
"learning_rate": 4.778359147471538e-05,
|
| 937 |
+
"loss": 1.0065,
|
| 938 |
+
"step": 77000
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 20.38,
|
| 942 |
+
"learning_rate": 4.776704395022505e-05,
|
| 943 |
+
"loss": 1.005,
|
| 944 |
+
"step": 77500
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 20.52,
|
| 948 |
+
"learning_rate": 4.7750529520783695e-05,
|
| 949 |
+
"loss": 1.0025,
|
| 950 |
+
"step": 78000
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 20.65,
|
| 954 |
+
"learning_rate": 4.773398199629336e-05,
|
| 955 |
+
"loss": 1.0008,
|
| 956 |
+
"step": 78500
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 20.78,
|
| 960 |
+
"learning_rate": 4.771743447180302e-05,
|
| 961 |
+
"loss": 0.9965,
|
| 962 |
+
"step": 79000
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 20.91,
|
| 966 |
+
"learning_rate": 4.770088694731268e-05,
|
| 967 |
+
"loss": 0.9943,
|
| 968 |
+
"step": 79500
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 21.04,
|
| 972 |
+
"learning_rate": 4.7684372517871326e-05,
|
| 973 |
+
"loss": 0.9907,
|
| 974 |
+
"step": 80000
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 21.17,
|
| 978 |
+
"learning_rate": 4.766782499338099e-05,
|
| 979 |
+
"loss": 0.9882,
|
| 980 |
+
"step": 80500
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 21.3,
|
| 984 |
+
"learning_rate": 4.765127746889065e-05,
|
| 985 |
+
"loss": 0.9861,
|
| 986 |
+
"step": 81000
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 21.44,
|
| 990 |
+
"learning_rate": 4.763472994440032e-05,
|
| 991 |
+
"loss": 0.9809,
|
| 992 |
+
"step": 81500
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 21.57,
|
| 996 |
+
"learning_rate": 4.7618215514958964e-05,
|
| 997 |
+
"loss": 0.9793,
|
| 998 |
+
"step": 82000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 21.7,
|
| 1002 |
+
"learning_rate": 4.760166799046863e-05,
|
| 1003 |
+
"loss": 0.9778,
|
| 1004 |
+
"step": 82500
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 21.83,
|
| 1008 |
+
"learning_rate": 4.7585153561027276e-05,
|
| 1009 |
+
"loss": 0.9756,
|
| 1010 |
+
"step": 83000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 21.96,
|
| 1014 |
+
"learning_rate": 4.7568606036536936e-05,
|
| 1015 |
+
"loss": 0.9732,
|
| 1016 |
+
"step": 83500
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 22.09,
|
| 1020 |
+
"learning_rate": 4.7552058512046596e-05,
|
| 1021 |
+
"loss": 0.97,
|
| 1022 |
+
"step": 84000
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 22.23,
|
| 1026 |
+
"learning_rate": 4.753551098755626e-05,
|
| 1027 |
+
"loss": 0.9687,
|
| 1028 |
+
"step": 84500
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 22.36,
|
| 1032 |
+
"learning_rate": 4.751899655811491e-05,
|
| 1033 |
+
"loss": 0.965,
|
| 1034 |
+
"step": 85000
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 22.49,
|
| 1038 |
+
"learning_rate": 4.7502449033624574e-05,
|
| 1039 |
+
"loss": 0.9627,
|
| 1040 |
+
"step": 85500
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 22.62,
|
| 1044 |
+
"learning_rate": 4.7485901509134233e-05,
|
| 1045 |
+
"loss": 0.9628,
|
| 1046 |
+
"step": 86000
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 22.75,
|
| 1050 |
+
"learning_rate": 4.74693539846439e-05,
|
| 1051 |
+
"loss": 0.9605,
|
| 1052 |
+
"step": 86500
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 22.88,
|
| 1056 |
+
"learning_rate": 4.7452806460153566e-05,
|
| 1057 |
+
"loss": 0.9565,
|
| 1058 |
+
"step": 87000
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 23.01,
|
| 1062 |
+
"learning_rate": 4.7436258935663226e-05,
|
| 1063 |
+
"loss": 0.9553,
|
| 1064 |
+
"step": 87500
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 23.15,
|
| 1068 |
+
"learning_rate": 4.741971141117289e-05,
|
| 1069 |
+
"loss": 0.9518,
|
| 1070 |
+
"step": 88000
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 23.28,
|
| 1074 |
+
"learning_rate": 4.740316388668256e-05,
|
| 1075 |
+
"loss": 0.9504,
|
| 1076 |
+
"step": 88500
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 23.41,
|
| 1080 |
+
"learning_rate": 4.73866494572412e-05,
|
| 1081 |
+
"loss": 0.9482,
|
| 1082 |
+
"step": 89000
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 23.54,
|
| 1086 |
+
"learning_rate": 4.7370101932750864e-05,
|
| 1087 |
+
"loss": 0.946,
|
| 1088 |
+
"step": 89500
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 23.67,
|
| 1092 |
+
"learning_rate": 4.7353554408260523e-05,
|
| 1093 |
+
"loss": 0.9442,
|
| 1094 |
+
"step": 90000
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 23.8,
|
| 1098 |
+
"learning_rate": 4.733700688377019e-05,
|
| 1099 |
+
"loss": 0.9428,
|
| 1100 |
+
"step": 90500
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 23.93,
|
| 1104 |
+
"learning_rate": 4.7320492454328835e-05,
|
| 1105 |
+
"loss": 0.9421,
|
| 1106 |
+
"step": 91000
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 24.07,
|
| 1110 |
+
"learning_rate": 4.73039449298385e-05,
|
| 1111 |
+
"loss": 0.9389,
|
| 1112 |
+
"step": 91500
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 24.2,
|
| 1116 |
+
"learning_rate": 4.728739740534816e-05,
|
| 1117 |
+
"loss": 0.9374,
|
| 1118 |
+
"step": 92000
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 24.33,
|
| 1122 |
+
"learning_rate": 4.727084988085783e-05,
|
| 1123 |
+
"loss": 0.9337,
|
| 1124 |
+
"step": 92500
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 24.46,
|
| 1128 |
+
"learning_rate": 4.7254302356367494e-05,
|
| 1129 |
+
"loss": 0.9332,
|
| 1130 |
+
"step": 93000
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 24.59,
|
| 1134 |
+
"learning_rate": 4.7237754831877154e-05,
|
| 1135 |
+
"loss": 0.931,
|
| 1136 |
+
"step": 93500
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 24.72,
|
| 1140 |
+
"learning_rate": 4.722124040243579e-05,
|
| 1141 |
+
"loss": 0.9301,
|
| 1142 |
+
"step": 94000
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 24.86,
|
| 1146 |
+
"learning_rate": 4.720469287794546e-05,
|
| 1147 |
+
"loss": 0.9296,
|
| 1148 |
+
"step": 94500
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 24.99,
|
| 1152 |
+
"learning_rate": 4.7188145353455125e-05,
|
| 1153 |
+
"loss": 0.926,
|
| 1154 |
+
"step": 95000
|
| 1155 |
+
}
|
| 1156 |
+
],
|
| 1157 |
+
"max_steps": 1520800,
|
| 1158 |
+
"num_train_epochs": 400,
|
| 1159 |
+
"total_flos": 2.561490423429831e+19,
|
| 1160 |
+
"trial_name": null,
|
| 1161 |
+
"trial_params": null
|
| 1162 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6bd1965bd43c23a129e14c29cc491fa9ea519ee5bd59ab7a894ee6b05c9b1d36
|
| 3 |
+
size 3503
|