Hugging Face's logo

jed351
/
gpt2_tiny_zh-hk-shikoto

Text Generation
Transformers
PyTorch
gpt2
Generated from Trainer
Eval Results
text-generation-inference
Inference Endpoints
Model card Files Community
1
jed351 commited on
Commit
86a5699
•
1 Parent(s): 53ab4c5

Upload 15 files

Browse files
Files changed (15) hide show
  1. README.md +72 -0
  2. added_tokens.json +127 -0
  3. all_results.json +15 -0
  4. config.json +41 -0
  5. eval_results.json +10 -0
  6. merges.txt +0 -0
  7. pytorch_model.bin +3 -0
  8. special_tokens_map.json +7 -0
  9. tokenizer.json +0 -0
  10. tokenizer_config.json +14 -0
  11. train_results.json +8 -0
  12. trainer_state.json +4825 -0
  13. training_args.bin +3 -0
  14. vocab.json +0 -0
  15. vocab.txt +0 -0
README.md ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_trainer
4
+ datasets:
5
+ - jed351/shikoto_zh_hk
6
+ metrics:
7
+ - accuracy
8
+ model-index:
9
+ - name: gpt2-shikoto
10
+ results:
11
+ - task:
12
+ name: Causal Language Modeling
13
+ type: text-generation
14
+ dataset:
15
+ name: jed351/shikoto_zh_hk
16
+ type: jed351/shikoto_zh_hk
17
+ metrics:
18
+ - name: Accuracy
19
+ type: accuracy
20
+ value: 0.37381769930940056
21
+ ---
22
+
23
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
24
+ should probably proofread and complete it, then remove this comment. -->
25
+
26
+ # gpt2-shikoto
27
+
28
+ This model is a fine-tuned version of [jed351/gpt2-tiny-zh-hk](https://huggingface.co/jed351/gpt2-tiny-zh-hk) on the jed351/shikoto_zh_hk dataset.
29
+ It achieves the following results on the evaluation set:
30
+ - Loss: 3.2965
31
+ - Accuracy: 0.3738
32
+
33
+ ## Model description
34
+
35
+ More information needed
36
+
37
+ ## Intended uses & limitations
38
+
39
+ More information needed
40
+
41
+ ## Training and evaluation data
42
+
43
+ More information needed
44
+
45
+ ## Training procedure
46
+
47
+ ### Training hyperparameters
48
+
49
+ The following hyperparameters were used during training:
50
+ - learning_rate: 5e-05
51
+ - train_batch_size: 20
52
+ - eval_batch_size: 20
53
+ - seed: 42
54
+ - distributed_type: multi-GPU
55
+ - num_devices: 2
56
+ - total_train_batch_size: 40
57
+ - total_eval_batch_size: 40
58
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
59
+ - lr_scheduler_type: linear
60
+ - training_steps: 400000
61
+ - mixed_precision_training: Native AMP
62
+
63
+ ### Training results
64
+
65
+
66
+
67
+ ### Framework versions
68
+
69
+ - Transformers 4.26.0.dev0
70
+ - Pytorch 1.13.1
71
+ - Datasets 2.8.0
72
+ - Tokenizers 0.13.2
added_tokens.json ADDED
@@ -0,0 +1,127 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "㖭": 21128,
3
+ "㨂": 21129,
4
+ "㩒": 21130,
5
+ "䏙": 21131,
6
+ "䚗": 21132,
7
+ "䟕": 21133,
8
+ "乸": 21134,
9
+ "仼": 21135,
10
+ "冚": 21136,
11
+ "冧": 21137,
12
+ "厠": 21138,
13
+ "厹": 21139,
14
+ "吔": 21140,
15
+ "呔": 21141,
16
+ "咇": 21142,
17
+ "唒": 21143,
18
+ "唓": 21144,
19
+ "唞": 21145,
20
+ "啩": 21146,
21
+ "喼": 21147,
22
+ "嗄": 21148,
23
+ "嗌": 21149,
24
+ "嗮": 21150,
25
+ "嗱": 21151,
26
+ "嘥": 21152,
27
+ "噃": 21153,
28
+ "噉": 21154,
29
+ "噏": 21155,
30
+ "噚": 21156,
31
+ "嚹": 21157,
32
+ "嚿": 21158,
33
+ "囇": 21159,
34
+ "囯": 21160,
35
+ "妺": 21161,
36
+ "嬲": 21162,
37
+ "孭": 21163,
38
+ "屙": 21164,
39
+ "廸": 21165,
40
+ "彊": 21166,
41
+ "戇": 21167,
42
+ "戥": 21168,
43
+ "扚": 21169,
44
+ "扠": 21170,
45
+ "扺": 21171,
46
+ "扻": 21172,
47
+ "抆": 21173,
48
+ "抌": 21174,
49
+ "抺": 21175,
50
+ "捽": 21176,
51
+ "掕": 21177,
52
+ "掟": 21178,
53
+ "掹": 21179,
54
+ "揦": 21180,
55
+ "揼": 21181,
56
+ "揾": 21182,
57
+ "搣": 21183,
58
+ "搥": 21184,
59
+ "搲": 21185,
60
+ "摙": 21186,
61
+ "摷": 21187,
62
+ "撘": 21188,
63
+ "撳": 21189,
64
+ "攋": 21190,
65
+ "攰": 21191,
66
+ "曱": 21192,
67
+ "栢": 21193,
68
+ "梘": 21194,
69
+ "櫈": 21195,
70
+ "欅": 21196,
71
+ "殻": 21197,
72
+ "炆": 21198,
73
+ "烚": 21199,
74
+ "燶": 21200,
75
+ "猁": 21201,
76
+ "猢": 21202,
77
+ "獌": 21203,
78
+ "甴": 21204,
79
+ "痾": 21205,
80
+ "瘜": 21206,
81
+ "癈": 21207,
82
+ "癪": 21208,
83
+ "砵": 21209,
84
+ "祼": 21210,
85
+ "罅": 21211,
86
+ "罉": 21212,
87
+ "脷": 21213,
88
+ "蚑": 21214,
89
+ "蝻": 21215,
90
+ "袓": 21216,
91
+ "裇": 21217,
92
+ "褸": 21218,
93
+ "趷": 21219,
94
+ "跣": 21220,
95
+ "踎": 21221,
96
+ "踭": 21222,
97
+ "踼": 21223,
98
+ "躝": 21224,
99
+ "軚": 21225,
100
+ "鋭": 21226,
101
+ "錡": 21227,
102
+ "鍠": 21228,
103
+ "鎅": 21229,
104
+ "鏝": 21230,
105
+ "閂": 21231,
106
+ "閪": 21232,
107
+ "餸": 21233,
108
+ "騮": 21234,
109
+ "骹": 21235,
110
+ "髀": 21236,
111
+ "鰂": 21237,
112
+ "麖": 21238,
113
+ "黐": 21239,
114
+ "𠝹": 21240,
115
+ "𠵱": 21241,
116
+ "𠹌": 21242,
117
+ "𠺌": 21243,
118
+ "𠺘": 21244,
119
+ "𠻹": 21245,
120
+ "𠿪": 21246,
121
+ "𡃁": 21247,
122
+ "𢃇": 21248,
123
+ "𥄫": 21249,
124
+ "𨋢": 21250,
125
+ "𨳒": 21251,
126
+ "𪘲": 21252
127
+ }
all_results.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 36.08,
3
+ "eval_accuracy": 0.37381769930940056,
4
+ "eval_loss": 3.296532154083252,
5
+ "eval_runtime": 116.1554,
6
+ "eval_samples": 26482,
7
+ "eval_samples_per_second": 227.988,
8
+ "eval_steps_per_second": 5.708,
9
+ "perplexity": 27.018779305888142,
10
+ "train_loss": 3.451316015625,
11
+ "train_runtime": 66733.7733,
12
+ "train_samples": 443474,
13
+ "train_samples_per_second": 239.759,
14
+ "train_steps_per_second": 5.994
15
+ }
config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "jed351/gpt2-tiny-zh-hk",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 101,
9
+ "embd_pdrop": 0.1,
10
+ "eos_token_id": 102,
11
+ "gradient_checkpointing": false,
12
+ "initializer_range": 0.02,
13
+ "layer_norm_epsilon": 1e-05,
14
+ "model_type": "gpt2",
15
+ "n_ctx": 1024,
16
+ "n_embd": 312,
17
+ "n_head": 12,
18
+ "n_inner": null,
19
+ "n_layer": 4,
20
+ "n_positions": 1024,
21
+ "reorder_and_upcast_attn": false,
22
+ "resid_pdrop": 0.1,
23
+ "scale_attn_by_inverse_layer_idx": false,
24
+ "scale_attn_weights": true,
25
+ "summary_activation": null,
26
+ "summary_first_dropout": 0.1,
27
+ "summary_proj_to_labels": true,
28
+ "summary_type": "cls_index",
29
+ "summary_use_proj": true,
30
+ "task_specific_params": {
31
+ "text-generation": {
32
+ "do_sample": true,
33
+ "max_length": 50
34
+ }
35
+ },
36
+ "tokenizer_class": "BertTokenizerFast",
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.26.0.dev0",
39
+ "use_cache": true,
40
+ "vocab_size": 21253
41
+ }
eval_results.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 36.08,
3
+ "eval_accuracy": 0.37381769930940056,
4
+ "eval_loss": 3.296532154083252,
5
+ "eval_runtime": 116.1554,
6
+ "eval_samples": 26482,
7
+ "eval_samples_per_second": 227.988,
8
+ "eval_steps_per_second": 5.708,
9
+ "perplexity": 27.018779305888142
10
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48150ff9d1f73d4addaac5e7e67b037512c373687a4e807c338273b8a5c79d82
3
+ size 50771749
special_tokens_map.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "mask_token": "[MASK]",
4
+ "pad_token": "[PAD]",
5
+ "sep_token": "[SEP]",
6
+ "unk_token": "[UNK]"
7
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "do_lower_case": false,
4
+ "mask_token": "[MASK]",
5
+ "model_max_length": 512,
6
+ "name_or_path": "jed351/gpt2-tiny-zh-hk",
7
+ "pad_token": "[PAD]",
8
+ "sep_token": "[SEP]",
9
+ "special_tokens_map_file": null,
10
+ "strip_accents": null,
11
+ "tokenize_chinese_chars": true,
12
+ "tokenizer_class": "BertTokenizer",
13
+ "unk_token": "[UNK]"
14
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 36.08,
3
+ "train_loss": 3.451316015625,
4
+ "train_runtime": 66733.7733,
5
+ "train_samples": 443474,
6
+ "train_samples_per_second": 239.759,
7
+ "train_steps_per_second": 5.994
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,4825 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 36.078289889059256,
5
+ "global_step": 400000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.05,
12
+ "learning_rate": 4.9937625e-05,
13
+ "loss": 4.4121,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.09,
18
+ "learning_rate": 4.9875125e-05,
19
+ "loss": 4.1777,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.14,
24
+ "learning_rate": 4.9812625e-05,
25
+ "loss": 4.0899,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.18,
30
+ "learning_rate": 4.9750125000000005e-05,
31
+ "loss": 4.029,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.23,
36
+ "learning_rate": 4.9687625000000004e-05,
37
+ "loss": 3.9892,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.27,
42
+ "learning_rate": 4.9625125e-05,
43
+ "loss": 3.9563,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.32,
48
+ "learning_rate": 4.9562625e-05,
49
+ "loss": 3.9319,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.36,
54
+ "learning_rate": 4.9500125000000006e-05,
55
+ "loss": 3.9077,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.41,
60
+ "learning_rate": 4.9437625000000004e-05,
61
+ "loss": 3.8883,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.45,
66
+ "learning_rate": 4.9375125e-05,
67
+ "loss": 3.8724,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.5,
72
+ "learning_rate": 4.9312625e-05,
73
+ "loss": 3.8566,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.54,
78
+ "learning_rate": 4.9250125e-05,
79
+ "loss": 3.8431,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.59,
84
+ "learning_rate": 4.9187625000000005e-05,
85
+ "loss": 3.8284,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.63,
90
+ "learning_rate": 4.9125125e-05,
91
+ "loss": 3.8168,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.68,
96
+ "learning_rate": 4.9062625e-05,
97
+ "loss": 3.8098,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.72,
102
+ "learning_rate": 4.900025e-05,
103
+ "loss": 3.7967,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.77,
108
+ "learning_rate": 4.893775000000001e-05,
109
+ "loss": 3.7899,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.81,
114
+ "learning_rate": 4.887525e-05,
115
+ "loss": 3.7773,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.86,
120
+ "learning_rate": 4.8812750000000005e-05,
121
+ "loss": 3.7716,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.9,
126
+ "learning_rate": 4.8750375000000005e-05,
127
+ "loss": 3.7606,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.95,
132
+ "learning_rate": 4.8687875000000004e-05,
133
+ "loss": 3.755,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.99,
138
+ "learning_rate": 4.8625375e-05,
139
+ "loss": 3.7495,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 1.04,
144
+ "learning_rate": 4.8562875e-05,
145
+ "loss": 3.7424,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 1.08,
150
+ "learning_rate": 4.8500375000000006e-05,
151
+ "loss": 3.73,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 1.13,
156
+ "learning_rate": 4.8438e-05,
157
+ "loss": 3.7284,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 1.17,
162
+ "learning_rate": 4.8375500000000005e-05,
163
+ "loss": 3.7267,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 1.22,
168
+ "learning_rate": 4.8313e-05,
169
+ "loss": 3.7213,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 1.26,
174
+ "learning_rate": 4.82505e-05,
175
+ "loss": 3.7122,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 1.31,
180
+ "learning_rate": 4.8188125e-05,
181
+ "loss": 3.7055,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 1.35,
186
+ "learning_rate": 4.8125625e-05,
187
+ "loss": 3.7036,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 1.4,
192
+ "learning_rate": 4.8063125000000006e-05,
193
+ "loss": 3.7003,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 1.44,
198
+ "learning_rate": 4.8000625000000004e-05,
199
+ "loss": 3.6933,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 1.49,
204
+ "learning_rate": 4.7938125e-05,
205
+ "loss": 3.6898,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 1.53,
210
+ "learning_rate": 4.7875625e-05,
211
+ "loss": 3.6875,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 1.58,
216
+ "learning_rate": 4.781325e-05,
217
+ "loss": 3.6848,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 1.62,
222
+ "learning_rate": 4.775075e-05,
223
+ "loss": 3.6827,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 1.67,
228
+ "learning_rate": 4.768825e-05,
229
+ "loss": 3.6729,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 1.71,
234
+ "learning_rate": 4.7625750000000004e-05,
235
+ "loss": 3.6718,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 1.76,
240
+ "learning_rate": 4.7563375000000004e-05,
241
+ "loss": 3.6674,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 1.8,
246
+ "learning_rate": 4.7500875e-05,
247
+ "loss": 3.6626,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 1.85,
252
+ "learning_rate": 4.7438375e-05,
253
+ "loss": 3.659,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 1.89,
258
+ "learning_rate": 4.7375875e-05,
259
+ "loss": 3.6569,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 1.94,
264
+ "learning_rate": 4.7313375000000005e-05,
265
+ "loss": 3.6508,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 1.98,
270
+ "learning_rate": 4.7250875e-05,
271
+ "loss": 3.6536,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 2.03,
276
+ "learning_rate": 4.7188375e-05,
277
+ "loss": 3.6467,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 2.07,
282
+ "learning_rate": 4.7125875e-05,
283
+ "loss": 3.6441,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 2.12,
288
+ "learning_rate": 4.7063375000000006e-05,
289
+ "loss": 3.6382,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 2.16,
294
+ "learning_rate": 4.7000875e-05,
295
+ "loss": 3.6401,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 2.21,
300
+ "learning_rate": 4.6938500000000005e-05,
301
+ "loss": 3.6339,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 2.25,
306
+ "learning_rate": 4.6876125000000005e-05,
307
+ "loss": 3.633,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 2.3,
312
+ "learning_rate": 4.6813625e-05,
313
+ "loss": 3.6301,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 2.35,
318
+ "learning_rate": 4.6751125e-05,
319
+ "loss": 3.6277,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 2.39,
324
+ "learning_rate": 4.6688625e-05,
325
+ "loss": 3.6243,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 2.44,
330
+ "learning_rate": 4.6626125000000006e-05,
331
+ "loss": 3.6222,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 2.48,
336
+ "learning_rate": 4.6563625e-05,
337
+ "loss": 3.6189,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 2.53,
342
+ "learning_rate": 4.6501125e-05,
343
+ "loss": 3.6232,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 2.57,
348
+ "learning_rate": 4.6438625e-05,
349
+ "loss": 3.6158,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 2.62,
354
+ "learning_rate": 4.6376125000000007e-05,
355
+ "loss": 3.6137,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 2.66,
360
+ "learning_rate": 4.6313625000000005e-05,
361
+ "loss": 3.6101,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 2.71,
366
+ "learning_rate": 4.625125e-05,
367
+ "loss": 3.6095,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 2.75,
372
+ "learning_rate": 4.6188750000000004e-05,
373
+ "loss": 3.6098,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 2.8,
378
+ "learning_rate": 4.612625e-05,
379
+ "loss": 3.6091,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 2.84,
384
+ "learning_rate": 4.606375e-05,
385
+ "loss": 3.6047,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 2.89,
390
+ "learning_rate": 4.6001375e-05,
391
+ "loss": 3.603,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 2.93,
396
+ "learning_rate": 4.593887500000001e-05,
397
+ "loss": 3.6038,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 2.98,
402
+ "learning_rate": 4.5876375000000005e-05,
403
+ "loss": 3.592,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 3.02,
408
+ "learning_rate": 4.5813875000000004e-05,
409
+ "loss": 3.5909,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 3.07,
414
+ "learning_rate": 4.5751375e-05,
415
+ "loss": 3.5902,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 3.11,
420
+ "learning_rate": 4.5688875e-05,
421
+ "loss": 3.5931,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 3.16,
426
+ "learning_rate": 4.56265e-05,
427
+ "loss": 3.5901,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 3.2,
432
+ "learning_rate": 4.5564e-05,
433
+ "loss": 3.5858,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 3.25,
438
+ "learning_rate": 4.5501500000000005e-05,
439
+ "loss": 3.5849,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 3.29,
444
+ "learning_rate": 4.5439e-05,
445
+ "loss": 3.5811,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 3.34,
450
+ "learning_rate": 4.53765e-05,
451
+ "loss": 3.5823,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 3.38,
456
+ "learning_rate": 4.5314e-05,
457
+ "loss": 3.5768,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 3.43,
462
+ "learning_rate": 4.5251625e-05,
463
+ "loss": 3.5796,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 3.47,
468
+ "learning_rate": 4.5189125000000006e-05,
469
+ "loss": 3.581,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 3.52,
474
+ "learning_rate": 4.5126625e-05,
475
+ "loss": 3.5798,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 3.56,
480
+ "learning_rate": 4.5064125e-05,
481
+ "loss": 3.5744,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 3.61,
486
+ "learning_rate": 4.5001625e-05,
487
+ "loss": 3.5701,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 3.65,
492
+ "learning_rate": 4.4939125000000007e-05,
493
+ "loss": 3.5754,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 3.7,
498
+ "learning_rate": 4.487675e-05,
499
+ "loss": 3.5716,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 3.74,
504
+ "learning_rate": 4.481425e-05,
505
+ "loss": 3.5696,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 3.79,
510
+ "learning_rate": 4.4751750000000004e-05,
511
+ "loss": 3.5696,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 3.83,
516
+ "learning_rate": 4.468925e-05,
517
+ "loss": 3.565,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 3.88,
522
+ "learning_rate": 4.462675e-05,
523
+ "loss": 3.5641,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 3.92,
528
+ "learning_rate": 4.456425e-05,
529
+ "loss": 3.5667,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 3.97,
534
+ "learning_rate": 4.4501750000000005e-05,
535
+ "loss": 3.5658,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 4.01,
540
+ "learning_rate": 4.443925e-05,
541
+ "loss": 3.5593,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 4.06,
546
+ "learning_rate": 4.4376875000000004e-05,
547
+ "loss": 3.5572,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 4.1,
552
+ "learning_rate": 4.4314500000000004e-05,
553
+ "loss": 3.5618,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 4.15,
558
+ "learning_rate": 4.4252e-05,
559
+ "loss": 3.5545,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 4.19,
564
+ "learning_rate": 4.41895e-05,
565
+ "loss": 3.5556,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 4.24,
570
+ "learning_rate": 4.4127e-05,
571
+ "loss": 3.5497,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 4.28,
576
+ "learning_rate": 4.406462500000001e-05,
577
+ "loss": 3.5504,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 4.33,
582
+ "learning_rate": 4.4002125e-05,
583
+ "loss": 3.5517,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 4.37,
588
+ "learning_rate": 4.3939625000000004e-05,
589
+ "loss": 3.5524,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 4.42,
594
+ "learning_rate": 4.3877125e-05,
595
+ "loss": 3.5483,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 4.46,
600
+ "learning_rate": 4.3814625e-05,
601
+ "loss": 3.5497,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 4.51,
606
+ "learning_rate": 4.375225e-05,
607
+ "loss": 3.5479,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 4.55,
612
+ "learning_rate": 4.368975e-05,
613
+ "loss": 3.5457,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 4.6,
618
+ "learning_rate": 4.3627250000000005e-05,
619
+ "loss": 3.5427,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 4.65,
624
+ "learning_rate": 4.356475e-05,
625
+ "loss": 3.5399,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 4.69,
630
+ "learning_rate": 4.3502375000000004e-05,
631
+ "loss": 3.5412,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 4.74,
636
+ "learning_rate": 4.3439875e-05,
637
+ "loss": 3.5412,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 4.78,
642
+ "learning_rate": 4.3377375e-05,
643
+ "loss": 3.5444,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 4.83,
648
+ "learning_rate": 4.3314875000000006e-05,
649
+ "loss": 3.5404,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 4.87,
654
+ "learning_rate": 4.32525e-05,
655
+ "loss": 3.5368,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 4.92,
660
+ "learning_rate": 4.3190000000000005e-05,
661
+ "loss": 3.5374,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 4.96,
666
+ "learning_rate": 4.3127500000000003e-05,
667
+ "loss": 3.5409,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 5.01,
672
+ "learning_rate": 4.3065e-05,
673
+ "loss": 3.533,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 5.05,
678
+ "learning_rate": 4.30025e-05,
679
+ "loss": 3.5265,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 5.1,
684
+ "learning_rate": 4.2940125e-05,
685
+ "loss": 3.53,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 5.14,
690
+ "learning_rate": 4.2877625000000006e-05,
691
+ "loss": 3.5315,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 5.19,
696
+ "learning_rate": 4.2815125e-05,
697
+ "loss": 3.531,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 5.23,
702
+ "learning_rate": 4.2752625e-05,
703
+ "loss": 3.5327,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 5.28,
708
+ "learning_rate": 4.2690125e-05,
709
+ "loss": 3.526,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 5.32,
714
+ "learning_rate": 4.262762500000001e-05,
715
+ "loss": 3.5263,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 5.37,
720
+ "learning_rate": 4.2565125e-05,
721
+ "loss": 3.5248,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 5.41,
726
+ "learning_rate": 4.250275e-05,
727
+ "loss": 3.5257,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 5.46,
732
+ "learning_rate": 4.2440250000000004e-05,
733
+ "loss": 3.5275,
734
+ "step": 60500
735
+ },
736
+ {
737
+ "epoch": 5.5,
738
+ "learning_rate": 4.237775e-05,
739
+ "loss": 3.5263,
740
+ "step": 61000
741
+ },
742
+ {
743
+ "epoch": 5.55,
744
+ "learning_rate": 4.231525e-05,
745
+ "loss": 3.5228,
746
+ "step": 61500
747
+ },
748
+ {
749
+ "epoch": 5.59,
750
+ "learning_rate": 4.225275e-05,
751
+ "loss": 3.5231,
752
+ "step": 62000
753
+ },
754
+ {
755
+ "epoch": 5.64,
756
+ "learning_rate": 4.219037500000001e-05,
757
+ "loss": 3.5251,
758
+ "step": 62500
759
+ },
760
+ {
761
+ "epoch": 5.68,
762
+ "learning_rate": 4.2127875e-05,
763
+ "loss": 3.5208,
764
+ "step": 63000
765
+ },
766
+ {
767
+ "epoch": 5.73,
768
+ "learning_rate": 4.2065375000000004e-05,
769
+ "loss": 3.5222,
770
+ "step": 63500
771
+ },
772
+ {
773
+ "epoch": 5.77,
774
+ "learning_rate": 4.2002875e-05,
775
+ "loss": 3.522,
776
+ "step": 64000
777
+ },
778
+ {
779
+ "epoch": 5.82,
780
+ "learning_rate": 4.1940499999999996e-05,
781
+ "loss": 3.5171,
782
+ "step": 64500
783
+ },
784
+ {
785
+ "epoch": 5.86,
786
+ "learning_rate": 4.1878e-05,
787
+ "loss": 3.52,
788
+ "step": 65000
789
+ },
790
+ {
791
+ "epoch": 5.91,
792
+ "learning_rate": 4.18155e-05,
793
+ "loss": 3.5124,
794
+ "step": 65500
795
+ },
796
+ {
797
+ "epoch": 5.95,
798
+ "learning_rate": 4.1753000000000005e-05,
799
+ "loss": 3.5136,
800
+ "step": 66000
801
+ },
802
+ {
803
+ "epoch": 6.0,
804
+ "learning_rate": 4.1690500000000003e-05,
805
+ "loss": 3.5168,
806
+ "step": 66500
807
+ },
808
+ {
809
+ "epoch": 6.04,
810
+ "learning_rate": 4.162825e-05,
811
+ "loss": 3.5105,
812
+ "step": 67000
813
+ },
814
+ {
815
+ "epoch": 6.09,
816
+ "learning_rate": 4.1565750000000004e-05,
817
+ "loss": 3.5102,
818
+ "step": 67500
819
+ },
820
+ {
821
+ "epoch": 6.13,
822
+ "learning_rate": 4.150325e-05,
823
+ "loss": 3.505,
824
+ "step": 68000
825
+ },
826
+ {
827
+ "epoch": 6.18,
828
+ "learning_rate": 4.144075e-05,
829
+ "loss": 3.5073,
830
+ "step": 68500
831
+ },
832
+ {
833
+ "epoch": 6.22,
834
+ "learning_rate": 4.137825e-05,
835
+ "loss": 3.5137,
836
+ "step": 69000
837
+ },
838
+ {
839
+ "epoch": 6.27,
840
+ "learning_rate": 4.1315750000000005e-05,
841
+ "loss": 3.5111,
842
+ "step": 69500
843
+ },
844
+ {
845
+ "epoch": 6.31,
846
+ "learning_rate": 4.1253250000000004e-05,
847
+ "loss": 3.513,
848
+ "step": 70000
849
+ },
850
+ {
851
+ "epoch": 6.36,
852
+ "learning_rate": 4.119075e-05,
853
+ "loss": 3.5095,
854
+ "step": 70500
855
+ },
856
+ {
857
+ "epoch": 6.4,
858
+ "learning_rate": 4.112825e-05,
859
+ "loss": 3.5081,
860
+ "step": 71000
861
+ },
862
+ {
863
+ "epoch": 6.45,
864
+ "learning_rate": 4.1065875e-05,
865
+ "loss": 3.5059,
866
+ "step": 71500
867
+ },
868
+ {
869
+ "epoch": 6.49,
870
+ "learning_rate": 4.1003375e-05,
871
+ "loss": 3.5076,
872
+ "step": 72000
873
+ },
874
+ {
875
+ "epoch": 6.54,
876
+ "learning_rate": 4.0940875e-05,
877
+ "loss": 3.5035,
878
+ "step": 72500
879
+ },
880
+ {
881
+ "epoch": 6.58,
882
+ "learning_rate": 4.0878375e-05,
883
+ "loss": 3.5087,
884
+ "step": 73000
885
+ },
886
+ {
887
+ "epoch": 6.63,
888
+ "learning_rate": 4.0816000000000004e-05,
889
+ "loss": 3.5044,
890
+ "step": 73500
891
+ },
892
+ {
893
+ "epoch": 6.67,
894
+ "learning_rate": 4.07535e-05,
895
+ "loss": 3.503,
896
+ "step": 74000
897
+ },
898
+ {
899
+ "epoch": 6.72,
900
+ "learning_rate": 4.0691e-05,
901
+ "loss": 3.5033,
902
+ "step": 74500
903
+ },
904
+ {
905
+ "epoch": 6.76,
906
+ "learning_rate": 4.0628500000000006e-05,
907
+ "loss": 3.5035,
908
+ "step": 75000
909
+ },
910
+ {
911
+ "epoch": 6.81,
912
+ "learning_rate": 4.0566000000000004e-05,
913
+ "loss": 3.5037,
914
+ "step": 75500
915
+ },
916
+ {
917
+ "epoch": 6.85,
918
+ "learning_rate": 4.05035e-05,
919
+ "loss": 3.4997,
920
+ "step": 76000
921
+ },
922
+ {
923
+ "epoch": 6.9,
924
+ "learning_rate": 4.0441e-05,
925
+ "loss": 3.4997,
926
+ "step": 76500
927
+ },
928
+ {
929
+ "epoch": 6.95,
930
+ "learning_rate": 4.03785e-05,
931
+ "loss": 3.4969,
932
+ "step": 77000
933
+ },
934
+ {
935
+ "epoch": 6.99,
936
+ "learning_rate": 4.031612500000001e-05,
937
+ "loss": 3.4987,
938
+ "step": 77500
939
+ },
940
+ {
941
+ "epoch": 7.04,
942
+ "learning_rate": 4.0253625e-05,
943
+ "loss": 3.5014,
944
+ "step": 78000
945
+ },
946
+ {
947
+ "epoch": 7.08,
948
+ "learning_rate": 4.0191125000000004e-05,
949
+ "loss": 3.4939,
950
+ "step": 78500
951
+ },
952
+ {
953
+ "epoch": 7.13,
954
+ "learning_rate": 4.0128625e-05,
955
+ "loss": 3.4961,
956
+ "step": 79000
957
+ },
958
+ {
959
+ "epoch": 7.17,
960
+ "learning_rate": 4.006625e-05,
961
+ "loss": 3.4964,
962
+ "step": 79500
963
+ },
964
+ {
965
+ "epoch": 7.22,
966
+ "learning_rate": 4.000375e-05,
967
+ "loss": 3.4956,
968
+ "step": 80000
969
+ },
970
+ {
971
+ "epoch": 7.26,
972
+ "learning_rate": 3.994125e-05,
973
+ "loss": 3.4967,
974
+ "step": 80500
975
+ },
976
+ {
977
+ "epoch": 7.31,
978
+ "learning_rate": 3.9878750000000005e-05,
979
+ "loss": 3.4964,
980
+ "step": 81000
981
+ },
982
+ {
983
+ "epoch": 7.35,
984
+ "learning_rate": 3.9816375e-05,
985
+ "loss": 3.4924,
986
+ "step": 81500
987
+ },
988
+ {
989
+ "epoch": 7.4,
990
+ "learning_rate": 3.9753875000000004e-05,
991
+ "loss": 3.4903,
992
+ "step": 82000
993
+ },
994
+ {
995
+ "epoch": 7.44,
996
+ "learning_rate": 3.9691375e-05,
997
+ "loss": 3.4869,
998
+ "step": 82500
999
+ },
1000
+ {
1001
+ "epoch": 7.49,
1002
+ "learning_rate": 3.9628875e-05,
1003
+ "loss": 3.4938,
1004
+ "step": 83000
1005
+ },
1006
+ {
1007
+ "epoch": 7.53,
1008
+ "learning_rate": 3.9566375e-05,
1009
+ "loss": 3.4892,
1010
+ "step": 83500
1011
+ },
1012
+ {
1013
+ "epoch": 7.58,
1014
+ "learning_rate": 3.9503875e-05,
1015
+ "loss": 3.4888,
1016
+ "step": 84000
1017
+ },
1018
+ {
1019
+ "epoch": 7.62,
1020
+ "learning_rate": 3.9441500000000005e-05,
1021
+ "loss": 3.4882,
1022
+ "step": 84500
1023
+ },
1024
+ {
1025
+ "epoch": 7.67,
1026
+ "learning_rate": 3.9379e-05,
1027
+ "loss": 3.4885,
1028
+ "step": 85000
1029
+ },
1030
+ {
1031
+ "epoch": 7.71,
1032
+ "learning_rate": 3.93165e-05,
1033
+ "loss": 3.4893,
1034
+ "step": 85500
1035
+ },
1036
+ {
1037
+ "epoch": 7.76,
1038
+ "learning_rate": 3.9254e-05,
1039
+ "loss": 3.4895,
1040
+ "step": 86000
1041
+ },
1042
+ {
1043
+ "epoch": 7.8,
1044
+ "learning_rate": 3.9191500000000006e-05,
1045
+ "loss": 3.4839,
1046
+ "step": 86500
1047
+ },
1048
+ {
1049
+ "epoch": 7.85,
1050
+ "learning_rate": 3.9129125e-05,
1051
+ "loss": 3.4906,
1052
+ "step": 87000
1053
+ },
1054
+ {
1055
+ "epoch": 7.89,
1056
+ "learning_rate": 3.9066625e-05,
1057
+ "loss": 3.4846,
1058
+ "step": 87500
1059
+ },
1060
+ {
1061
+ "epoch": 7.94,
1062
+ "learning_rate": 3.9004125e-05,
1063
+ "loss": 3.487,
1064
+ "step": 88000
1065
+ },
1066
+ {
1067
+ "epoch": 7.98,
1068
+ "learning_rate": 3.8941625e-05,
1069
+ "loss": 3.4836,
1070
+ "step": 88500
1071
+ },
1072
+ {
1073
+ "epoch": 8.03,
1074
+ "learning_rate": 3.8879125e-05,
1075
+ "loss": 3.4813,
1076
+ "step": 89000
1077
+ },
1078
+ {
1079
+ "epoch": 8.07,
1080
+ "learning_rate": 3.881675e-05,
1081
+ "loss": 3.4797,
1082
+ "step": 89500
1083
+ },
1084
+ {
1085
+ "epoch": 8.12,
1086
+ "learning_rate": 3.8754250000000006e-05,
1087
+ "loss": 3.4797,
1088
+ "step": 90000
1089
+ },
1090
+ {
1091
+ "epoch": 8.16,
1092
+ "learning_rate": 3.8691750000000004e-05,
1093
+ "loss": 3.4799,
1094
+ "step": 90500
1095
+ },
1096
+ {
1097
+ "epoch": 8.21,
1098
+ "learning_rate": 3.862925e-05,
1099
+ "loss": 3.4791,
1100
+ "step": 91000
1101
+ },
1102
+ {
1103
+ "epoch": 8.25,
1104
+ "learning_rate": 3.8566875e-05,
1105
+ "loss": 3.4776,
1106
+ "step": 91500
1107
+ },
1108
+ {
1109
+ "epoch": 8.3,
1110
+ "learning_rate": 3.8504375e-05,
1111
+ "loss": 3.4789,
1112
+ "step": 92000
1113
+ },
1114
+ {
1115
+ "epoch": 8.34,
1116
+ "learning_rate": 3.8441875e-05,
1117
+ "loss": 3.4818,
1118
+ "step": 92500
1119
+ },
1120
+ {
1121
+ "epoch": 8.39,
1122
+ "learning_rate": 3.8379375e-05,
1123
+ "loss": 3.477,
1124
+ "step": 93000
1125
+ },
1126
+ {
1127
+ "epoch": 8.43,
1128
+ "learning_rate": 3.8316875000000004e-05,
1129
+ "loss": 3.4787,
1130
+ "step": 93500
1131
+ },
1132
+ {
1133
+ "epoch": 8.48,
1134
+ "learning_rate": 3.8254500000000004e-05,
1135
+ "loss": 3.4792,
1136
+ "step": 94000
1137
+ },
1138
+ {
1139
+ "epoch": 8.52,
1140
+ "learning_rate": 3.8192e-05,
1141
+ "loss": 3.4806,
1142
+ "step": 94500
1143
+ },
1144
+ {
1145
+ "epoch": 8.57,
1146
+ "learning_rate": 3.81295e-05,
1147
+ "loss": 3.48,
1148
+ "step": 95000
1149
+ },
1150
+ {
1151
+ "epoch": 8.61,
1152
+ "learning_rate": 3.8067e-05,
1153
+ "loss": 3.4793,
1154
+ "step": 95500
1155
+ },
1156
+ {
1157
+ "epoch": 8.66,
1158
+ "learning_rate": 3.8004625e-05,
1159
+ "loss": 3.4767,
1160
+ "step": 96000
1161
+ },
1162
+ {
1163
+ "epoch": 8.7,
1164
+ "learning_rate": 3.7942125e-05,
1165
+ "loss": 3.4792,
1166
+ "step": 96500
1167
+ },
1168
+ {
1169
+ "epoch": 8.75,
1170
+ "learning_rate": 3.7879625000000004e-05,
1171
+ "loss": 3.4751,
1172
+ "step": 97000
1173
+ },
1174
+ {
1175
+ "epoch": 8.79,
1176
+ "learning_rate": 3.7817125e-05,
1177
+ "loss": 3.4747,
1178
+ "step": 97500
1179
+ },
1180
+ {
1181
+ "epoch": 8.84,
1182
+ "learning_rate": 3.775475e-05,
1183
+ "loss": 3.4771,
1184
+ "step": 98000
1185
+ },
1186
+ {
1187
+ "epoch": 8.88,
1188
+ "learning_rate": 3.769225e-05,
1189
+ "loss": 3.4755,
1190
+ "step": 98500
1191
+ },
1192
+ {
1193
+ "epoch": 8.93,
1194
+ "learning_rate": 3.762975e-05,
1195
+ "loss": 3.4746,
1196
+ "step": 99000
1197
+ },
1198
+ {
1199
+ "epoch": 8.97,
1200
+ "learning_rate": 3.7567250000000005e-05,
1201
+ "loss": 3.4749,
1202
+ "step": 99500
1203
+ },
1204
+ {
1205
+ "epoch": 9.02,
1206
+ "learning_rate": 3.7504875e-05,
1207
+ "loss": 3.4744,
1208
+ "step": 100000
1209
+ },
1210
+ {
1211
+ "epoch": 9.06,
1212
+ "learning_rate": 3.7442375000000004e-05,
1213
+ "loss": 3.4658,
1214
+ "step": 100500
1215
+ },
1216
+ {
1217
+ "epoch": 9.11,
1218
+ "learning_rate": 3.7379875e-05,
1219
+ "loss": 3.4723,
1220
+ "step": 101000
1221
+ },
1222
+ {
1223
+ "epoch": 9.15,
1224
+ "learning_rate": 3.7317499999999996e-05,
1225
+ "loss": 3.4672,
1226
+ "step": 101500
1227
+ },
1228
+ {
1229
+ "epoch": 9.2,
1230
+ "learning_rate": 3.7255e-05,
1231
+ "loss": 3.4716,
1232
+ "step": 102000
1233
+ },
1234
+ {
1235
+ "epoch": 9.25,
1236
+ "learning_rate": 3.71925e-05,
1237
+ "loss": 3.4695,
1238
+ "step": 102500
1239
+ },
1240
+ {
1241
+ "epoch": 9.29,
1242
+ "learning_rate": 3.7130000000000005e-05,
1243
+ "loss": 3.474,
1244
+ "step": 103000
1245
+ },
1246
+ {
1247
+ "epoch": 9.34,
1248
+ "learning_rate": 3.70675e-05,
1249
+ "loss": 3.4701,
1250
+ "step": 103500
1251
+ },
1252
+ {
1253
+ "epoch": 9.38,
1254
+ "learning_rate": 3.7005e-05,
1255
+ "loss": 3.4665,
1256
+ "step": 104000
1257
+ },
1258
+ {
1259
+ "epoch": 9.43,
1260
+ "learning_rate": 3.69425e-05,
1261
+ "loss": 3.468,
1262
+ "step": 104500
1263
+ },
1264
+ {
1265
+ "epoch": 9.47,
1266
+ "learning_rate": 3.6880000000000006e-05,
1267
+ "loss": 3.4654,
1268
+ "step": 105000
1269
+ },
1270
+ {
1271
+ "epoch": 9.52,
1272
+ "learning_rate": 3.68175e-05,
1273
+ "loss": 3.4674,
1274
+ "step": 105500
1275
+ },
1276
+ {
1277
+ "epoch": 9.56,
1278
+ "learning_rate": 3.6755125e-05,
1279
+ "loss": 3.4652,
1280
+ "step": 106000
1281
+ },
1282
+ {
1283
+ "epoch": 9.61,
1284
+ "learning_rate": 3.6692625000000003e-05,
1285
+ "loss": 3.4673,
1286
+ "step": 106500
1287
+ },
1288
+ {
1289
+ "epoch": 9.65,
1290
+ "learning_rate": 3.6630125e-05,
1291
+ "loss": 3.464,
1292
+ "step": 107000
1293
+ },
1294
+ {
1295
+ "epoch": 9.7,
1296
+ "learning_rate": 3.6567625e-05,
1297
+ "loss": 3.4656,
1298
+ "step": 107500
1299
+ },
1300
+ {
1301
+ "epoch": 9.74,
1302
+ "learning_rate": 3.6505125e-05,
1303
+ "loss": 3.4704,
1304
+ "step": 108000
1305
+ },
1306
+ {
1307
+ "epoch": 9.79,
1308
+ "learning_rate": 3.6442750000000006e-05,
1309
+ "loss": 3.4671,
1310
+ "step": 108500
1311
+ },
1312
+ {
1313
+ "epoch": 9.83,
1314
+ "learning_rate": 3.638025e-05,
1315
+ "loss": 3.462,
1316
+ "step": 109000
1317
+ },
1318
+ {
1319
+ "epoch": 9.88,
1320
+ "learning_rate": 3.631775e-05,
1321
+ "loss": 3.4675,
1322
+ "step": 109500
1323
+ },
1324
+ {
1325
+ "epoch": 9.92,
1326
+ "learning_rate": 3.625525e-05,
1327
+ "loss": 3.4658,
1328
+ "step": 110000
1329
+ },
1330
+ {
1331
+ "epoch": 9.97,
1332
+ "learning_rate": 3.6192875e-05,
1333
+ "loss": 3.4662,
1334
+ "step": 110500
1335
+ },
1336
+ {
1337
+ "epoch": 10.01,
1338
+ "learning_rate": 3.6130375e-05,
1339
+ "loss": 3.462,
1340
+ "step": 111000
1341
+ },
1342
+ {
1343
+ "epoch": 10.06,
1344
+ "learning_rate": 3.6068e-05,
1345
+ "loss": 3.4601,
1346
+ "step": 111500
1347
+ },
1348
+ {
1349
+ "epoch": 10.1,
1350
+ "learning_rate": 3.6005500000000006e-05,
1351
+ "loss": 3.4581,
1352
+ "step": 112000
1353
+ },
1354
+ {
1355
+ "epoch": 10.15,
1356
+ "learning_rate": 3.5943000000000005e-05,
1357
+ "loss": 3.4608,
1358
+ "step": 112500
1359
+ },
1360
+ {
1361
+ "epoch": 10.19,
1362
+ "learning_rate": 3.58805e-05,
1363
+ "loss": 3.4583,
1364
+ "step": 113000
1365
+ },
1366
+ {
1367
+ "epoch": 10.24,
1368
+ "learning_rate": 3.5818e-05,
1369
+ "loss": 3.4628,
1370
+ "step": 113500
1371
+ },
1372
+ {
1373
+ "epoch": 10.28,
1374
+ "learning_rate": 3.5755625e-05,
1375
+ "loss": 3.4592,
1376
+ "step": 114000
1377
+ },
1378
+ {
1379
+ "epoch": 10.33,
1380
+ "learning_rate": 3.5693125e-05,
1381
+ "loss": 3.4598,
1382
+ "step": 114500
1383
+ },
1384
+ {
1385
+ "epoch": 10.37,
1386
+ "learning_rate": 3.5630625e-05,
1387
+ "loss": 3.4596,
1388
+ "step": 115000
1389
+ },
1390
+ {
1391
+ "epoch": 10.42,
1392
+ "learning_rate": 3.5568125000000004e-05,
1393
+ "loss": 3.4609,
1394
+ "step": 115500
1395
+ },
1396
+ {
1397
+ "epoch": 10.46,
1398
+ "learning_rate": 3.5505625e-05,
1399
+ "loss": 3.4563,
1400
+ "step": 116000
1401
+ },
1402
+ {
1403
+ "epoch": 10.51,
1404
+ "learning_rate": 3.5443125e-05,
1405
+ "loss": 3.4563,
1406
+ "step": 116500
1407
+ },
1408
+ {
1409
+ "epoch": 10.55,
1410
+ "learning_rate": 3.5380625e-05,
1411
+ "loss": 3.4557,
1412
+ "step": 117000
1413
+ },
1414
+ {
1415
+ "epoch": 10.6,
1416
+ "learning_rate": 3.5318125000000005e-05,
1417
+ "loss": 3.458,
1418
+ "step": 117500
1419
+ },
1420
+ {
1421
+ "epoch": 10.64,
1422
+ "learning_rate": 3.5255750000000005e-05,
1423
+ "loss": 3.4586,
1424
+ "step": 118000
1425
+ },
1426
+ {
1427
+ "epoch": 10.69,
1428
+ "learning_rate": 3.519325e-05,
1429
+ "loss": 3.4586,
1430
+ "step": 118500
1431
+ },
1432
+ {
1433
+ "epoch": 10.73,
1434
+ "learning_rate": 3.513075e-05,
1435
+ "loss": 3.4584,
1436
+ "step": 119000
1437
+ },
1438
+ {
1439
+ "epoch": 10.78,
1440
+ "learning_rate": 3.506825e-05,
1441
+ "loss": 3.4515,
1442
+ "step": 119500
1443
+ },
1444
+ {
1445
+ "epoch": 10.82,
1446
+ "learning_rate": 3.5005875e-05,
1447
+ "loss": 3.4551,
1448
+ "step": 120000
1449
+ },
1450
+ {
1451
+ "epoch": 10.87,
1452
+ "learning_rate": 3.4943375e-05,
1453
+ "loss": 3.4565,
1454
+ "step": 120500
1455
+ },
1456
+ {
1457
+ "epoch": 10.91,
1458
+ "learning_rate": 3.4880875000000005e-05,
1459
+ "loss": 3.4596,
1460
+ "step": 121000
1461
+ },
1462
+ {
1463
+ "epoch": 10.96,
1464
+ "learning_rate": 3.4818375000000004e-05,
1465
+ "loss": 3.4562,
1466
+ "step": 121500
1467
+ },
1468
+ {
1469
+ "epoch": 11.0,
1470
+ "learning_rate": 3.4756e-05,
1471
+ "loss": 3.4551,
1472
+ "step": 122000
1473
+ },
1474
+ {
1475
+ "epoch": 11.05,
1476
+ "learning_rate": 3.4693625000000004e-05,
1477
+ "loss": 3.4466,
1478
+ "step": 122500
1479
+ },
1480
+ {
1481
+ "epoch": 11.09,
1482
+ "learning_rate": 3.4631125e-05,
1483
+ "loss": 3.4494,
1484
+ "step": 123000
1485
+ },
1486
+ {
1487
+ "epoch": 11.14,
1488
+ "learning_rate": 3.4568625e-05,
1489
+ "loss": 3.4485,
1490
+ "step": 123500
1491
+ },
1492
+ {
1493
+ "epoch": 11.18,
1494
+ "learning_rate": 3.4506125e-05,
1495
+ "loss": 3.4511,
1496
+ "step": 124000
1497
+ },
1498
+ {
1499
+ "epoch": 11.23,
1500
+ "learning_rate": 3.444375e-05,
1501
+ "loss": 3.4521,
1502
+ "step": 124500
1503
+ },
1504
+ {
1505
+ "epoch": 11.27,
1506
+ "learning_rate": 3.4381250000000006e-05,
1507
+ "loss": 3.4494,
1508
+ "step": 125000
1509
+ },
1510
+ {
1511
+ "epoch": 11.32,
1512
+ "learning_rate": 3.431875e-05,
1513
+ "loss": 3.4515,
1514
+ "step": 125500
1515
+ },
1516
+ {
1517
+ "epoch": 11.36,
1518
+ "learning_rate": 3.425625e-05,
1519
+ "loss": 3.4499,
1520
+ "step": 126000
1521
+ },
1522
+ {
1523
+ "epoch": 11.41,
1524
+ "learning_rate": 3.419375e-05,
1525
+ "loss": 3.4532,
1526
+ "step": 126500
1527
+ },
1528
+ {
1529
+ "epoch": 11.45,
1530
+ "learning_rate": 3.4131250000000006e-05,
1531
+ "loss": 3.4527,
1532
+ "step": 127000
1533
+ },
1534
+ {
1535
+ "epoch": 11.5,
1536
+ "learning_rate": 3.406875e-05,
1537
+ "loss": 3.4505,
1538
+ "step": 127500
1539
+ },
1540
+ {
1541
+ "epoch": 11.55,
1542
+ "learning_rate": 3.400625e-05,
1543
+ "loss": 3.4496,
1544
+ "step": 128000
1545
+ },
1546
+ {
1547
+ "epoch": 11.59,
1548
+ "learning_rate": 3.3943875000000004e-05,
1549
+ "loss": 3.4526,
1550
+ "step": 128500
1551
+ },
1552
+ {
1553
+ "epoch": 11.64,
1554
+ "learning_rate": 3.3881375e-05,
1555
+ "loss": 3.4503,
1556
+ "step": 129000
1557
+ },
1558
+ {
1559
+ "epoch": 11.68,
1560
+ "learning_rate": 3.3818875e-05,
1561
+ "loss": 3.4498,
1562
+ "step": 129500
1563
+ },
1564
+ {
1565
+ "epoch": 11.73,
1566
+ "learning_rate": 3.3756375e-05,
1567
+ "loss": 3.4492,
1568
+ "step": 130000
1569
+ },
1570
+ {
1571
+ "epoch": 11.77,
1572
+ "learning_rate": 3.3694000000000006e-05,
1573
+ "loss": 3.4471,
1574
+ "step": 130500
1575
+ },
1576
+ {
1577
+ "epoch": 11.82,
1578
+ "learning_rate": 3.36315e-05,
1579
+ "loss": 3.4504,
1580
+ "step": 131000
1581
+ },
1582
+ {
1583
+ "epoch": 11.86,
1584
+ "learning_rate": 3.3569e-05,
1585
+ "loss": 3.449,
1586
+ "step": 131500
1587
+ },
1588
+ {
1589
+ "epoch": 11.91,
1590
+ "learning_rate": 3.35065e-05,
1591
+ "loss": 3.4478,
1592
+ "step": 132000
1593
+ },
1594
+ {
1595
+ "epoch": 11.95,
1596
+ "learning_rate": 3.3444125e-05,
1597
+ "loss": 3.4483,
1598
+ "step": 132500
1599
+ },
1600
+ {
1601
+ "epoch": 12.0,
1602
+ "learning_rate": 3.3381625e-05,
1603
+ "loss": 3.4496,
1604
+ "step": 133000
1605
+ },
1606
+ {
1607
+ "epoch": 12.04,
1608
+ "learning_rate": 3.3319125e-05,
1609
+ "loss": 3.4443,
1610
+ "step": 133500
1611
+ },
1612
+ {
1613
+ "epoch": 12.09,
1614
+ "learning_rate": 3.3256625000000004e-05,
1615
+ "loss": 3.4451,
1616
+ "step": 134000
1617
+ },
1618
+ {
1619
+ "epoch": 12.13,
1620
+ "learning_rate": 3.3194125e-05,
1621
+ "loss": 3.4406,
1622
+ "step": 134500
1623
+ },
1624
+ {
1625
+ "epoch": 12.18,
1626
+ "learning_rate": 3.3131750000000003e-05,
1627
+ "loss": 3.4448,
1628
+ "step": 135000
1629
+ },
1630
+ {
1631
+ "epoch": 12.22,
1632
+ "learning_rate": 3.306925e-05,
1633
+ "loss": 3.4466,
1634
+ "step": 135500
1635
+ },
1636
+ {
1637
+ "epoch": 12.27,
1638
+ "learning_rate": 3.300675e-05,
1639
+ "loss": 3.4472,
1640
+ "step": 136000
1641
+ },
1642
+ {
1643
+ "epoch": 12.31,
1644
+ "learning_rate": 3.2944375e-05,
1645
+ "loss": 3.4445,
1646
+ "step": 136500
1647
+ },
1648
+ {
1649
+ "epoch": 12.36,
1650
+ "learning_rate": 3.2881875e-05,
1651
+ "loss": 3.4465,
1652
+ "step": 137000
1653
+ },
1654
+ {
1655
+ "epoch": 12.4,
1656
+ "learning_rate": 3.2819375000000005e-05,
1657
+ "loss": 3.445,
1658
+ "step": 137500
1659
+ },
1660
+ {
1661
+ "epoch": 12.45,
1662
+ "learning_rate": 3.2756875e-05,
1663
+ "loss": 3.4443,
1664
+ "step": 138000
1665
+ },
1666
+ {
1667
+ "epoch": 12.49,
1668
+ "learning_rate": 3.2694375e-05,
1669
+ "loss": 3.4402,
1670
+ "step": 138500
1671
+ },
1672
+ {
1673
+ "epoch": 12.54,
1674
+ "learning_rate": 3.2631875e-05,
1675
+ "loss": 3.4419,
1676
+ "step": 139000
1677
+ },
1678
+ {
1679
+ "epoch": 12.58,
1680
+ "learning_rate": 3.2569375000000005e-05,
1681
+ "loss": 3.4387,
1682
+ "step": 139500
1683
+ },
1684
+ {
1685
+ "epoch": 12.63,
1686
+ "learning_rate": 3.2506875000000004e-05,
1687
+ "loss": 3.4432,
1688
+ "step": 140000
1689
+ },
1690
+ {
1691
+ "epoch": 12.67,
1692
+ "learning_rate": 3.24445e-05,
1693
+ "loss": 3.4392,
1694
+ "step": 140500
1695
+ },
1696
+ {
1697
+ "epoch": 12.72,
1698
+ "learning_rate": 3.2382e-05,
1699
+ "loss": 3.442,
1700
+ "step": 141000
1701
+ },
1702
+ {
1703
+ "epoch": 12.76,
1704
+ "learning_rate": 3.23195e-05,
1705
+ "loss": 3.4443,
1706
+ "step": 141500
1707
+ },
1708
+ {
1709
+ "epoch": 12.81,
1710
+ "learning_rate": 3.2257000000000006e-05,
1711
+ "loss": 3.4405,
1712
+ "step": 142000
1713
+ },
1714
+ {
1715
+ "epoch": 12.85,
1716
+ "learning_rate": 3.21945e-05,
1717
+ "loss": 3.4427,
1718
+ "step": 142500
1719
+ },
1720
+ {
1721
+ "epoch": 12.9,
1722
+ "learning_rate": 3.2132e-05,
1723
+ "loss": 3.4398,
1724
+ "step": 143000
1725
+ },
1726
+ {
1727
+ "epoch": 12.94,
1728
+ "learning_rate": 3.20695e-05,
1729
+ "loss": 3.4392,
1730
+ "step": 143500
1731
+ },
1732
+ {
1733
+ "epoch": 12.99,
1734
+ "learning_rate": 3.2007e-05,
1735
+ "loss": 3.4396,
1736
+ "step": 144000
1737
+ },
1738
+ {
1739
+ "epoch": 13.03,
1740
+ "learning_rate": 3.1944625e-05,
1741
+ "loss": 3.4359,
1742
+ "step": 144500
1743
+ },
1744
+ {
1745
+ "epoch": 13.08,
1746
+ "learning_rate": 3.1882125e-05,
1747
+ "loss": 3.4373,
1748
+ "step": 145000
1749
+ },
1750
+ {
1751
+ "epoch": 13.12,
1752
+ "learning_rate": 3.1819625000000005e-05,
1753
+ "loss": 3.4352,
1754
+ "step": 145500
1755
+ },
1756
+ {
1757
+ "epoch": 13.17,
1758
+ "learning_rate": 3.1757125e-05,
1759
+ "loss": 3.4361,
1760
+ "step": 146000
1761
+ },
1762
+ {
1763
+ "epoch": 13.21,
1764
+ "learning_rate": 3.1694625e-05,
1765
+ "loss": 3.4386,
1766
+ "step": 146500
1767
+ },
1768
+ {
1769
+ "epoch": 13.26,
1770
+ "learning_rate": 3.163225e-05,
1771
+ "loss": 3.4412,
1772
+ "step": 147000
1773
+ },
1774
+ {
1775
+ "epoch": 13.3,
1776
+ "learning_rate": 3.156975e-05,
1777
+ "loss": 3.4356,
1778
+ "step": 147500
1779
+ },
1780
+ {
1781
+ "epoch": 13.35,
1782
+ "learning_rate": 3.150725e-05,
1783
+ "loss": 3.4377,
1784
+ "step": 148000
1785
+ },
1786
+ {
1787
+ "epoch": 13.39,
1788
+ "learning_rate": 3.144475e-05,
1789
+ "loss": 3.4351,
1790
+ "step": 148500
1791
+ },
1792
+ {
1793
+ "epoch": 13.44,
1794
+ "learning_rate": 3.1382375000000005e-05,
1795
+ "loss": 3.4402,
1796
+ "step": 149000
1797
+ },
1798
+ {
1799
+ "epoch": 13.48,
1800
+ "learning_rate": 3.1319875e-05,
1801
+ "loss": 3.4362,
1802
+ "step": 149500
1803
+ },
1804
+ {
1805
+ "epoch": 13.53,
1806
+ "learning_rate": 3.1257375e-05,
1807
+ "loss": 3.4402,
1808
+ "step": 150000
1809
+ },
1810
+ {
1811
+ "epoch": 13.57,
1812
+ "learning_rate": 3.1194875e-05,
1813
+ "loss": 3.4366,
1814
+ "step": 150500
1815
+ },
1816
+ {
1817
+ "epoch": 13.62,
1818
+ "learning_rate": 3.11325e-05,
1819
+ "loss": 3.4356,
1820
+ "step": 151000
1821
+ },
1822
+ {
1823
+ "epoch": 13.66,
1824
+ "learning_rate": 3.107e-05,
1825
+ "loss": 3.4332,
1826
+ "step": 151500
1827
+ },
1828
+ {
1829
+ "epoch": 13.71,
1830
+ "learning_rate": 3.10075e-05,
1831
+ "loss": 3.4343,
1832
+ "step": 152000
1833
+ },
1834
+ {
1835
+ "epoch": 13.75,
1836
+ "learning_rate": 3.0945e-05,
1837
+ "loss": 3.4366,
1838
+ "step": 152500
1839
+ },
1840
+ {
1841
+ "epoch": 13.8,
1842
+ "learning_rate": 3.08825e-05,
1843
+ "loss": 3.436,
1844
+ "step": 153000
1845
+ },
1846
+ {
1847
+ "epoch": 13.85,
1848
+ "learning_rate": 3.0820125e-05,
1849
+ "loss": 3.4278,
1850
+ "step": 153500
1851
+ },
1852
+ {
1853
+ "epoch": 13.89,
1854
+ "learning_rate": 3.0757625e-05,
1855
+ "loss": 3.4351,
1856
+ "step": 154000
1857
+ },
1858
+ {
1859
+ "epoch": 13.94,
1860
+ "learning_rate": 3.0695125000000005e-05,
1861
+ "loss": 3.4341,
1862
+ "step": 154500
1863
+ },
1864
+ {
1865
+ "epoch": 13.98,
1866
+ "learning_rate": 3.0632625000000004e-05,
1867
+ "loss": 3.4362,
1868
+ "step": 155000
1869
+ },
1870
+ {
1871
+ "epoch": 14.03,
1872
+ "learning_rate": 3.057025e-05,
1873
+ "loss": 3.4335,
1874
+ "step": 155500
1875
+ },
1876
+ {
1877
+ "epoch": 14.07,
1878
+ "learning_rate": 3.0507750000000003e-05,
1879
+ "loss": 3.4289,
1880
+ "step": 156000
1881
+ },
1882
+ {
1883
+ "epoch": 14.12,
1884
+ "learning_rate": 3.044525e-05,
1885
+ "loss": 3.4283,
1886
+ "step": 156500
1887
+ },
1888
+ {
1889
+ "epoch": 14.16,
1890
+ "learning_rate": 3.0382875e-05,
1891
+ "loss": 3.4298,
1892
+ "step": 157000
1893
+ },
1894
+ {
1895
+ "epoch": 14.21,
1896
+ "learning_rate": 3.0320375e-05,
1897
+ "loss": 3.4257,
1898
+ "step": 157500
1899
+ },
1900
+ {
1901
+ "epoch": 14.25,
1902
+ "learning_rate": 3.0257875000000002e-05,
1903
+ "loss": 3.4315,
1904
+ "step": 158000
1905
+ },
1906
+ {
1907
+ "epoch": 14.3,
1908
+ "learning_rate": 3.0195375000000004e-05,
1909
+ "loss": 3.4319,
1910
+ "step": 158500
1911
+ },
1912
+ {
1913
+ "epoch": 14.34,
1914
+ "learning_rate": 3.0132875e-05,
1915
+ "loss": 3.4311,
1916
+ "step": 159000
1917
+ },
1918
+ {
1919
+ "epoch": 14.39,
1920
+ "learning_rate": 3.0070375e-05,
1921
+ "loss": 3.4312,
1922
+ "step": 159500
1923
+ },
1924
+ {
1925
+ "epoch": 14.43,
1926
+ "learning_rate": 3.0007875000000003e-05,
1927
+ "loss": 3.4293,
1928
+ "step": 160000
1929
+ },
1930
+ {
1931
+ "epoch": 14.48,
1932
+ "learning_rate": 2.9945375000000005e-05,
1933
+ "loss": 3.4291,
1934
+ "step": 160500
1935
+ },
1936
+ {
1937
+ "epoch": 14.52,
1938
+ "learning_rate": 2.9882875e-05,
1939
+ "loss": 3.4323,
1940
+ "step": 161000
1941
+ },
1942
+ {
1943
+ "epoch": 14.57,
1944
+ "learning_rate": 2.98205e-05,
1945
+ "loss": 3.4282,
1946
+ "step": 161500
1947
+ },
1948
+ {
1949
+ "epoch": 14.61,
1950
+ "learning_rate": 2.9758000000000002e-05,
1951
+ "loss": 3.4317,
1952
+ "step": 162000
1953
+ },
1954
+ {
1955
+ "epoch": 14.66,
1956
+ "learning_rate": 2.9695500000000004e-05,
1957
+ "loss": 3.4295,
1958
+ "step": 162500
1959
+ },
1960
+ {
1961
+ "epoch": 14.7,
1962
+ "learning_rate": 2.9633e-05,
1963
+ "loss": 3.429,
1964
+ "step": 163000
1965
+ },
1966
+ {
1967
+ "epoch": 14.75,
1968
+ "learning_rate": 2.9570625000000003e-05,
1969
+ "loss": 3.4297,
1970
+ "step": 163500
1971
+ },
1972
+ {
1973
+ "epoch": 14.79,
1974
+ "learning_rate": 2.9508125000000005e-05,
1975
+ "loss": 3.4336,
1976
+ "step": 164000
1977
+ },
1978
+ {
1979
+ "epoch": 14.84,
1980
+ "learning_rate": 2.9445625e-05,
1981
+ "loss": 3.4315,
1982
+ "step": 164500
1983
+ },
1984
+ {
1985
+ "epoch": 14.88,
1986
+ "learning_rate": 2.9383125e-05,
1987
+ "loss": 3.4266,
1988
+ "step": 165000
1989
+ },
1990
+ {
1991
+ "epoch": 14.93,
1992
+ "learning_rate": 2.9320750000000002e-05,
1993
+ "loss": 3.4312,
1994
+ "step": 165500
1995
+ },
1996
+ {
1997
+ "epoch": 14.97,
1998
+ "learning_rate": 2.9258250000000004e-05,
1999
+ "loss": 3.4297,
2000
+ "step": 166000
2001
+ },
2002
+ {
2003
+ "epoch": 15.02,
2004
+ "learning_rate": 2.919575e-05,
2005
+ "loss": 3.4267,
2006
+ "step": 166500
2007
+ },
2008
+ {
2009
+ "epoch": 15.06,
2010
+ "learning_rate": 2.913325e-05,
2011
+ "loss": 3.4227,
2012
+ "step": 167000
2013
+ },
2014
+ {
2015
+ "epoch": 15.11,
2016
+ "learning_rate": 2.9070750000000003e-05,
2017
+ "loss": 3.4243,
2018
+ "step": 167500
2019
+ },
2020
+ {
2021
+ "epoch": 15.15,
2022
+ "learning_rate": 2.9008375e-05,
2023
+ "loss": 3.4238,
2024
+ "step": 168000
2025
+ },
2026
+ {
2027
+ "epoch": 15.2,
2028
+ "learning_rate": 2.8945875000000002e-05,
2029
+ "loss": 3.4273,
2030
+ "step": 168500
2031
+ },
2032
+ {
2033
+ "epoch": 15.24,
2034
+ "learning_rate": 2.8883375e-05,
2035
+ "loss": 3.4242,
2036
+ "step": 169000
2037
+ },
2038
+ {
2039
+ "epoch": 15.29,
2040
+ "learning_rate": 2.8820875000000002e-05,
2041
+ "loss": 3.4284,
2042
+ "step": 169500
2043
+ },
2044
+ {
2045
+ "epoch": 15.33,
2046
+ "learning_rate": 2.87585e-05,
2047
+ "loss": 3.4299,
2048
+ "step": 170000
2049
+ },
2050
+ {
2051
+ "epoch": 15.38,
2052
+ "learning_rate": 2.8696e-05,
2053
+ "loss": 3.4234,
2054
+ "step": 170500
2055
+ },
2056
+ {
2057
+ "epoch": 15.42,
2058
+ "learning_rate": 2.8633500000000003e-05,
2059
+ "loss": 3.4272,
2060
+ "step": 171000
2061
+ },
2062
+ {
2063
+ "epoch": 15.47,
2064
+ "learning_rate": 2.8571000000000005e-05,
2065
+ "loss": 3.4253,
2066
+ "step": 171500
2067
+ },
2068
+ {
2069
+ "epoch": 15.51,
2070
+ "learning_rate": 2.85085e-05,
2071
+ "loss": 3.4262,
2072
+ "step": 172000
2073
+ },
2074
+ {
2075
+ "epoch": 15.56,
2076
+ "learning_rate": 2.8446000000000002e-05,
2077
+ "loss": 3.4278,
2078
+ "step": 172500
2079
+ },
2080
+ {
2081
+ "epoch": 15.6,
2082
+ "learning_rate": 2.83835e-05,
2083
+ "loss": 3.4224,
2084
+ "step": 173000
2085
+ },
2086
+ {
2087
+ "epoch": 15.65,
2088
+ "learning_rate": 2.8321000000000002e-05,
2089
+ "loss": 3.4207,
2090
+ "step": 173500
2091
+ },
2092
+ {
2093
+ "epoch": 15.69,
2094
+ "learning_rate": 2.8258625e-05,
2095
+ "loss": 3.4205,
2096
+ "step": 174000
2097
+ },
2098
+ {
2099
+ "epoch": 15.74,
2100
+ "learning_rate": 2.8196125e-05,
2101
+ "loss": 3.4271,
2102
+ "step": 174500
2103
+ },
2104
+ {
2105
+ "epoch": 15.78,
2106
+ "learning_rate": 2.8133625000000003e-05,
2107
+ "loss": 3.4237,
2108
+ "step": 175000
2109
+ },
2110
+ {
2111
+ "epoch": 15.83,
2112
+ "learning_rate": 2.8071125000000005e-05,
2113
+ "loss": 3.4261,
2114
+ "step": 175500
2115
+ },
2116
+ {
2117
+ "epoch": 15.87,
2118
+ "learning_rate": 2.8008750000000002e-05,
2119
+ "loss": 3.4213,
2120
+ "step": 176000
2121
+ },
2122
+ {
2123
+ "epoch": 15.92,
2124
+ "learning_rate": 2.794625e-05,
2125
+ "loss": 3.4281,
2126
+ "step": 176500
2127
+ },
2128
+ {
2129
+ "epoch": 15.96,
2130
+ "learning_rate": 2.7883750000000002e-05,
2131
+ "loss": 3.4257,
2132
+ "step": 177000
2133
+ },
2134
+ {
2135
+ "epoch": 16.01,
2136
+ "learning_rate": 2.7821249999999997e-05,
2137
+ "loss": 3.4206,
2138
+ "step": 177500
2139
+ },
2140
+ {
2141
+ "epoch": 16.05,
2142
+ "learning_rate": 2.7758875e-05,
2143
+ "loss": 3.4193,
2144
+ "step": 178000
2145
+ },
2146
+ {
2147
+ "epoch": 16.1,
2148
+ "learning_rate": 2.7696375000000003e-05,
2149
+ "loss": 3.4245,
2150
+ "step": 178500
2151
+ },
2152
+ {
2153
+ "epoch": 16.15,
2154
+ "learning_rate": 2.7633875000000005e-05,
2155
+ "loss": 3.4176,
2156
+ "step": 179000
2157
+ },
2158
+ {
2159
+ "epoch": 16.19,
2160
+ "learning_rate": 2.7571375e-05,
2161
+ "loss": 3.4192,
2162
+ "step": 179500
2163
+ },
2164
+ {
2165
+ "epoch": 16.24,
2166
+ "learning_rate": 2.7508875000000002e-05,
2167
+ "loss": 3.4212,
2168
+ "step": 180000
2169
+ },
2170
+ {
2171
+ "epoch": 16.28,
2172
+ "learning_rate": 2.7446500000000002e-05,
2173
+ "loss": 3.4189,
2174
+ "step": 180500
2175
+ },
2176
+ {
2177
+ "epoch": 16.33,
2178
+ "learning_rate": 2.7383999999999997e-05,
2179
+ "loss": 3.418,
2180
+ "step": 181000
2181
+ },
2182
+ {
2183
+ "epoch": 16.37,
2184
+ "learning_rate": 2.73215e-05,
2185
+ "loss": 3.4228,
2186
+ "step": 181500
2187
+ },
2188
+ {
2189
+ "epoch": 16.42,
2190
+ "learning_rate": 2.7259e-05,
2191
+ "loss": 3.4215,
2192
+ "step": 182000
2193
+ },
2194
+ {
2195
+ "epoch": 16.46,
2196
+ "learning_rate": 2.7196625000000005e-05,
2197
+ "loss": 3.4161,
2198
+ "step": 182500
2199
+ },
2200
+ {
2201
+ "epoch": 16.51,
2202
+ "learning_rate": 2.7134125e-05,
2203
+ "loss": 3.4222,
2204
+ "step": 183000
2205
+ },
2206
+ {
2207
+ "epoch": 16.55,
2208
+ "learning_rate": 2.7071625000000002e-05,
2209
+ "loss": 3.4198,
2210
+ "step": 183500
2211
+ },
2212
+ {
2213
+ "epoch": 16.6,
2214
+ "learning_rate": 2.7009125e-05,
2215
+ "loss": 3.4212,
2216
+ "step": 184000
2217
+ },
2218
+ {
2219
+ "epoch": 16.64,
2220
+ "learning_rate": 2.6946750000000004e-05,
2221
+ "loss": 3.4148,
2222
+ "step": 184500
2223
+ },
2224
+ {
2225
+ "epoch": 16.69,
2226
+ "learning_rate": 2.6884375e-05,
2227
+ "loss": 3.4235,
2228
+ "step": 185000
2229
+ },
2230
+ {
2231
+ "epoch": 16.73,
2232
+ "learning_rate": 2.6821875000000003e-05,
2233
+ "loss": 3.4216,
2234
+ "step": 185500
2235
+ },
2236
+ {
2237
+ "epoch": 16.78,
2238
+ "learning_rate": 2.6759375e-05,
2239
+ "loss": 3.4211,
2240
+ "step": 186000
2241
+ },
2242
+ {
2243
+ "epoch": 16.82,
2244
+ "learning_rate": 2.6696875e-05,
2245
+ "loss": 3.4161,
2246
+ "step": 186500
2247
+ },
2248
+ {
2249
+ "epoch": 16.87,
2250
+ "learning_rate": 2.66345e-05,
2251
+ "loss": 3.4201,
2252
+ "step": 187000
2253
+ },
2254
+ {
2255
+ "epoch": 16.91,
2256
+ "learning_rate": 2.6572000000000002e-05,
2257
+ "loss": 3.4175,
2258
+ "step": 187500
2259
+ },
2260
+ {
2261
+ "epoch": 16.96,
2262
+ "learning_rate": 2.6509500000000004e-05,
2263
+ "loss": 3.421,
2264
+ "step": 188000
2265
+ },
2266
+ {
2267
+ "epoch": 17.0,
2268
+ "learning_rate": 2.6447e-05,
2269
+ "loss": 3.4195,
2270
+ "step": 188500
2271
+ },
2272
+ {
2273
+ "epoch": 17.05,
2274
+ "learning_rate": 2.63845e-05,
2275
+ "loss": 3.4126,
2276
+ "step": 189000
2277
+ },
2278
+ {
2279
+ "epoch": 17.09,
2280
+ "learning_rate": 2.6322000000000003e-05,
2281
+ "loss": 3.4153,
2282
+ "step": 189500
2283
+ },
2284
+ {
2285
+ "epoch": 17.14,
2286
+ "learning_rate": 2.6259500000000005e-05,
2287
+ "loss": 3.4149,
2288
+ "step": 190000
2289
+ },
2290
+ {
2291
+ "epoch": 17.18,
2292
+ "learning_rate": 2.6197e-05,
2293
+ "loss": 3.4204,
2294
+ "step": 190500
2295
+ },
2296
+ {
2297
+ "epoch": 17.23,
2298
+ "learning_rate": 2.6134625e-05,
2299
+ "loss": 3.412,
2300
+ "step": 191000
2301
+ },
2302
+ {
2303
+ "epoch": 17.27,
2304
+ "learning_rate": 2.6072125000000002e-05,
2305
+ "loss": 3.4135,
2306
+ "step": 191500
2307
+ },
2308
+ {
2309
+ "epoch": 17.32,
2310
+ "learning_rate": 2.6009624999999997e-05,
2311
+ "loss": 3.4191,
2312
+ "step": 192000
2313
+ },
2314
+ {
2315
+ "epoch": 17.36,
2316
+ "learning_rate": 2.5947125e-05,
2317
+ "loss": 3.4165,
2318
+ "step": 192500
2319
+ },
2320
+ {
2321
+ "epoch": 17.41,
2322
+ "learning_rate": 2.5884750000000003e-05,
2323
+ "loss": 3.4179,
2324
+ "step": 193000
2325
+ },
2326
+ {
2327
+ "epoch": 17.45,
2328
+ "learning_rate": 2.5822250000000005e-05,
2329
+ "loss": 3.4153,
2330
+ "step": 193500
2331
+ },
2332
+ {
2333
+ "epoch": 17.5,
2334
+ "learning_rate": 2.575975e-05,
2335
+ "loss": 3.4167,
2336
+ "step": 194000
2337
+ },
2338
+ {
2339
+ "epoch": 17.54,
2340
+ "learning_rate": 2.5697250000000002e-05,
2341
+ "loss": 3.4153,
2342
+ "step": 194500
2343
+ },
2344
+ {
2345
+ "epoch": 17.59,
2346
+ "learning_rate": 2.5634875000000002e-05,
2347
+ "loss": 3.4107,
2348
+ "step": 195000
2349
+ },
2350
+ {
2351
+ "epoch": 17.63,
2352
+ "learning_rate": 2.5572375000000004e-05,
2353
+ "loss": 3.4158,
2354
+ "step": 195500
2355
+ },
2356
+ {
2357
+ "epoch": 17.68,
2358
+ "learning_rate": 2.5509875e-05,
2359
+ "loss": 3.4149,
2360
+ "step": 196000
2361
+ },
2362
+ {
2363
+ "epoch": 17.72,
2364
+ "learning_rate": 2.5447500000000003e-05,
2365
+ "loss": 3.4192,
2366
+ "step": 196500
2367
+ },
2368
+ {
2369
+ "epoch": 17.77,
2370
+ "learning_rate": 2.5385000000000002e-05,
2371
+ "loss": 3.4126,
2372
+ "step": 197000
2373
+ },
2374
+ {
2375
+ "epoch": 17.81,
2376
+ "learning_rate": 2.53225e-05,
2377
+ "loss": 3.4128,
2378
+ "step": 197500
2379
+ },
2380
+ {
2381
+ "epoch": 17.86,
2382
+ "learning_rate": 2.526e-05,
2383
+ "loss": 3.4174,
2384
+ "step": 198000
2385
+ },
2386
+ {
2387
+ "epoch": 17.9,
2388
+ "learning_rate": 2.51975e-05,
2389
+ "loss": 3.4133,
2390
+ "step": 198500
2391
+ },
2392
+ {
2393
+ "epoch": 17.95,
2394
+ "learning_rate": 2.5135125000000004e-05,
2395
+ "loss": 3.4146,
2396
+ "step": 199000
2397
+ },
2398
+ {
2399
+ "epoch": 17.99,
2400
+ "learning_rate": 2.5072625e-05,
2401
+ "loss": 3.4183,
2402
+ "step": 199500
2403
+ },
2404
+ {
2405
+ "epoch": 18.04,
2406
+ "learning_rate": 2.5010125e-05,
2407
+ "loss": 3.4109,
2408
+ "step": 200000
2409
+ },
2410
+ {
2411
+ "epoch": 18.08,
2412
+ "learning_rate": 2.4947625000000003e-05,
2413
+ "loss": 3.4156,
2414
+ "step": 200500
2415
+ },
2416
+ {
2417
+ "epoch": 18.13,
2418
+ "learning_rate": 2.488525e-05,
2419
+ "loss": 3.4113,
2420
+ "step": 201000
2421
+ },
2422
+ {
2423
+ "epoch": 18.17,
2424
+ "learning_rate": 2.482275e-05,
2425
+ "loss": 3.4134,
2426
+ "step": 201500
2427
+ },
2428
+ {
2429
+ "epoch": 18.22,
2430
+ "learning_rate": 2.476025e-05,
2431
+ "loss": 3.4117,
2432
+ "step": 202000
2433
+ },
2434
+ {
2435
+ "epoch": 18.26,
2436
+ "learning_rate": 2.469775e-05,
2437
+ "loss": 3.4121,
2438
+ "step": 202500
2439
+ },
2440
+ {
2441
+ "epoch": 18.31,
2442
+ "learning_rate": 2.4635375000000003e-05,
2443
+ "loss": 3.413,
2444
+ "step": 203000
2445
+ },
2446
+ {
2447
+ "epoch": 18.35,
2448
+ "learning_rate": 2.4572875e-05,
2449
+ "loss": 3.412,
2450
+ "step": 203500
2451
+ },
2452
+ {
2453
+ "epoch": 18.4,
2454
+ "learning_rate": 2.4510375000000003e-05,
2455
+ "loss": 3.4081,
2456
+ "step": 204000
2457
+ },
2458
+ {
2459
+ "epoch": 18.45,
2460
+ "learning_rate": 2.4447875000000002e-05,
2461
+ "loss": 3.4136,
2462
+ "step": 204500
2463
+ },
2464
+ {
2465
+ "epoch": 18.49,
2466
+ "learning_rate": 2.4385375e-05,
2467
+ "loss": 3.4101,
2468
+ "step": 205000
2469
+ },
2470
+ {
2471
+ "epoch": 18.54,
2472
+ "learning_rate": 2.4322875000000002e-05,
2473
+ "loss": 3.4107,
2474
+ "step": 205500
2475
+ },
2476
+ {
2477
+ "epoch": 18.58,
2478
+ "learning_rate": 2.4260375e-05,
2479
+ "loss": 3.4144,
2480
+ "step": 206000
2481
+ },
2482
+ {
2483
+ "epoch": 18.63,
2484
+ "learning_rate": 2.4197875e-05,
2485
+ "loss": 3.41,
2486
+ "step": 206500
2487
+ },
2488
+ {
2489
+ "epoch": 18.67,
2490
+ "learning_rate": 2.4135375e-05,
2491
+ "loss": 3.4122,
2492
+ "step": 207000
2493
+ },
2494
+ {
2495
+ "epoch": 18.72,
2496
+ "learning_rate": 2.4073e-05,
2497
+ "loss": 3.4093,
2498
+ "step": 207500
2499
+ },
2500
+ {
2501
+ "epoch": 18.76,
2502
+ "learning_rate": 2.4010625000000002e-05,
2503
+ "loss": 3.4078,
2504
+ "step": 208000
2505
+ },
2506
+ {
2507
+ "epoch": 18.81,
2508
+ "learning_rate": 2.3948125e-05,
2509
+ "loss": 3.4094,
2510
+ "step": 208500
2511
+ },
2512
+ {
2513
+ "epoch": 18.85,
2514
+ "learning_rate": 2.3885625e-05,
2515
+ "loss": 3.4083,
2516
+ "step": 209000
2517
+ },
2518
+ {
2519
+ "epoch": 18.9,
2520
+ "learning_rate": 2.3823125e-05,
2521
+ "loss": 3.4111,
2522
+ "step": 209500
2523
+ },
2524
+ {
2525
+ "epoch": 18.94,
2526
+ "learning_rate": 2.3760625e-05,
2527
+ "loss": 3.4105,
2528
+ "step": 210000
2529
+ },
2530
+ {
2531
+ "epoch": 18.99,
2532
+ "learning_rate": 2.3698125e-05,
2533
+ "loss": 3.4112,
2534
+ "step": 210500
2535
+ },
2536
+ {
2537
+ "epoch": 19.03,
2538
+ "learning_rate": 2.3635625e-05,
2539
+ "loss": 3.4081,
2540
+ "step": 211000
2541
+ },
2542
+ {
2543
+ "epoch": 19.08,
2544
+ "learning_rate": 2.3573125e-05,
2545
+ "loss": 3.4082,
2546
+ "step": 211500
2547
+ },
2548
+ {
2549
+ "epoch": 19.12,
2550
+ "learning_rate": 2.3510750000000002e-05,
2551
+ "loss": 3.4088,
2552
+ "step": 212000
2553
+ },
2554
+ {
2555
+ "epoch": 19.17,
2556
+ "learning_rate": 2.3448250000000004e-05,
2557
+ "loss": 3.4083,
2558
+ "step": 212500
2559
+ },
2560
+ {
2561
+ "epoch": 19.21,
2562
+ "learning_rate": 2.3385875e-05,
2563
+ "loss": 3.4091,
2564
+ "step": 213000
2565
+ },
2566
+ {
2567
+ "epoch": 19.26,
2568
+ "learning_rate": 2.3323375e-05,
2569
+ "loss": 3.4074,
2570
+ "step": 213500
2571
+ },
2572
+ {
2573
+ "epoch": 19.3,
2574
+ "learning_rate": 2.3260875e-05,
2575
+ "loss": 3.4093,
2576
+ "step": 214000
2577
+ },
2578
+ {
2579
+ "epoch": 19.35,
2580
+ "learning_rate": 2.3198375e-05,
2581
+ "loss": 3.4044,
2582
+ "step": 214500
2583
+ },
2584
+ {
2585
+ "epoch": 19.39,
2586
+ "learning_rate": 2.3135875e-05,
2587
+ "loss": 3.4078,
2588
+ "step": 215000
2589
+ },
2590
+ {
2591
+ "epoch": 19.44,
2592
+ "learning_rate": 2.3073375e-05,
2593
+ "loss": 3.4072,
2594
+ "step": 215500
2595
+ },
2596
+ {
2597
+ "epoch": 19.48,
2598
+ "learning_rate": 2.3011e-05,
2599
+ "loss": 3.4091,
2600
+ "step": 216000
2601
+ },
2602
+ {
2603
+ "epoch": 19.53,
2604
+ "learning_rate": 2.29485e-05,
2605
+ "loss": 3.4021,
2606
+ "step": 216500
2607
+ },
2608
+ {
2609
+ "epoch": 19.57,
2610
+ "learning_rate": 2.2886e-05,
2611
+ "loss": 3.4095,
2612
+ "step": 217000
2613
+ },
2614
+ {
2615
+ "epoch": 19.62,
2616
+ "learning_rate": 2.28235e-05,
2617
+ "loss": 3.4028,
2618
+ "step": 217500
2619
+ },
2620
+ {
2621
+ "epoch": 19.66,
2622
+ "learning_rate": 2.2761e-05,
2623
+ "loss": 3.407,
2624
+ "step": 218000
2625
+ },
2626
+ {
2627
+ "epoch": 19.71,
2628
+ "learning_rate": 2.26985e-05,
2629
+ "loss": 3.4132,
2630
+ "step": 218500
2631
+ },
2632
+ {
2633
+ "epoch": 19.75,
2634
+ "learning_rate": 2.2636e-05,
2635
+ "loss": 3.4077,
2636
+ "step": 219000
2637
+ },
2638
+ {
2639
+ "epoch": 19.8,
2640
+ "learning_rate": 2.25735e-05,
2641
+ "loss": 3.4138,
2642
+ "step": 219500
2643
+ },
2644
+ {
2645
+ "epoch": 19.84,
2646
+ "learning_rate": 2.2511125000000004e-05,
2647
+ "loss": 3.4038,
2648
+ "step": 220000
2649
+ },
2650
+ {
2651
+ "epoch": 19.89,
2652
+ "learning_rate": 2.244875e-05,
2653
+ "loss": 3.4052,
2654
+ "step": 220500
2655
+ },
2656
+ {
2657
+ "epoch": 19.93,
2658
+ "learning_rate": 2.238625e-05,
2659
+ "loss": 3.4074,
2660
+ "step": 221000
2661
+ },
2662
+ {
2663
+ "epoch": 19.98,
2664
+ "learning_rate": 2.232375e-05,
2665
+ "loss": 3.4074,
2666
+ "step": 221500
2667
+ },
2668
+ {
2669
+ "epoch": 20.02,
2670
+ "learning_rate": 2.226125e-05,
2671
+ "loss": 3.404,
2672
+ "step": 222000
2673
+ },
2674
+ {
2675
+ "epoch": 20.07,
2676
+ "learning_rate": 2.219875e-05,
2677
+ "loss": 3.4,
2678
+ "step": 222500
2679
+ },
2680
+ {
2681
+ "epoch": 20.11,
2682
+ "learning_rate": 2.213625e-05,
2683
+ "loss": 3.4001,
2684
+ "step": 223000
2685
+ },
2686
+ {
2687
+ "epoch": 20.16,
2688
+ "learning_rate": 2.2073750000000002e-05,
2689
+ "loss": 3.3976,
2690
+ "step": 223500
2691
+ },
2692
+ {
2693
+ "epoch": 20.2,
2694
+ "learning_rate": 2.2011375000000002e-05,
2695
+ "loss": 3.4042,
2696
+ "step": 224000
2697
+ },
2698
+ {
2699
+ "epoch": 20.25,
2700
+ "learning_rate": 2.1948875e-05,
2701
+ "loss": 3.4025,
2702
+ "step": 224500
2703
+ },
2704
+ {
2705
+ "epoch": 20.29,
2706
+ "learning_rate": 2.1886375e-05,
2707
+ "loss": 3.4024,
2708
+ "step": 225000
2709
+ },
2710
+ {
2711
+ "epoch": 20.34,
2712
+ "learning_rate": 2.1823875e-05,
2713
+ "loss": 3.4071,
2714
+ "step": 225500
2715
+ },
2716
+ {
2717
+ "epoch": 20.38,
2718
+ "learning_rate": 2.1761375e-05,
2719
+ "loss": 3.3998,
2720
+ "step": 226000
2721
+ },
2722
+ {
2723
+ "epoch": 20.43,
2724
+ "learning_rate": 2.1698875e-05,
2725
+ "loss": 3.4028,
2726
+ "step": 226500
2727
+ },
2728
+ {
2729
+ "epoch": 20.47,
2730
+ "learning_rate": 2.1636375e-05,
2731
+ "loss": 3.4084,
2732
+ "step": 227000
2733
+ },
2734
+ {
2735
+ "epoch": 20.52,
2736
+ "learning_rate": 2.1573875000000002e-05,
2737
+ "loss": 3.4082,
2738
+ "step": 227500
2739
+ },
2740
+ {
2741
+ "epoch": 20.56,
2742
+ "learning_rate": 2.1511500000000002e-05,
2743
+ "loss": 3.4051,
2744
+ "step": 228000
2745
+ },
2746
+ {
2747
+ "epoch": 20.61,
2748
+ "learning_rate": 2.1449e-05,
2749
+ "loss": 3.4011,
2750
+ "step": 228500
2751
+ },
2752
+ {
2753
+ "epoch": 20.65,
2754
+ "learning_rate": 2.1386625e-05,
2755
+ "loss": 3.4041,
2756
+ "step": 229000
2757
+ },
2758
+ {
2759
+ "epoch": 20.7,
2760
+ "learning_rate": 2.1324125e-05,
2761
+ "loss": 3.4057,
2762
+ "step": 229500
2763
+ },
2764
+ {
2765
+ "epoch": 20.75,
2766
+ "learning_rate": 2.1261625e-05,
2767
+ "loss": 3.4035,
2768
+ "step": 230000
2769
+ },
2770
+ {
2771
+ "epoch": 20.79,
2772
+ "learning_rate": 2.1199125e-05,
2773
+ "loss": 3.404,
2774
+ "step": 230500
2775
+ },
2776
+ {
2777
+ "epoch": 20.84,
2778
+ "learning_rate": 2.1136625000000002e-05,
2779
+ "loss": 3.4056,
2780
+ "step": 231000
2781
+ },
2782
+ {
2783
+ "epoch": 20.88,
2784
+ "learning_rate": 2.1074125e-05,
2785
+ "loss": 3.4059,
2786
+ "step": 231500
2787
+ },
2788
+ {
2789
+ "epoch": 20.93,
2790
+ "learning_rate": 2.101175e-05,
2791
+ "loss": 3.4038,
2792
+ "step": 232000
2793
+ },
2794
+ {
2795
+ "epoch": 20.97,
2796
+ "learning_rate": 2.094925e-05,
2797
+ "loss": 3.4031,
2798
+ "step": 232500
2799
+ },
2800
+ {
2801
+ "epoch": 21.02,
2802
+ "learning_rate": 2.088675e-05,
2803
+ "loss": 3.4027,
2804
+ "step": 233000
2805
+ },
2806
+ {
2807
+ "epoch": 21.06,
2808
+ "learning_rate": 2.082425e-05,
2809
+ "loss": 3.3979,
2810
+ "step": 233500
2811
+ },
2812
+ {
2813
+ "epoch": 21.11,
2814
+ "learning_rate": 2.076175e-05,
2815
+ "loss": 3.4026,
2816
+ "step": 234000
2817
+ },
2818
+ {
2819
+ "epoch": 21.15,
2820
+ "learning_rate": 2.069925e-05,
2821
+ "loss": 3.4007,
2822
+ "step": 234500
2823
+ },
2824
+ {
2825
+ "epoch": 21.2,
2826
+ "learning_rate": 2.0636750000000002e-05,
2827
+ "loss": 3.3966,
2828
+ "step": 235000
2829
+ },
2830
+ {
2831
+ "epoch": 21.24,
2832
+ "learning_rate": 2.057425e-05,
2833
+ "loss": 3.3991,
2834
+ "step": 235500
2835
+ },
2836
+ {
2837
+ "epoch": 21.29,
2838
+ "learning_rate": 2.0511875e-05,
2839
+ "loss": 3.3993,
2840
+ "step": 236000
2841
+ },
2842
+ {
2843
+ "epoch": 21.33,
2844
+ "learning_rate": 2.0449375000000003e-05,
2845
+ "loss": 3.4017,
2846
+ "step": 236500
2847
+ },
2848
+ {
2849
+ "epoch": 21.38,
2850
+ "learning_rate": 2.0386875e-05,
2851
+ "loss": 3.3995,
2852
+ "step": 237000
2853
+ },
2854
+ {
2855
+ "epoch": 21.42,
2856
+ "learning_rate": 2.0324375e-05,
2857
+ "loss": 3.4066,
2858
+ "step": 237500
2859
+ },
2860
+ {
2861
+ "epoch": 21.47,
2862
+ "learning_rate": 2.0262e-05,
2863
+ "loss": 3.3995,
2864
+ "step": 238000
2865
+ },
2866
+ {
2867
+ "epoch": 21.51,
2868
+ "learning_rate": 2.0199500000000002e-05,
2869
+ "loss": 3.4038,
2870
+ "step": 238500
2871
+ },
2872
+ {
2873
+ "epoch": 21.56,
2874
+ "learning_rate": 2.0137e-05,
2875
+ "loss": 3.3997,
2876
+ "step": 239000
2877
+ },
2878
+ {
2879
+ "epoch": 21.6,
2880
+ "learning_rate": 2.0074500000000002e-05,
2881
+ "loss": 3.3984,
2882
+ "step": 239500
2883
+ },
2884
+ {
2885
+ "epoch": 21.65,
2886
+ "learning_rate": 2.0012125e-05,
2887
+ "loss": 3.4012,
2888
+ "step": 240000
2889
+ },
2890
+ {
2891
+ "epoch": 21.69,
2892
+ "learning_rate": 1.9949625e-05,
2893
+ "loss": 3.3983,
2894
+ "step": 240500
2895
+ },
2896
+ {
2897
+ "epoch": 21.74,
2898
+ "learning_rate": 1.9887125e-05,
2899
+ "loss": 3.4043,
2900
+ "step": 241000
2901
+ },
2902
+ {
2903
+ "epoch": 21.78,
2904
+ "learning_rate": 1.9824625e-05,
2905
+ "loss": 3.3983,
2906
+ "step": 241500
2907
+ },
2908
+ {
2909
+ "epoch": 21.83,
2910
+ "learning_rate": 1.9762250000000002e-05,
2911
+ "loss": 3.4018,
2912
+ "step": 242000
2913
+ },
2914
+ {
2915
+ "epoch": 21.87,
2916
+ "learning_rate": 1.969975e-05,
2917
+ "loss": 3.4004,
2918
+ "step": 242500
2919
+ },
2920
+ {
2921
+ "epoch": 21.92,
2922
+ "learning_rate": 1.9637250000000002e-05,
2923
+ "loss": 3.4026,
2924
+ "step": 243000
2925
+ },
2926
+ {
2927
+ "epoch": 21.96,
2928
+ "learning_rate": 1.9574875e-05,
2929
+ "loss": 3.4022,
2930
+ "step": 243500
2931
+ },
2932
+ {
2933
+ "epoch": 22.01,
2934
+ "learning_rate": 1.9512375e-05,
2935
+ "loss": 3.3985,
2936
+ "step": 244000
2937
+ },
2938
+ {
2939
+ "epoch": 22.05,
2940
+ "learning_rate": 1.9449875e-05,
2941
+ "loss": 3.3935,
2942
+ "step": 244500
2943
+ },
2944
+ {
2945
+ "epoch": 22.1,
2946
+ "learning_rate": 1.9387375000000002e-05,
2947
+ "loss": 3.4019,
2948
+ "step": 245000
2949
+ },
2950
+ {
2951
+ "epoch": 22.14,
2952
+ "learning_rate": 1.9324875e-05,
2953
+ "loss": 3.396,
2954
+ "step": 245500
2955
+ },
2956
+ {
2957
+ "epoch": 22.19,
2958
+ "learning_rate": 1.9262375000000002e-05,
2959
+ "loss": 3.3981,
2960
+ "step": 246000
2961
+ },
2962
+ {
2963
+ "epoch": 22.23,
2964
+ "learning_rate": 1.9200125e-05,
2965
+ "loss": 3.3935,
2966
+ "step": 246500
2967
+ },
2968
+ {
2969
+ "epoch": 22.28,
2970
+ "learning_rate": 1.9137625e-05,
2971
+ "loss": 3.396,
2972
+ "step": 247000
2973
+ },
2974
+ {
2975
+ "epoch": 22.32,
2976
+ "learning_rate": 1.9075125e-05,
2977
+ "loss": 3.3978,
2978
+ "step": 247500
2979
+ },
2980
+ {
2981
+ "epoch": 22.37,
2982
+ "learning_rate": 1.9012625e-05,
2983
+ "loss": 3.3973,
2984
+ "step": 248000
2985
+ },
2986
+ {
2987
+ "epoch": 22.41,
2988
+ "learning_rate": 1.8950125000000002e-05,
2989
+ "loss": 3.3974,
2990
+ "step": 248500
2991
+ },
2992
+ {
2993
+ "epoch": 22.46,
2994
+ "learning_rate": 1.8887625e-05,
2995
+ "loss": 3.3993,
2996
+ "step": 249000
2997
+ },
2998
+ {
2999
+ "epoch": 22.5,
3000
+ "learning_rate": 1.882525e-05,
3001
+ "loss": 3.4011,
3002
+ "step": 249500
3003
+ },
3004
+ {
3005
+ "epoch": 22.55,
3006
+ "learning_rate": 1.8762750000000003e-05,
3007
+ "loss": 3.3976,
3008
+ "step": 250000
3009
+ },
3010
+ {
3011
+ "epoch": 22.59,
3012
+ "learning_rate": 1.870025e-05,
3013
+ "loss": 3.3981,
3014
+ "step": 250500
3015
+ },
3016
+ {
3017
+ "epoch": 22.64,
3018
+ "learning_rate": 1.863775e-05,
3019
+ "loss": 3.4044,
3020
+ "step": 251000
3021
+ },
3022
+ {
3023
+ "epoch": 22.68,
3024
+ "learning_rate": 1.8575249999999998e-05,
3025
+ "loss": 3.3969,
3026
+ "step": 251500
3027
+ },
3028
+ {
3029
+ "epoch": 22.73,
3030
+ "learning_rate": 1.851275e-05,
3031
+ "loss": 3.3933,
3032
+ "step": 252000
3033
+ },
3034
+ {
3035
+ "epoch": 22.77,
3036
+ "learning_rate": 1.8450250000000002e-05,
3037
+ "loss": 3.3981,
3038
+ "step": 252500
3039
+ },
3040
+ {
3041
+ "epoch": 22.82,
3042
+ "learning_rate": 1.838775e-05,
3043
+ "loss": 3.3936,
3044
+ "step": 253000
3045
+ },
3046
+ {
3047
+ "epoch": 22.86,
3048
+ "learning_rate": 1.8325250000000002e-05,
3049
+ "loss": 3.3964,
3050
+ "step": 253500
3051
+ },
3052
+ {
3053
+ "epoch": 22.91,
3054
+ "learning_rate": 1.826275e-05,
3055
+ "loss": 3.397,
3056
+ "step": 254000
3057
+ },
3058
+ {
3059
+ "epoch": 22.95,
3060
+ "learning_rate": 1.8200250000000002e-05,
3061
+ "loss": 3.3963,
3062
+ "step": 254500
3063
+ },
3064
+ {
3065
+ "epoch": 23.0,
3066
+ "learning_rate": 1.813775e-05,
3067
+ "loss": 3.3977,
3068
+ "step": 255000
3069
+ },
3070
+ {
3071
+ "epoch": 23.05,
3072
+ "learning_rate": 1.8075375e-05,
3073
+ "loss": 3.3935,
3074
+ "step": 255500
3075
+ },
3076
+ {
3077
+ "epoch": 23.09,
3078
+ "learning_rate": 1.8013000000000002e-05,
3079
+ "loss": 3.3964,
3080
+ "step": 256000
3081
+ },
3082
+ {
3083
+ "epoch": 23.14,
3084
+ "learning_rate": 1.79505e-05,
3085
+ "loss": 3.401,
3086
+ "step": 256500
3087
+ },
3088
+ {
3089
+ "epoch": 23.18,
3090
+ "learning_rate": 1.7888125e-05,
3091
+ "loss": 3.3938,
3092
+ "step": 257000
3093
+ },
3094
+ {
3095
+ "epoch": 23.23,
3096
+ "learning_rate": 1.7825625000000003e-05,
3097
+ "loss": 3.3931,
3098
+ "step": 257500
3099
+ },
3100
+ {
3101
+ "epoch": 23.27,
3102
+ "learning_rate": 1.7763125e-05,
3103
+ "loss": 3.3949,
3104
+ "step": 258000
3105
+ },
3106
+ {
3107
+ "epoch": 23.32,
3108
+ "learning_rate": 1.7700625e-05,
3109
+ "loss": 3.396,
3110
+ "step": 258500
3111
+ },
3112
+ {
3113
+ "epoch": 23.36,
3114
+ "learning_rate": 1.7638124999999998e-05,
3115
+ "loss": 3.3942,
3116
+ "step": 259000
3117
+ },
3118
+ {
3119
+ "epoch": 23.41,
3120
+ "learning_rate": 1.7575625e-05,
3121
+ "loss": 3.3946,
3122
+ "step": 259500
3123
+ },
3124
+ {
3125
+ "epoch": 23.45,
3126
+ "learning_rate": 1.7513125e-05,
3127
+ "loss": 3.3992,
3128
+ "step": 260000
3129
+ },
3130
+ {
3131
+ "epoch": 23.5,
3132
+ "learning_rate": 1.7450625e-05,
3133
+ "loss": 3.391,
3134
+ "step": 260500
3135
+ },
3136
+ {
3137
+ "epoch": 23.54,
3138
+ "learning_rate": 1.7388125e-05,
3139
+ "loss": 3.3936,
3140
+ "step": 261000
3141
+ },
3142
+ {
3143
+ "epoch": 23.59,
3144
+ "learning_rate": 1.7325625e-05,
3145
+ "loss": 3.3946,
3146
+ "step": 261500
3147
+ },
3148
+ {
3149
+ "epoch": 23.63,
3150
+ "learning_rate": 1.7263125000000003e-05,
3151
+ "loss": 3.3935,
3152
+ "step": 262000
3153
+ },
3154
+ {
3155
+ "epoch": 23.68,
3156
+ "learning_rate": 1.7200625e-05,
3157
+ "loss": 3.3954,
3158
+ "step": 262500
3159
+ },
3160
+ {
3161
+ "epoch": 23.72,
3162
+ "learning_rate": 1.713825e-05,
3163
+ "loss": 3.3972,
3164
+ "step": 263000
3165
+ },
3166
+ {
3167
+ "epoch": 23.77,
3168
+ "learning_rate": 1.707575e-05,
3169
+ "loss": 3.3912,
3170
+ "step": 263500
3171
+ },
3172
+ {
3173
+ "epoch": 23.81,
3174
+ "learning_rate": 1.701325e-05,
3175
+ "loss": 3.3925,
3176
+ "step": 264000
3177
+ },
3178
+ {
3179
+ "epoch": 23.86,
3180
+ "learning_rate": 1.6950875000000002e-05,
3181
+ "loss": 3.3921,
3182
+ "step": 264500
3183
+ },
3184
+ {
3185
+ "epoch": 23.9,
3186
+ "learning_rate": 1.6888375e-05,
3187
+ "loss": 3.3894,
3188
+ "step": 265000
3189
+ },
3190
+ {
3191
+ "epoch": 23.95,
3192
+ "learning_rate": 1.6825875000000003e-05,
3193
+ "loss": 3.3949,
3194
+ "step": 265500
3195
+ },
3196
+ {
3197
+ "epoch": 23.99,
3198
+ "learning_rate": 1.6763375e-05,
3199
+ "loss": 3.3925,
3200
+ "step": 266000
3201
+ },
3202
+ {
3203
+ "epoch": 24.04,
3204
+ "learning_rate": 1.6700875000000003e-05,
3205
+ "loss": 3.3923,
3206
+ "step": 266500
3207
+ },
3208
+ {
3209
+ "epoch": 24.08,
3210
+ "learning_rate": 1.66385e-05,
3211
+ "loss": 3.3892,
3212
+ "step": 267000
3213
+ },
3214
+ {
3215
+ "epoch": 24.13,
3216
+ "learning_rate": 1.6576e-05,
3217
+ "loss": 3.3893,
3218
+ "step": 267500
3219
+ },
3220
+ {
3221
+ "epoch": 24.17,
3222
+ "learning_rate": 1.65135e-05,
3223
+ "loss": 3.3872,
3224
+ "step": 268000
3225
+ },
3226
+ {
3227
+ "epoch": 24.22,
3228
+ "learning_rate": 1.6451e-05,
3229
+ "loss": 3.3936,
3230
+ "step": 268500
3231
+ },
3232
+ {
3233
+ "epoch": 24.26,
3234
+ "learning_rate": 1.63885e-05,
3235
+ "loss": 3.3879,
3236
+ "step": 269000
3237
+ },
3238
+ {
3239
+ "epoch": 24.31,
3240
+ "learning_rate": 1.6326e-05,
3241
+ "loss": 3.3888,
3242
+ "step": 269500
3243
+ },
3244
+ {
3245
+ "epoch": 24.35,
3246
+ "learning_rate": 1.6263625000000003e-05,
3247
+ "loss": 3.392,
3248
+ "step": 270000
3249
+ },
3250
+ {
3251
+ "epoch": 24.4,
3252
+ "learning_rate": 1.6201125e-05,
3253
+ "loss": 3.3913,
3254
+ "step": 270500
3255
+ },
3256
+ {
3257
+ "epoch": 24.44,
3258
+ "learning_rate": 1.6138625e-05,
3259
+ "loss": 3.3901,
3260
+ "step": 271000
3261
+ },
3262
+ {
3263
+ "epoch": 24.49,
3264
+ "learning_rate": 1.6076125000000002e-05,
3265
+ "loss": 3.391,
3266
+ "step": 271500
3267
+ },
3268
+ {
3269
+ "epoch": 24.53,
3270
+ "learning_rate": 1.6013625e-05,
3271
+ "loss": 3.393,
3272
+ "step": 272000
3273
+ },
3274
+ {
3275
+ "epoch": 24.58,
3276
+ "learning_rate": 1.5951125e-05,
3277
+ "loss": 3.3953,
3278
+ "step": 272500
3279
+ },
3280
+ {
3281
+ "epoch": 24.62,
3282
+ "learning_rate": 1.5888625e-05,
3283
+ "loss": 3.3892,
3284
+ "step": 273000
3285
+ },
3286
+ {
3287
+ "epoch": 24.67,
3288
+ "learning_rate": 1.5826125e-05,
3289
+ "loss": 3.3964,
3290
+ "step": 273500
3291
+ },
3292
+ {
3293
+ "epoch": 24.71,
3294
+ "learning_rate": 1.5763750000000003e-05,
3295
+ "loss": 3.3945,
3296
+ "step": 274000
3297
+ },
3298
+ {
3299
+ "epoch": 24.76,
3300
+ "learning_rate": 1.570125e-05,
3301
+ "loss": 3.3964,
3302
+ "step": 274500
3303
+ },
3304
+ {
3305
+ "epoch": 24.8,
3306
+ "learning_rate": 1.5638875e-05,
3307
+ "loss": 3.3937,
3308
+ "step": 275000
3309
+ },
3310
+ {
3311
+ "epoch": 24.85,
3312
+ "learning_rate": 1.5576375e-05,
3313
+ "loss": 3.3897,
3314
+ "step": 275500
3315
+ },
3316
+ {
3317
+ "epoch": 24.89,
3318
+ "learning_rate": 1.5513875e-05,
3319
+ "loss": 3.3916,
3320
+ "step": 276000
3321
+ },
3322
+ {
3323
+ "epoch": 24.94,
3324
+ "learning_rate": 1.5451375e-05,
3325
+ "loss": 3.3917,
3326
+ "step": 276500
3327
+ },
3328
+ {
3329
+ "epoch": 24.98,
3330
+ "learning_rate": 1.5389e-05,
3331
+ "loss": 3.3941,
3332
+ "step": 277000
3333
+ },
3334
+ {
3335
+ "epoch": 25.03,
3336
+ "learning_rate": 1.5326500000000003e-05,
3337
+ "loss": 3.3929,
3338
+ "step": 277500
3339
+ },
3340
+ {
3341
+ "epoch": 25.07,
3342
+ "learning_rate": 1.5264e-05,
3343
+ "loss": 3.3903,
3344
+ "step": 278000
3345
+ },
3346
+ {
3347
+ "epoch": 25.12,
3348
+ "learning_rate": 1.5201500000000002e-05,
3349
+ "loss": 3.3851,
3350
+ "step": 278500
3351
+ },
3352
+ {
3353
+ "epoch": 25.16,
3354
+ "learning_rate": 1.5139e-05,
3355
+ "loss": 3.3887,
3356
+ "step": 279000
3357
+ },
3358
+ {
3359
+ "epoch": 25.21,
3360
+ "learning_rate": 1.5076500000000002e-05,
3361
+ "loss": 3.3908,
3362
+ "step": 279500
3363
+ },
3364
+ {
3365
+ "epoch": 25.25,
3366
+ "learning_rate": 1.5014e-05,
3367
+ "loss": 3.3888,
3368
+ "step": 280000
3369
+ },
3370
+ {
3371
+ "epoch": 25.3,
3372
+ "learning_rate": 1.4951625e-05,
3373
+ "loss": 3.392,
3374
+ "step": 280500
3375
+ },
3376
+ {
3377
+ "epoch": 25.34,
3378
+ "learning_rate": 1.4889125000000001e-05,
3379
+ "loss": 3.3876,
3380
+ "step": 281000
3381
+ },
3382
+ {
3383
+ "epoch": 25.39,
3384
+ "learning_rate": 1.4826625e-05,
3385
+ "loss": 3.3851,
3386
+ "step": 281500
3387
+ },
3388
+ {
3389
+ "epoch": 25.44,
3390
+ "learning_rate": 1.4764125000000002e-05,
3391
+ "loss": 3.3887,
3392
+ "step": 282000
3393
+ },
3394
+ {
3395
+ "epoch": 25.48,
3396
+ "learning_rate": 1.4701625e-05,
3397
+ "loss": 3.3898,
3398
+ "step": 282500
3399
+ },
3400
+ {
3401
+ "epoch": 25.53,
3402
+ "learning_rate": 1.4639125e-05,
3403
+ "loss": 3.3924,
3404
+ "step": 283000
3405
+ },
3406
+ {
3407
+ "epoch": 25.57,
3408
+ "learning_rate": 1.4576625000000002e-05,
3409
+ "loss": 3.3869,
3410
+ "step": 283500
3411
+ },
3412
+ {
3413
+ "epoch": 25.62,
3414
+ "learning_rate": 1.4514125e-05,
3415
+ "loss": 3.3871,
3416
+ "step": 284000
3417
+ },
3418
+ {
3419
+ "epoch": 25.66,
3420
+ "learning_rate": 1.445175e-05,
3421
+ "loss": 3.3909,
3422
+ "step": 284500
3423
+ },
3424
+ {
3425
+ "epoch": 25.71,
3426
+ "learning_rate": 1.4389250000000001e-05,
3427
+ "loss": 3.3883,
3428
+ "step": 285000
3429
+ },
3430
+ {
3431
+ "epoch": 25.75,
3432
+ "learning_rate": 1.432675e-05,
3433
+ "loss": 3.3925,
3434
+ "step": 285500
3435
+ },
3436
+ {
3437
+ "epoch": 25.8,
3438
+ "learning_rate": 1.4264375000000002e-05,
3439
+ "loss": 3.39,
3440
+ "step": 286000
3441
+ },
3442
+ {
3443
+ "epoch": 25.84,
3444
+ "learning_rate": 1.4201875e-05,
3445
+ "loss": 3.3893,
3446
+ "step": 286500
3447
+ },
3448
+ {
3449
+ "epoch": 25.89,
3450
+ "learning_rate": 1.4139375000000002e-05,
3451
+ "loss": 3.3943,
3452
+ "step": 287000
3453
+ },
3454
+ {
3455
+ "epoch": 25.93,
3456
+ "learning_rate": 1.4076875e-05,
3457
+ "loss": 3.386,
3458
+ "step": 287500
3459
+ },
3460
+ {
3461
+ "epoch": 25.98,
3462
+ "learning_rate": 1.40145e-05,
3463
+ "loss": 3.3857,
3464
+ "step": 288000
3465
+ },
3466
+ {
3467
+ "epoch": 26.02,
3468
+ "learning_rate": 1.3952000000000001e-05,
3469
+ "loss": 3.385,
3470
+ "step": 288500
3471
+ },
3472
+ {
3473
+ "epoch": 26.07,
3474
+ "learning_rate": 1.38895e-05,
3475
+ "loss": 3.3878,
3476
+ "step": 289000
3477
+ },
3478
+ {
3479
+ "epoch": 26.11,
3480
+ "learning_rate": 1.3827000000000002e-05,
3481
+ "loss": 3.385,
3482
+ "step": 289500
3483
+ },
3484
+ {
3485
+ "epoch": 26.16,
3486
+ "learning_rate": 1.37645e-05,
3487
+ "loss": 3.385,
3488
+ "step": 290000
3489
+ },
3490
+ {
3491
+ "epoch": 26.2,
3492
+ "learning_rate": 1.3702125000000002e-05,
3493
+ "loss": 3.3881,
3494
+ "step": 290500
3495
+ },
3496
+ {
3497
+ "epoch": 26.25,
3498
+ "learning_rate": 1.3639625e-05,
3499
+ "loss": 3.3883,
3500
+ "step": 291000
3501
+ },
3502
+ {
3503
+ "epoch": 26.29,
3504
+ "learning_rate": 1.3577125000000001e-05,
3505
+ "loss": 3.3865,
3506
+ "step": 291500
3507
+ },
3508
+ {
3509
+ "epoch": 26.34,
3510
+ "learning_rate": 1.3514625e-05,
3511
+ "loss": 3.3826,
3512
+ "step": 292000
3513
+ },
3514
+ {
3515
+ "epoch": 26.38,
3516
+ "learning_rate": 1.3452125000000001e-05,
3517
+ "loss": 3.3918,
3518
+ "step": 292500
3519
+ },
3520
+ {
3521
+ "epoch": 26.43,
3522
+ "learning_rate": 1.3389625e-05,
3523
+ "loss": 3.3854,
3524
+ "step": 293000
3525
+ },
3526
+ {
3527
+ "epoch": 26.47,
3528
+ "learning_rate": 1.3327125000000002e-05,
3529
+ "loss": 3.3873,
3530
+ "step": 293500
3531
+ },
3532
+ {
3533
+ "epoch": 26.52,
3534
+ "learning_rate": 1.326475e-05,
3535
+ "loss": 3.3906,
3536
+ "step": 294000
3537
+ },
3538
+ {
3539
+ "epoch": 26.56,
3540
+ "learning_rate": 1.3202249999999999e-05,
3541
+ "loss": 3.3904,
3542
+ "step": 294500
3543
+ },
3544
+ {
3545
+ "epoch": 26.61,
3546
+ "learning_rate": 1.313975e-05,
3547
+ "loss": 3.3852,
3548
+ "step": 295000
3549
+ },
3550
+ {
3551
+ "epoch": 26.65,
3552
+ "learning_rate": 1.3077250000000003e-05,
3553
+ "loss": 3.3843,
3554
+ "step": 295500
3555
+ },
3556
+ {
3557
+ "epoch": 26.7,
3558
+ "learning_rate": 1.3014750000000001e-05,
3559
+ "loss": 3.3857,
3560
+ "step": 296000
3561
+ },
3562
+ {
3563
+ "epoch": 26.74,
3564
+ "learning_rate": 1.2952250000000001e-05,
3565
+ "loss": 3.388,
3566
+ "step": 296500
3567
+ },
3568
+ {
3569
+ "epoch": 26.79,
3570
+ "learning_rate": 1.2889875000000002e-05,
3571
+ "loss": 3.3881,
3572
+ "step": 297000
3573
+ },
3574
+ {
3575
+ "epoch": 26.83,
3576
+ "learning_rate": 1.2827375e-05,
3577
+ "loss": 3.3875,
3578
+ "step": 297500
3579
+ },
3580
+ {
3581
+ "epoch": 26.88,
3582
+ "learning_rate": 1.2764875e-05,
3583
+ "loss": 3.3834,
3584
+ "step": 298000
3585
+ },
3586
+ {
3587
+ "epoch": 26.92,
3588
+ "learning_rate": 1.2702375e-05,
3589
+ "loss": 3.3867,
3590
+ "step": 298500
3591
+ },
3592
+ {
3593
+ "epoch": 26.97,
3594
+ "learning_rate": 1.2639875e-05,
3595
+ "loss": 3.3892,
3596
+ "step": 299000
3597
+ },
3598
+ {
3599
+ "epoch": 27.01,
3600
+ "learning_rate": 1.2577375e-05,
3601
+ "loss": 3.3863,
3602
+ "step": 299500
3603
+ },
3604
+ {
3605
+ "epoch": 27.06,
3606
+ "learning_rate": 1.2514875000000001e-05,
3607
+ "loss": 3.3841,
3608
+ "step": 300000
3609
+ },
3610
+ {
3611
+ "epoch": 27.1,
3612
+ "learning_rate": 1.2452375000000001e-05,
3613
+ "loss": 3.3849,
3614
+ "step": 300500
3615
+ },
3616
+ {
3617
+ "epoch": 27.15,
3618
+ "learning_rate": 1.239e-05,
3619
+ "loss": 3.383,
3620
+ "step": 301000
3621
+ },
3622
+ {
3623
+ "epoch": 27.19,
3624
+ "learning_rate": 1.23275e-05,
3625
+ "loss": 3.3832,
3626
+ "step": 301500
3627
+ },
3628
+ {
3629
+ "epoch": 27.24,
3630
+ "learning_rate": 1.2265e-05,
3631
+ "loss": 3.3854,
3632
+ "step": 302000
3633
+ },
3634
+ {
3635
+ "epoch": 27.28,
3636
+ "learning_rate": 1.22025e-05,
3637
+ "loss": 3.3843,
3638
+ "step": 302500
3639
+ },
3640
+ {
3641
+ "epoch": 27.33,
3642
+ "learning_rate": 1.214025e-05,
3643
+ "loss": 3.3839,
3644
+ "step": 303000
3645
+ },
3646
+ {
3647
+ "epoch": 27.37,
3648
+ "learning_rate": 1.207775e-05,
3649
+ "loss": 3.3902,
3650
+ "step": 303500
3651
+ },
3652
+ {
3653
+ "epoch": 27.42,
3654
+ "learning_rate": 1.201525e-05,
3655
+ "loss": 3.3849,
3656
+ "step": 304000
3657
+ },
3658
+ {
3659
+ "epoch": 27.46,
3660
+ "learning_rate": 1.1952875e-05,
3661
+ "loss": 3.3838,
3662
+ "step": 304500
3663
+ },
3664
+ {
3665
+ "epoch": 27.51,
3666
+ "learning_rate": 1.1890375e-05,
3667
+ "loss": 3.3831,
3668
+ "step": 305000
3669
+ },
3670
+ {
3671
+ "epoch": 27.55,
3672
+ "learning_rate": 1.1827875e-05,
3673
+ "loss": 3.387,
3674
+ "step": 305500
3675
+ },
3676
+ {
3677
+ "epoch": 27.6,
3678
+ "learning_rate": 1.1765375e-05,
3679
+ "loss": 3.3852,
3680
+ "step": 306000
3681
+ },
3682
+ {
3683
+ "epoch": 27.64,
3684
+ "learning_rate": 1.1703e-05,
3685
+ "loss": 3.3805,
3686
+ "step": 306500
3687
+ },
3688
+ {
3689
+ "epoch": 27.69,
3690
+ "learning_rate": 1.16405e-05,
3691
+ "loss": 3.385,
3692
+ "step": 307000
3693
+ },
3694
+ {
3695
+ "epoch": 27.74,
3696
+ "learning_rate": 1.1578e-05,
3697
+ "loss": 3.385,
3698
+ "step": 307500
3699
+ },
3700
+ {
3701
+ "epoch": 27.78,
3702
+ "learning_rate": 1.15155e-05,
3703
+ "loss": 3.3858,
3704
+ "step": 308000
3705
+ },
3706
+ {
3707
+ "epoch": 27.83,
3708
+ "learning_rate": 1.1453000000000002e-05,
3709
+ "loss": 3.3839,
3710
+ "step": 308500
3711
+ },
3712
+ {
3713
+ "epoch": 27.87,
3714
+ "learning_rate": 1.13905e-05,
3715
+ "loss": 3.3851,
3716
+ "step": 309000
3717
+ },
3718
+ {
3719
+ "epoch": 27.92,
3720
+ "learning_rate": 1.1328e-05,
3721
+ "loss": 3.3807,
3722
+ "step": 309500
3723
+ },
3724
+ {
3725
+ "epoch": 27.96,
3726
+ "learning_rate": 1.12655e-05,
3727
+ "loss": 3.3885,
3728
+ "step": 310000
3729
+ },
3730
+ {
3731
+ "epoch": 28.01,
3732
+ "learning_rate": 1.1203125000000001e-05,
3733
+ "loss": 3.3837,
3734
+ "step": 310500
3735
+ },
3736
+ {
3737
+ "epoch": 28.05,
3738
+ "learning_rate": 1.1140625e-05,
3739
+ "loss": 3.3781,
3740
+ "step": 311000
3741
+ },
3742
+ {
3743
+ "epoch": 28.1,
3744
+ "learning_rate": 1.1078125e-05,
3745
+ "loss": 3.382,
3746
+ "step": 311500
3747
+ },
3748
+ {
3749
+ "epoch": 28.14,
3750
+ "learning_rate": 1.1015625e-05,
3751
+ "loss": 3.3815,
3752
+ "step": 312000
3753
+ },
3754
+ {
3755
+ "epoch": 28.19,
3756
+ "learning_rate": 1.0953125e-05,
3757
+ "loss": 3.3853,
3758
+ "step": 312500
3759
+ },
3760
+ {
3761
+ "epoch": 28.23,
3762
+ "learning_rate": 1.0890625e-05,
3763
+ "loss": 3.3849,
3764
+ "step": 313000
3765
+ },
3766
+ {
3767
+ "epoch": 28.28,
3768
+ "learning_rate": 1.0828250000000001e-05,
3769
+ "loss": 3.3825,
3770
+ "step": 313500
3771
+ },
3772
+ {
3773
+ "epoch": 28.32,
3774
+ "learning_rate": 1.0765875e-05,
3775
+ "loss": 3.3811,
3776
+ "step": 314000
3777
+ },
3778
+ {
3779
+ "epoch": 28.37,
3780
+ "learning_rate": 1.0703375e-05,
3781
+ "loss": 3.3775,
3782
+ "step": 314500
3783
+ },
3784
+ {
3785
+ "epoch": 28.41,
3786
+ "learning_rate": 1.0640875e-05,
3787
+ "loss": 3.3851,
3788
+ "step": 315000
3789
+ },
3790
+ {
3791
+ "epoch": 28.46,
3792
+ "learning_rate": 1.0578375e-05,
3793
+ "loss": 3.388,
3794
+ "step": 315500
3795
+ },
3796
+ {
3797
+ "epoch": 28.5,
3798
+ "learning_rate": 1.0515875e-05,
3799
+ "loss": 3.3806,
3800
+ "step": 316000
3801
+ },
3802
+ {
3803
+ "epoch": 28.55,
3804
+ "learning_rate": 1.0453375e-05,
3805
+ "loss": 3.3841,
3806
+ "step": 316500
3807
+ },
3808
+ {
3809
+ "epoch": 28.59,
3810
+ "learning_rate": 1.0390875e-05,
3811
+ "loss": 3.3856,
3812
+ "step": 317000
3813
+ },
3814
+ {
3815
+ "epoch": 28.64,
3816
+ "learning_rate": 1.0328375e-05,
3817
+ "loss": 3.3835,
3818
+ "step": 317500
3819
+ },
3820
+ {
3821
+ "epoch": 28.68,
3822
+ "learning_rate": 1.0265875e-05,
3823
+ "loss": 3.38,
3824
+ "step": 318000
3825
+ },
3826
+ {
3827
+ "epoch": 28.73,
3828
+ "learning_rate": 1.02035e-05,
3829
+ "loss": 3.3829,
3830
+ "step": 318500
3831
+ },
3832
+ {
3833
+ "epoch": 28.77,
3834
+ "learning_rate": 1.0141e-05,
3835
+ "loss": 3.3794,
3836
+ "step": 319000
3837
+ },
3838
+ {
3839
+ "epoch": 28.82,
3840
+ "learning_rate": 1.0078625e-05,
3841
+ "loss": 3.3837,
3842
+ "step": 319500
3843
+ },
3844
+ {
3845
+ "epoch": 28.86,
3846
+ "learning_rate": 1.0016125e-05,
3847
+ "loss": 3.3798,
3848
+ "step": 320000
3849
+ },
3850
+ {
3851
+ "epoch": 28.91,
3852
+ "learning_rate": 9.953625e-06,
3853
+ "loss": 3.3836,
3854
+ "step": 320500
3855
+ },
3856
+ {
3857
+ "epoch": 28.95,
3858
+ "learning_rate": 9.891250000000001e-06,
3859
+ "loss": 3.3841,
3860
+ "step": 321000
3861
+ },
3862
+ {
3863
+ "epoch": 29.0,
3864
+ "learning_rate": 9.82875e-06,
3865
+ "loss": 3.38,
3866
+ "step": 321500
3867
+ },
3868
+ {
3869
+ "epoch": 29.04,
3870
+ "learning_rate": 9.76625e-06,
3871
+ "loss": 3.375,
3872
+ "step": 322000
3873
+ },
3874
+ {
3875
+ "epoch": 29.09,
3876
+ "learning_rate": 9.70375e-06,
3877
+ "loss": 3.3763,
3878
+ "step": 322500
3879
+ },
3880
+ {
3881
+ "epoch": 29.13,
3882
+ "learning_rate": 9.64125e-06,
3883
+ "loss": 3.3831,
3884
+ "step": 323000
3885
+ },
3886
+ {
3887
+ "epoch": 29.18,
3888
+ "learning_rate": 9.57875e-06,
3889
+ "loss": 3.3844,
3890
+ "step": 323500
3891
+ },
3892
+ {
3893
+ "epoch": 29.22,
3894
+ "learning_rate": 9.51625e-06,
3895
+ "loss": 3.3822,
3896
+ "step": 324000
3897
+ },
3898
+ {
3899
+ "epoch": 29.27,
3900
+ "learning_rate": 9.45375e-06,
3901
+ "loss": 3.3815,
3902
+ "step": 324500
3903
+ },
3904
+ {
3905
+ "epoch": 29.31,
3906
+ "learning_rate": 9.39125e-06,
3907
+ "loss": 3.382,
3908
+ "step": 325000
3909
+ },
3910
+ {
3911
+ "epoch": 29.36,
3912
+ "learning_rate": 9.32875e-06,
3913
+ "loss": 3.375,
3914
+ "step": 325500
3915
+ },
3916
+ {
3917
+ "epoch": 29.4,
3918
+ "learning_rate": 9.26625e-06,
3919
+ "loss": 3.3796,
3920
+ "step": 326000
3921
+ },
3922
+ {
3923
+ "epoch": 29.45,
3924
+ "learning_rate": 9.20375e-06,
3925
+ "loss": 3.3799,
3926
+ "step": 326500
3927
+ },
3928
+ {
3929
+ "epoch": 29.49,
3930
+ "learning_rate": 9.14125e-06,
3931
+ "loss": 3.3861,
3932
+ "step": 327000
3933
+ },
3934
+ {
3935
+ "epoch": 29.54,
3936
+ "learning_rate": 9.078875e-06,
3937
+ "loss": 3.3837,
3938
+ "step": 327500
3939
+ },
3940
+ {
3941
+ "epoch": 29.58,
3942
+ "learning_rate": 9.016375e-06,
3943
+ "loss": 3.3795,
3944
+ "step": 328000
3945
+ },
3946
+ {
3947
+ "epoch": 29.63,
3948
+ "learning_rate": 8.953875e-06,
3949
+ "loss": 3.3787,
3950
+ "step": 328500
3951
+ },
3952
+ {
3953
+ "epoch": 29.67,
3954
+ "learning_rate": 8.891375000000001e-06,
3955
+ "loss": 3.3819,
3956
+ "step": 329000
3957
+ },
3958
+ {
3959
+ "epoch": 29.72,
3960
+ "learning_rate": 8.829e-06,
3961
+ "loss": 3.3802,
3962
+ "step": 329500
3963
+ },
3964
+ {
3965
+ "epoch": 29.76,
3966
+ "learning_rate": 8.766625e-06,
3967
+ "loss": 3.383,
3968
+ "step": 330000
3969
+ },
3970
+ {
3971
+ "epoch": 29.81,
3972
+ "learning_rate": 8.704125e-06,
3973
+ "loss": 3.3832,
3974
+ "step": 330500
3975
+ },
3976
+ {
3977
+ "epoch": 29.85,
3978
+ "learning_rate": 8.641625e-06,
3979
+ "loss": 3.3784,
3980
+ "step": 331000
3981
+ },
3982
+ {
3983
+ "epoch": 29.9,
3984
+ "learning_rate": 8.579125e-06,
3985
+ "loss": 3.3788,
3986
+ "step": 331500
3987
+ },
3988
+ {
3989
+ "epoch": 29.94,
3990
+ "learning_rate": 8.516625e-06,
3991
+ "loss": 3.3797,
3992
+ "step": 332000
3993
+ },
3994
+ {
3995
+ "epoch": 29.99,
3996
+ "learning_rate": 8.454125000000001e-06,
3997
+ "loss": 3.3809,
3998
+ "step": 332500
3999
+ },
4000
+ {
4001
+ "epoch": 30.04,
4002
+ "learning_rate": 8.391625e-06,
4003
+ "loss": 3.3792,
4004
+ "step": 333000
4005
+ },
4006
+ {
4007
+ "epoch": 30.08,
4008
+ "learning_rate": 8.329125e-06,
4009
+ "loss": 3.3784,
4010
+ "step": 333500
4011
+ },
4012
+ {
4013
+ "epoch": 30.13,
4014
+ "learning_rate": 8.266625e-06,
4015
+ "loss": 3.377,
4016
+ "step": 334000
4017
+ },
4018
+ {
4019
+ "epoch": 30.17,
4020
+ "learning_rate": 8.204125e-06,
4021
+ "loss": 3.3774,
4022
+ "step": 334500
4023
+ },
4024
+ {
4025
+ "epoch": 30.22,
4026
+ "learning_rate": 8.14175e-06,
4027
+ "loss": 3.3822,
4028
+ "step": 335000
4029
+ },
4030
+ {
4031
+ "epoch": 30.26,
4032
+ "learning_rate": 8.07925e-06,
4033
+ "loss": 3.382,
4034
+ "step": 335500
4035
+ },
4036
+ {
4037
+ "epoch": 30.31,
4038
+ "learning_rate": 8.01675e-06,
4039
+ "loss": 3.3753,
4040
+ "step": 336000
4041
+ },
4042
+ {
4043
+ "epoch": 30.35,
4044
+ "learning_rate": 7.954250000000001e-06,
4045
+ "loss": 3.3784,
4046
+ "step": 336500
4047
+ },
4048
+ {
4049
+ "epoch": 30.4,
4050
+ "learning_rate": 7.891750000000001e-06,
4051
+ "loss": 3.3807,
4052
+ "step": 337000
4053
+ },
4054
+ {
4055
+ "epoch": 30.44,
4056
+ "learning_rate": 7.829375e-06,
4057
+ "loss": 3.3787,
4058
+ "step": 337500
4059
+ },
4060
+ {
4061
+ "epoch": 30.49,
4062
+ "learning_rate": 7.766875e-06,
4063
+ "loss": 3.3751,
4064
+ "step": 338000
4065
+ },
4066
+ {
4067
+ "epoch": 30.53,
4068
+ "learning_rate": 7.704375e-06,
4069
+ "loss": 3.3808,
4070
+ "step": 338500
4071
+ },
4072
+ {
4073
+ "epoch": 30.58,
4074
+ "learning_rate": 7.641875e-06,
4075
+ "loss": 3.3817,
4076
+ "step": 339000
4077
+ },
4078
+ {
4079
+ "epoch": 30.62,
4080
+ "learning_rate": 7.579375e-06,
4081
+ "loss": 3.3808,
4082
+ "step": 339500
4083
+ },
4084
+ {
4085
+ "epoch": 30.67,
4086
+ "learning_rate": 7.517e-06,
4087
+ "loss": 3.3784,
4088
+ "step": 340000
4089
+ },
4090
+ {
4091
+ "epoch": 30.71,
4092
+ "learning_rate": 7.4545e-06,
4093
+ "loss": 3.3774,
4094
+ "step": 340500
4095
+ },
4096
+ {
4097
+ "epoch": 30.76,
4098
+ "learning_rate": 7.3920000000000005e-06,
4099
+ "loss": 3.3788,
4100
+ "step": 341000
4101
+ },
4102
+ {
4103
+ "epoch": 30.8,
4104
+ "learning_rate": 7.3295e-06,
4105
+ "loss": 3.377,
4106
+ "step": 341500
4107
+ },
4108
+ {
4109
+ "epoch": 30.85,
4110
+ "learning_rate": 7.267e-06,
4111
+ "loss": 3.3779,
4112
+ "step": 342000
4113
+ },
4114
+ {
4115
+ "epoch": 30.89,
4116
+ "learning_rate": 7.2045e-06,
4117
+ "loss": 3.3786,
4118
+ "step": 342500
4119
+ },
4120
+ {
4121
+ "epoch": 30.94,
4122
+ "learning_rate": 7.142e-06,
4123
+ "loss": 3.3765,
4124
+ "step": 343000
4125
+ },
4126
+ {
4127
+ "epoch": 30.98,
4128
+ "learning_rate": 7.0795e-06,
4129
+ "loss": 3.3755,
4130
+ "step": 343500
4131
+ },
4132
+ {
4133
+ "epoch": 31.03,
4134
+ "learning_rate": 7.017125000000001e-06,
4135
+ "loss": 3.3787,
4136
+ "step": 344000
4137
+ },
4138
+ {
4139
+ "epoch": 31.07,
4140
+ "learning_rate": 6.954625e-06,
4141
+ "loss": 3.3739,
4142
+ "step": 344500
4143
+ },
4144
+ {
4145
+ "epoch": 31.12,
4146
+ "learning_rate": 6.89225e-06,
4147
+ "loss": 3.3776,
4148
+ "step": 345000
4149
+ },
4150
+ {
4151
+ "epoch": 31.16,
4152
+ "learning_rate": 6.82975e-06,
4153
+ "loss": 3.3787,
4154
+ "step": 345500
4155
+ },
4156
+ {
4157
+ "epoch": 31.21,
4158
+ "learning_rate": 6.76725e-06,
4159
+ "loss": 3.3725,
4160
+ "step": 346000
4161
+ },
4162
+ {
4163
+ "epoch": 31.25,
4164
+ "learning_rate": 6.70475e-06,
4165
+ "loss": 3.3765,
4166
+ "step": 346500
4167
+ },
4168
+ {
4169
+ "epoch": 31.3,
4170
+ "learning_rate": 6.642375000000001e-06,
4171
+ "loss": 3.3796,
4172
+ "step": 347000
4173
+ },
4174
+ {
4175
+ "epoch": 31.34,
4176
+ "learning_rate": 6.579875e-06,
4177
+ "loss": 3.3784,
4178
+ "step": 347500
4179
+ },
4180
+ {
4181
+ "epoch": 31.39,
4182
+ "learning_rate": 6.5173750000000004e-06,
4183
+ "loss": 3.3795,
4184
+ "step": 348000
4185
+ },
4186
+ {
4187
+ "epoch": 31.43,
4188
+ "learning_rate": 6.454875000000001e-06,
4189
+ "loss": 3.3788,
4190
+ "step": 348500
4191
+ },
4192
+ {
4193
+ "epoch": 31.48,
4194
+ "learning_rate": 6.3925e-06,
4195
+ "loss": 3.3806,
4196
+ "step": 349000
4197
+ },
4198
+ {
4199
+ "epoch": 31.52,
4200
+ "learning_rate": 6.3299999999999995e-06,
4201
+ "loss": 3.3773,
4202
+ "step": 349500
4203
+ },
4204
+ {
4205
+ "epoch": 31.57,
4206
+ "learning_rate": 6.2675e-06,
4207
+ "loss": 3.3744,
4208
+ "step": 350000
4209
+ },
4210
+ {
4211
+ "epoch": 31.61,
4212
+ "learning_rate": 6.205000000000001e-06,
4213
+ "loss": 3.378,
4214
+ "step": 350500
4215
+ },
4216
+ {
4217
+ "epoch": 31.66,
4218
+ "learning_rate": 6.1425e-06,
4219
+ "loss": 3.3752,
4220
+ "step": 351000
4221
+ },
4222
+ {
4223
+ "epoch": 31.7,
4224
+ "learning_rate": 6.08e-06,
4225
+ "loss": 3.3774,
4226
+ "step": 351500
4227
+ },
4228
+ {
4229
+ "epoch": 31.75,
4230
+ "learning_rate": 6.0175e-06,
4231
+ "loss": 3.3778,
4232
+ "step": 352000
4233
+ },
4234
+ {
4235
+ "epoch": 31.79,
4236
+ "learning_rate": 5.955000000000001e-06,
4237
+ "loss": 3.3737,
4238
+ "step": 352500
4239
+ },
4240
+ {
4241
+ "epoch": 31.84,
4242
+ "learning_rate": 5.892625e-06,
4243
+ "loss": 3.3813,
4244
+ "step": 353000
4245
+ },
4246
+ {
4247
+ "epoch": 31.88,
4248
+ "learning_rate": 5.830250000000001e-06,
4249
+ "loss": 3.3745,
4250
+ "step": 353500
4251
+ },
4252
+ {
4253
+ "epoch": 31.93,
4254
+ "learning_rate": 5.76775e-06,
4255
+ "loss": 3.3786,
4256
+ "step": 354000
4257
+ },
4258
+ {
4259
+ "epoch": 31.97,
4260
+ "learning_rate": 5.70525e-06,
4261
+ "loss": 3.3741,
4262
+ "step": 354500
4263
+ },
4264
+ {
4265
+ "epoch": 32.02,
4266
+ "learning_rate": 5.64275e-06,
4267
+ "loss": 3.374,
4268
+ "step": 355000
4269
+ },
4270
+ {
4271
+ "epoch": 32.06,
4272
+ "learning_rate": 5.58025e-06,
4273
+ "loss": 3.3763,
4274
+ "step": 355500
4275
+ },
4276
+ {
4277
+ "epoch": 32.11,
4278
+ "learning_rate": 5.51775e-06,
4279
+ "loss": 3.3787,
4280
+ "step": 356000
4281
+ },
4282
+ {
4283
+ "epoch": 32.15,
4284
+ "learning_rate": 5.45525e-06,
4285
+ "loss": 3.3719,
4286
+ "step": 356500
4287
+ },
4288
+ {
4289
+ "epoch": 32.2,
4290
+ "learning_rate": 5.39275e-06,
4291
+ "loss": 3.3775,
4292
+ "step": 357000
4293
+ },
4294
+ {
4295
+ "epoch": 32.24,
4296
+ "learning_rate": 5.33025e-06,
4297
+ "loss": 3.3785,
4298
+ "step": 357500
4299
+ },
4300
+ {
4301
+ "epoch": 32.29,
4302
+ "learning_rate": 5.2677500000000005e-06,
4303
+ "loss": 3.3763,
4304
+ "step": 358000
4305
+ },
4306
+ {
4307
+ "epoch": 32.34,
4308
+ "learning_rate": 5.205375e-06,
4309
+ "loss": 3.3768,
4310
+ "step": 358500
4311
+ },
4312
+ {
4313
+ "epoch": 32.38,
4314
+ "learning_rate": 5.142875e-06,
4315
+ "loss": 3.3751,
4316
+ "step": 359000
4317
+ },
4318
+ {
4319
+ "epoch": 32.43,
4320
+ "learning_rate": 5.0803750000000005e-06,
4321
+ "loss": 3.3749,
4322
+ "step": 359500
4323
+ },
4324
+ {
4325
+ "epoch": 32.47,
4326
+ "learning_rate": 5.017875000000001e-06,
4327
+ "loss": 3.377,
4328
+ "step": 360000
4329
+ },
4330
+ {
4331
+ "epoch": 32.52,
4332
+ "learning_rate": 4.9555e-06,
4333
+ "loss": 3.3744,
4334
+ "step": 360500
4335
+ },
4336
+ {
4337
+ "epoch": 32.56,
4338
+ "learning_rate": 4.893e-06,
4339
+ "loss": 3.3732,
4340
+ "step": 361000
4341
+ },
4342
+ {
4343
+ "epoch": 32.61,
4344
+ "learning_rate": 4.830500000000001e-06,
4345
+ "loss": 3.374,
4346
+ "step": 361500
4347
+ },
4348
+ {
4349
+ "epoch": 32.65,
4350
+ "learning_rate": 4.768e-06,
4351
+ "loss": 3.3755,
4352
+ "step": 362000
4353
+ },
4354
+ {
4355
+ "epoch": 32.7,
4356
+ "learning_rate": 4.7055e-06,
4357
+ "loss": 3.3753,
4358
+ "step": 362500
4359
+ },
4360
+ {
4361
+ "epoch": 32.74,
4362
+ "learning_rate": 4.643e-06,
4363
+ "loss": 3.38,
4364
+ "step": 363000
4365
+ },
4366
+ {
4367
+ "epoch": 32.79,
4368
+ "learning_rate": 4.5805000000000004e-06,
4369
+ "loss": 3.3785,
4370
+ "step": 363500
4371
+ },
4372
+ {
4373
+ "epoch": 32.83,
4374
+ "learning_rate": 4.518e-06,
4375
+ "loss": 3.3712,
4376
+ "step": 364000
4377
+ },
4378
+ {
4379
+ "epoch": 32.88,
4380
+ "learning_rate": 4.4555e-06,
4381
+ "loss": 3.3703,
4382
+ "step": 364500
4383
+ },
4384
+ {
4385
+ "epoch": 32.92,
4386
+ "learning_rate": 4.393125e-06,
4387
+ "loss": 3.375,
4388
+ "step": 365000
4389
+ },
4390
+ {
4391
+ "epoch": 32.97,
4392
+ "learning_rate": 4.330625e-06,
4393
+ "loss": 3.3708,
4394
+ "step": 365500
4395
+ },
4396
+ {
4397
+ "epoch": 33.01,
4398
+ "learning_rate": 4.26825e-06,
4399
+ "loss": 3.3748,
4400
+ "step": 366000
4401
+ },
4402
+ {
4403
+ "epoch": 33.06,
4404
+ "learning_rate": 4.20575e-06,
4405
+ "loss": 3.3734,
4406
+ "step": 366500
4407
+ },
4408
+ {
4409
+ "epoch": 33.1,
4410
+ "learning_rate": 4.14325e-06,
4411
+ "loss": 3.3797,
4412
+ "step": 367000
4413
+ },
4414
+ {
4415
+ "epoch": 33.15,
4416
+ "learning_rate": 4.08075e-06,
4417
+ "loss": 3.3736,
4418
+ "step": 367500
4419
+ },
4420
+ {
4421
+ "epoch": 33.19,
4422
+ "learning_rate": 4.018250000000001e-06,
4423
+ "loss": 3.374,
4424
+ "step": 368000
4425
+ },
4426
+ {
4427
+ "epoch": 33.24,
4428
+ "learning_rate": 3.95575e-06,
4429
+ "loss": 3.374,
4430
+ "step": 368500
4431
+ },
4432
+ {
4433
+ "epoch": 33.28,
4434
+ "learning_rate": 3.89325e-06,
4435
+ "loss": 3.3729,
4436
+ "step": 369000
4437
+ },
4438
+ {
4439
+ "epoch": 33.33,
4440
+ "learning_rate": 3.830875000000001e-06,
4441
+ "loss": 3.3742,
4442
+ "step": 369500
4443
+ },
4444
+ {
4445
+ "epoch": 33.37,
4446
+ "learning_rate": 3.7683750000000006e-06,
4447
+ "loss": 3.3769,
4448
+ "step": 370000
4449
+ },
4450
+ {
4451
+ "epoch": 33.42,
4452
+ "learning_rate": 3.7058750000000003e-06,
4453
+ "loss": 3.3745,
4454
+ "step": 370500
4455
+ },
4456
+ {
4457
+ "epoch": 33.46,
4458
+ "learning_rate": 3.643375e-06,
4459
+ "loss": 3.3769,
4460
+ "step": 371000
4461
+ },
4462
+ {
4463
+ "epoch": 33.51,
4464
+ "learning_rate": 3.5808750000000003e-06,
4465
+ "loss": 3.3707,
4466
+ "step": 371500
4467
+ },
4468
+ {
4469
+ "epoch": 33.55,
4470
+ "learning_rate": 3.518375e-06,
4471
+ "loss": 3.3746,
4472
+ "step": 372000
4473
+ },
4474
+ {
4475
+ "epoch": 33.6,
4476
+ "learning_rate": 3.455875e-06,
4477
+ "loss": 3.3734,
4478
+ "step": 372500
4479
+ },
4480
+ {
4481
+ "epoch": 33.64,
4482
+ "learning_rate": 3.3935e-06,
4483
+ "loss": 3.3732,
4484
+ "step": 373000
4485
+ },
4486
+ {
4487
+ "epoch": 33.69,
4488
+ "learning_rate": 3.331e-06,
4489
+ "loss": 3.3715,
4490
+ "step": 373500
4491
+ },
4492
+ {
4493
+ "epoch": 33.73,
4494
+ "learning_rate": 3.2685e-06,
4495
+ "loss": 3.3767,
4496
+ "step": 374000
4497
+ },
4498
+ {
4499
+ "epoch": 33.78,
4500
+ "learning_rate": 3.206e-06,
4501
+ "loss": 3.3736,
4502
+ "step": 374500
4503
+ },
4504
+ {
4505
+ "epoch": 33.82,
4506
+ "learning_rate": 3.1435e-06,
4507
+ "loss": 3.3745,
4508
+ "step": 375000
4509
+ },
4510
+ {
4511
+ "epoch": 33.87,
4512
+ "learning_rate": 3.0810000000000002e-06,
4513
+ "loss": 3.3703,
4514
+ "step": 375500
4515
+ },
4516
+ {
4517
+ "epoch": 33.91,
4518
+ "learning_rate": 3.0185e-06,
4519
+ "loss": 3.379,
4520
+ "step": 376000
4521
+ },
4522
+ {
4523
+ "epoch": 33.96,
4524
+ "learning_rate": 2.956e-06,
4525
+ "loss": 3.3694,
4526
+ "step": 376500
4527
+ },
4528
+ {
4529
+ "epoch": 34.0,
4530
+ "learning_rate": 2.8935e-06,
4531
+ "loss": 3.3717,
4532
+ "step": 377000
4533
+ },
4534
+ {
4535
+ "epoch": 34.05,
4536
+ "learning_rate": 2.831e-06,
4537
+ "loss": 3.3719,
4538
+ "step": 377500
4539
+ },
4540
+ {
4541
+ "epoch": 34.09,
4542
+ "learning_rate": 2.7685000000000003e-06,
4543
+ "loss": 3.3733,
4544
+ "step": 378000
4545
+ },
4546
+ {
4547
+ "epoch": 34.14,
4548
+ "learning_rate": 2.706e-06,
4549
+ "loss": 3.3738,
4550
+ "step": 378500
4551
+ },
4552
+ {
4553
+ "epoch": 34.18,
4554
+ "learning_rate": 2.6436250000000005e-06,
4555
+ "loss": 3.3772,
4556
+ "step": 379000
4557
+ },
4558
+ {
4559
+ "epoch": 34.23,
4560
+ "learning_rate": 2.581125e-06,
4561
+ "loss": 3.368,
4562
+ "step": 379500
4563
+ },
4564
+ {
4565
+ "epoch": 34.27,
4566
+ "learning_rate": 2.518625e-06,
4567
+ "loss": 3.3728,
4568
+ "step": 380000
4569
+ },
4570
+ {
4571
+ "epoch": 34.32,
4572
+ "learning_rate": 2.456125e-06,
4573
+ "loss": 3.3764,
4574
+ "step": 380500
4575
+ },
4576
+ {
4577
+ "epoch": 34.36,
4578
+ "learning_rate": 2.393625e-06,
4579
+ "loss": 3.3724,
4580
+ "step": 381000
4581
+ },
4582
+ {
4583
+ "epoch": 34.41,
4584
+ "learning_rate": 2.331125e-06,
4585
+ "loss": 3.3744,
4586
+ "step": 381500
4587
+ },
4588
+ {
4589
+ "epoch": 34.45,
4590
+ "learning_rate": 2.2686250000000002e-06,
4591
+ "loss": 3.3719,
4592
+ "step": 382000
4593
+ },
4594
+ {
4595
+ "epoch": 34.5,
4596
+ "learning_rate": 2.2061250000000004e-06,
4597
+ "loss": 3.3702,
4598
+ "step": 382500
4599
+ },
4600
+ {
4601
+ "epoch": 34.54,
4602
+ "learning_rate": 2.143875e-06,
4603
+ "loss": 3.3728,
4604
+ "step": 383000
4605
+ },
4606
+ {
4607
+ "epoch": 34.59,
4608
+ "learning_rate": 2.081375e-06,
4609
+ "loss": 3.3709,
4610
+ "step": 383500
4611
+ },
4612
+ {
4613
+ "epoch": 34.64,
4614
+ "learning_rate": 2.018875e-06,
4615
+ "loss": 3.3701,
4616
+ "step": 384000
4617
+ },
4618
+ {
4619
+ "epoch": 34.68,
4620
+ "learning_rate": 1.9563750000000004e-06,
4621
+ "loss": 3.3747,
4622
+ "step": 384500
4623
+ },
4624
+ {
4625
+ "epoch": 34.73,
4626
+ "learning_rate": 1.894e-06,
4627
+ "loss": 3.373,
4628
+ "step": 385000
4629
+ },
4630
+ {
4631
+ "epoch": 34.77,
4632
+ "learning_rate": 1.8315000000000002e-06,
4633
+ "loss": 3.3739,
4634
+ "step": 385500
4635
+ },
4636
+ {
4637
+ "epoch": 34.82,
4638
+ "learning_rate": 1.7690000000000001e-06,
4639
+ "loss": 3.3696,
4640
+ "step": 386000
4641
+ },
4642
+ {
4643
+ "epoch": 34.86,
4644
+ "learning_rate": 1.7065e-06,
4645
+ "loss": 3.3722,
4646
+ "step": 386500
4647
+ },
4648
+ {
4649
+ "epoch": 34.91,
4650
+ "learning_rate": 1.644e-06,
4651
+ "loss": 3.3705,
4652
+ "step": 387000
4653
+ },
4654
+ {
4655
+ "epoch": 34.95,
4656
+ "learning_rate": 1.5815e-06,
4657
+ "loss": 3.3753,
4658
+ "step": 387500
4659
+ },
4660
+ {
4661
+ "epoch": 35.0,
4662
+ "learning_rate": 1.5190000000000002e-06,
4663
+ "loss": 3.3714,
4664
+ "step": 388000
4665
+ },
4666
+ {
4667
+ "epoch": 35.04,
4668
+ "learning_rate": 1.4565e-06,
4669
+ "loss": 3.3704,
4670
+ "step": 388500
4671
+ },
4672
+ {
4673
+ "epoch": 35.09,
4674
+ "learning_rate": 1.394e-06,
4675
+ "loss": 3.3735,
4676
+ "step": 389000
4677
+ },
4678
+ {
4679
+ "epoch": 35.13,
4680
+ "learning_rate": 1.3315e-06,
4681
+ "loss": 3.3706,
4682
+ "step": 389500
4683
+ },
4684
+ {
4685
+ "epoch": 35.18,
4686
+ "learning_rate": 1.269e-06,
4687
+ "loss": 3.3742,
4688
+ "step": 390000
4689
+ },
4690
+ {
4691
+ "epoch": 35.22,
4692
+ "learning_rate": 1.2067500000000001e-06,
4693
+ "loss": 3.3698,
4694
+ "step": 390500
4695
+ },
4696
+ {
4697
+ "epoch": 35.27,
4698
+ "learning_rate": 1.14425e-06,
4699
+ "loss": 3.3695,
4700
+ "step": 391000
4701
+ },
4702
+ {
4703
+ "epoch": 35.31,
4704
+ "learning_rate": 1.08175e-06,
4705
+ "loss": 3.3717,
4706
+ "step": 391500
4707
+ },
4708
+ {
4709
+ "epoch": 35.36,
4710
+ "learning_rate": 1.01925e-06,
4711
+ "loss": 3.375,
4712
+ "step": 392000
4713
+ },
4714
+ {
4715
+ "epoch": 35.4,
4716
+ "learning_rate": 9.5675e-07,
4717
+ "loss": 3.3746,
4718
+ "step": 392500
4719
+ },
4720
+ {
4721
+ "epoch": 35.45,
4722
+ "learning_rate": 8.943750000000001e-07,
4723
+ "loss": 3.3715,
4724
+ "step": 393000
4725
+ },
4726
+ {
4727
+ "epoch": 35.49,
4728
+ "learning_rate": 8.31875e-07,
4729
+ "loss": 3.3717,
4730
+ "step": 393500
4731
+ },
4732
+ {
4733
+ "epoch": 35.54,
4734
+ "learning_rate": 7.69375e-07,
4735
+ "loss": 3.3714,
4736
+ "step": 394000
4737
+ },
4738
+ {
4739
+ "epoch": 35.58,
4740
+ "learning_rate": 7.068750000000001e-07,
4741
+ "loss": 3.3704,
4742
+ "step": 394500
4743
+ },
4744
+ {
4745
+ "epoch": 35.63,
4746
+ "learning_rate": 6.44375e-07,
4747
+ "loss": 3.3733,
4748
+ "step": 395000
4749
+ },
4750
+ {
4751
+ "epoch": 35.67,
4752
+ "learning_rate": 5.81875e-07,
4753
+ "loss": 3.3687,
4754
+ "step": 395500
4755
+ },
4756
+ {
4757
+ "epoch": 35.72,
4758
+ "learning_rate": 5.19375e-07,
4759
+ "loss": 3.3722,
4760
+ "step": 396000
4761
+ },
4762
+ {
4763
+ "epoch": 35.76,
4764
+ "learning_rate": 4.56875e-07,
4765
+ "loss": 3.375,
4766
+ "step": 396500
4767
+ },
4768
+ {
4769
+ "epoch": 35.81,
4770
+ "learning_rate": 3.9437500000000004e-07,
4771
+ "loss": 3.3661,
4772
+ "step": 397000
4773
+ },
4774
+ {
4775
+ "epoch": 35.85,
4776
+ "learning_rate": 3.31875e-07,
4777
+ "loss": 3.3716,
4778
+ "step": 397500
4779
+ },
4780
+ {
4781
+ "epoch": 35.9,
4782
+ "learning_rate": 2.695e-07,
4783
+ "loss": 3.3689,
4784
+ "step": 398000
4785
+ },
4786
+ {
4787
+ "epoch": 35.94,
4788
+ "learning_rate": 2.07e-07,
4789
+ "loss": 3.3683,
4790
+ "step": 398500
4791
+ },
4792
+ {
4793
+ "epoch": 35.99,
4794
+ "learning_rate": 1.4450000000000003e-07,
4795
+ "loss": 3.3696,
4796
+ "step": 399000
4797
+ },
4798
+ {
4799
+ "epoch": 36.03,
4800
+ "learning_rate": 8.2e-08,
4801
+ "loss": 3.3741,
4802
+ "step": 399500
4803
+ },
4804
+ {
4805
+ "epoch": 36.08,
4806
+ "learning_rate": 1.9625000000000002e-08,
4807
+ "loss": 3.3696,
4808
+ "step": 400000
4809
+ },
4810
+ {
4811
+ "epoch": 36.08,
4812
+ "step": 400000,
4813
+ "total_flos": 2.3048832150195405e+17,
4814
+ "train_loss": 3.451316015625,
4815
+ "train_runtime": 66733.7733,
4816
+ "train_samples_per_second": 239.759,
4817
+ "train_steps_per_second": 5.994
4818
+ }
4819
+ ],
4820
+ "max_steps": 400000,
4821
+ "num_train_epochs": 37,
4822
+ "total_flos": 2.3048832150195405e+17,
4823
+ "trial_name": null,
4824
+ "trial_params": null
4825
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb34dff58822f9e32757ca6f4073b0ee8336279b8ad2ab64fbaa9262a1ce16c6
3
+ size 3515
vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
vocab.txt ADDED
The diff for this file is too large to render. See raw diff