Hugging Face's logo

bibekyess
/
qcpg-parabk2-t5-base

Text2Text Generation
Transformers
PyTorch
t5
text-generation-inference
Inference Endpoints
Model card Files Community
bibekyess commited on
Commit
0efcaba
•
1 Parent(s): f9423e3

Uploads demo qcpg for testing

Browse files
Files changed (8) hide show
  1. added_tokens.json +1 -0
  2. config.json +57 -0
  3. pytorch_model.bin +3 -0
  4. special_tokens_map.json +1 -0
  5. spiece.model +3 -0
  6. tokenizer.json +0 -0
  7. tokenizer_config.json +1 -0
  8. trainer_state.json +4024 -0
added_tokens.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"COND_SEMANTIC_SIM_45": 32129, "COND_LEXICAL_DIV_75": 32116, "COND_SYNTACTIC_DIV_80": 32157, "COND_SEMANTIC_SIM_15": 32123, "COND_SEMANTIC_SIM_80": 32137, "COND_LEXICAL_DIV_85": 32118, "COND_SYNTACTIC_DIV_60": 32153, "COND_SYNTACTIC_DIV_40": 32148, "COND_SYNTACTIC_DIV_10": 32142, "COND_SYNTACTIC_DIV_20": 32144, "COND_SEMANTIC_SIM_25": 32125, "COND_SEMANTIC_SIM_30": 32126, "COND_SYNTACTIC_DIV_45": 32149, "COND_LEXICAL_DIV_30": 32106, "COND_LEXICAL_DIV_20": 32104, "COND_LEXICAL_DIV_55": 32112, "COND_SYNTACTIC_DIV_75": 32156, "COND_LEXICAL_DIV_35": 32107, "COND_LEXICAL_DIV_5": 32110, "COND_SEMANTIC_SIM_95": 32140, "COND_LEXICAL_DIV_0": 32100, "COND_SYNTACTIC_DIV_0": 32141, "COND_LEXICAL_DIV_15": 32103, "COND_SYNTACTIC_DIV_70": 32155, "COND_LEXICAL_DIV_100": 32102, "COND_SEMANTIC_SIM_5": 32130, "COND_SEMANTIC_SIM_10": 32122, "COND_SYNTACTIC_DIV_55": 32152, "COND_LEXICAL_DIV_50": 32111, "COND_SEMANTIC_SIM_40": 32128, "COND_SEMANTIC_SIM_55": 32132, "COND_SYNTACTIC_DIV_35": 32147, "COND_SYNTACTIC_DIV_30": 32146, "COND_SEMANTIC_SIM_20": 32124, "COND_LEXICAL_DIV_80": 32117, "COND_LEXICAL_DIV_25": 32105, "COND_LEXICAL_DIV_40": 32108, "COND_SEMANTIC_SIM_65": 32134, "COND_LEXICAL_DIV_90": 32119, "COND_LEXICAL_DIV_10": 32101, "COND_LEXICAL_DIV_60": 32113, "COND_SEMANTIC_SIM_50": 32131, "COND_SYNTACTIC_DIV_65": 32154, "COND_SEMANTIC_SIM_60": 32133, "COND_SYNTACTIC_DIV_50": 32151, "COND_SEMANTIC_SIM_0": 32121, "COND_SYNTACTIC_DIV_5": 32150, "COND_LEXICAL_DIV_70": 32115, "COND_SYNTACTIC_DIV_15": 32143, "COND_SEMANTIC_SIM_70": 32135, "COND_SEMANTIC_SIM_35": 32127, "COND_LEXICAL_DIV_65": 32114, "COND_LEXICAL_DIV_95": 32120, "COND_SYNTACTIC_DIV_25": 32145, "COND_SEMANTIC_SIM_75": 32136, "COND_SEMANTIC_SIM_85": 32138, "COND_LEXICAL_DIV_45": 32109, "COND_SEMANTIC_SIM_90": 32139}
config.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "t5-base",
3
+ "architectures": [
4
+ "T5ForConditionalGeneration"
5
+ ],
6
+ "d_ff": 3072,
7
+ "d_kv": 64,
8
+ "d_model": 768,
9
+ "decoder_start_token_id": 0,
10
+ "dropout_rate": 0.1,
11
+ "eos_token_id": 1,
12
+ "feed_forward_proj": "relu",
13
+ "gradient_checkpointing": false,
14
+ "initializer_factor": 1.0,
15
+ "is_encoder_decoder": true,
16
+ "layer_norm_epsilon": 1e-06,
17
+ "model_type": "t5",
18
+ "n_positions": 512,
19
+ "num_decoder_layers": 12,
20
+ "num_heads": 12,
21
+ "num_layers": 12,
22
+ "output_past": true,
23
+ "pad_token_id": 0,
24
+ "relative_attention_num_buckets": 32,
25
+ "task_specific_params": {
26
+ "summarization": {
27
+ "early_stopping": true,
28
+ "length_penalty": 2.0,
29
+ "max_length": 200,
30
+ "min_length": 30,
31
+ "no_repeat_ngram_size": 3,
32
+ "num_beams": 4,
33
+ "prefix": "summarize: "
34
+ },
35
+ "translation_en_to_de": {
36
+ "early_stopping": true,
37
+ "max_length": 300,
38
+ "num_beams": 4,
39
+ "prefix": "translate English to German: "
40
+ },
41
+ "translation_en_to_fr": {
42
+ "early_stopping": true,
43
+ "max_length": 300,
44
+ "num_beams": 4,
45
+ "prefix": "translate English to French: "
46
+ },
47
+ "translation_en_to_ro": {
48
+ "early_stopping": true,
49
+ "max_length": 300,
50
+ "num_beams": 4,
51
+ "prefix": "translate English to Romanian: "
52
+ }
53
+ },
54
+ "transformers_version": "4.6.1",
55
+ "use_cache": true,
56
+ "vocab_size": 32158
57
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f65c2de7e551db0a564fa952fdde1558c794dfa061b8de8107ce5722ab5069c
3
+ size 891825169
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "additional_special_tokens": ["COND_LEXICAL_DIV_0", "COND_LEXICAL_DIV_10", "COND_LEXICAL_DIV_100", "COND_LEXICAL_DIV_15", "COND_LEXICAL_DIV_20", "COND_LEXICAL_DIV_25", "COND_LEXICAL_DIV_30", "COND_LEXICAL_DIV_35", "COND_LEXICAL_DIV_40", "COND_LEXICAL_DIV_45", "COND_LEXICAL_DIV_5", "COND_LEXICAL_DIV_50", "COND_LEXICAL_DIV_55", "COND_LEXICAL_DIV_60", "COND_LEXICAL_DIV_65", "COND_LEXICAL_DIV_70", "COND_LEXICAL_DIV_75", "COND_LEXICAL_DIV_80", "COND_LEXICAL_DIV_85", "COND_LEXICAL_DIV_90", "COND_LEXICAL_DIV_95", "COND_SEMANTIC_SIM_0", "COND_SEMANTIC_SIM_10", "COND_SEMANTIC_SIM_15", "COND_SEMANTIC_SIM_20", "COND_SEMANTIC_SIM_25", "COND_SEMANTIC_SIM_30", "COND_SEMANTIC_SIM_35", "COND_SEMANTIC_SIM_40", "COND_SEMANTIC_SIM_45", "COND_SEMANTIC_SIM_5", "COND_SEMANTIC_SIM_50", "COND_SEMANTIC_SIM_55", "COND_SEMANTIC_SIM_60", "COND_SEMANTIC_SIM_65", "COND_SEMANTIC_SIM_70", "COND_SEMANTIC_SIM_75", "COND_SEMANTIC_SIM_80", "COND_SEMANTIC_SIM_85", "COND_SEMANTIC_SIM_90", "COND_SEMANTIC_SIM_95", "COND_SYNTACTIC_DIV_0", "COND_SYNTACTIC_DIV_10", "COND_SYNTACTIC_DIV_15", "COND_SYNTACTIC_DIV_20", "COND_SYNTACTIC_DIV_25", "COND_SYNTACTIC_DIV_30", "COND_SYNTACTIC_DIV_35", "COND_SYNTACTIC_DIV_40", "COND_SYNTACTIC_DIV_45", "COND_SYNTACTIC_DIV_5", "COND_SYNTACTIC_DIV_50", "COND_SYNTACTIC_DIV_55", "COND_SYNTACTIC_DIV_60", "COND_SYNTACTIC_DIV_65", "COND_SYNTACTIC_DIV_70", "COND_SYNTACTIC_DIV_75", "COND_SYNTACTIC_DIV_80"]}
spiece.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d60acb128cf7b7f2536e8f38a5b18a05535c9e14c7a355904270e15b0945ea86
3
+ size 791656
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "extra_ids": 0, "additional_special_tokens": ["COND_LEXICAL_DIV_0", "COND_LEXICAL_DIV_10", "COND_LEXICAL_DIV_100", "COND_LEXICAL_DIV_15", "COND_LEXICAL_DIV_20", "COND_LEXICAL_DIV_25", "COND_LEXICAL_DIV_30", "COND_LEXICAL_DIV_35", "COND_LEXICAL_DIV_40", "COND_LEXICAL_DIV_45", "COND_LEXICAL_DIV_5", "COND_LEXICAL_DIV_50", "COND_LEXICAL_DIV_55", "COND_LEXICAL_DIV_60", "COND_LEXICAL_DIV_65", "COND_LEXICAL_DIV_70", "COND_LEXICAL_DIV_75", "COND_LEXICAL_DIV_80", "COND_LEXICAL_DIV_85", "COND_LEXICAL_DIV_90", "COND_LEXICAL_DIV_95", "COND_SEMANTIC_SIM_0", "COND_SEMANTIC_SIM_10", "COND_SEMANTIC_SIM_15", "COND_SEMANTIC_SIM_20", "COND_SEMANTIC_SIM_25", "COND_SEMANTIC_SIM_30", "COND_SEMANTIC_SIM_35", "COND_SEMANTIC_SIM_40", "COND_SEMANTIC_SIM_45", "COND_SEMANTIC_SIM_5", "COND_SEMANTIC_SIM_50", "COND_SEMANTIC_SIM_55", "COND_SEMANTIC_SIM_60", "COND_SEMANTIC_SIM_65", "COND_SEMANTIC_SIM_70", "COND_SEMANTIC_SIM_75", "COND_SEMANTIC_SIM_80", "COND_SEMANTIC_SIM_85", "COND_SEMANTIC_SIM_90", "COND_SEMANTIC_SIM_95", "COND_SYNTACTIC_DIV_0", "COND_SYNTACTIC_DIV_10", "COND_SYNTACTIC_DIV_15", "COND_SYNTACTIC_DIV_20", "COND_SYNTACTIC_DIV_25", "COND_SYNTACTIC_DIV_30", "COND_SYNTACTIC_DIV_35", "COND_SYNTACTIC_DIV_40", "COND_SYNTACTIC_DIV_45", "COND_SYNTACTIC_DIV_5", "COND_SYNTACTIC_DIV_50", "COND_SYNTACTIC_DIV_55", "COND_SYNTACTIC_DIV_60", "COND_SYNTACTIC_DIV_65", "COND_SYNTACTIC_DIV_70", "COND_SYNTACTIC_DIV_75", "COND_SYNTACTIC_DIV_80"], "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "t5-base"}
trainer_state.json ADDED
@@ -0,0 +1,4024 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.9555555555555557,
5
+ "global_step": 332500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 0.001,
13
+ "loss": 1.693,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 0.001,
19
+ "loss": 1.5462,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 0.001,
25
+ "loss": 1.5029,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.02,
30
+ "learning_rate": 0.001,
31
+ "loss": 1.4654,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.02,
36
+ "learning_rate": 0.001,
37
+ "loss": 1.4407,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.03,
42
+ "learning_rate": 0.001,
43
+ "loss": 1.3838,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.03,
48
+ "learning_rate": 0.001,
49
+ "loss": 1.37,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.04,
54
+ "learning_rate": 0.001,
55
+ "loss": 1.3818,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.04,
60
+ "learning_rate": 0.001,
61
+ "loss": 1.3539,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.04,
66
+ "learning_rate": 0.001,
67
+ "loss": 1.3605,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.05,
72
+ "learning_rate": 0.001,
73
+ "loss": 1.3485,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.05,
78
+ "learning_rate": 0.001,
79
+ "loss": 1.3284,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.06,
84
+ "learning_rate": 0.001,
85
+ "loss": 1.3229,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.06,
90
+ "learning_rate": 0.001,
91
+ "loss": 1.3203,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.07,
96
+ "learning_rate": 0.001,
97
+ "loss": 1.3039,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.07,
102
+ "learning_rate": 0.001,
103
+ "loss": 1.2939,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.08,
108
+ "learning_rate": 0.001,
109
+ "loss": 1.2585,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.08,
114
+ "learning_rate": 0.001,
115
+ "loss": 1.2824,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.08,
120
+ "learning_rate": 0.001,
121
+ "loss": 1.2786,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.09,
126
+ "learning_rate": 0.001,
127
+ "loss": 1.2889,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.09,
132
+ "learning_rate": 0.001,
133
+ "loss": 1.2513,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.1,
138
+ "learning_rate": 0.001,
139
+ "loss": 1.2813,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 0.1,
144
+ "learning_rate": 0.001,
145
+ "loss": 1.2363,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 0.11,
150
+ "learning_rate": 0.001,
151
+ "loss": 1.2591,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 0.11,
156
+ "learning_rate": 0.001,
157
+ "loss": 1.238,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 0.12,
162
+ "learning_rate": 0.001,
163
+ "loss": 1.2293,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 0.12,
168
+ "learning_rate": 0.001,
169
+ "loss": 1.2416,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 0.12,
174
+ "learning_rate": 0.001,
175
+ "loss": 1.2246,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 0.13,
180
+ "learning_rate": 0.001,
181
+ "loss": 1.2235,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 0.13,
186
+ "learning_rate": 0.001,
187
+ "loss": 1.2157,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 0.14,
192
+ "learning_rate": 0.001,
193
+ "loss": 1.2175,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 0.14,
198
+ "learning_rate": 0.001,
199
+ "loss": 1.2135,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 0.15,
204
+ "learning_rate": 0.001,
205
+ "loss": 1.2295,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 0.15,
210
+ "learning_rate": 0.001,
211
+ "loss": 1.2025,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 0.16,
216
+ "learning_rate": 0.001,
217
+ "loss": 1.1972,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 0.16,
222
+ "learning_rate": 0.001,
223
+ "loss": 1.2001,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 0.16,
228
+ "learning_rate": 0.001,
229
+ "loss": 1.1981,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 0.17,
234
+ "learning_rate": 0.001,
235
+ "loss": 1.1821,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 0.17,
240
+ "learning_rate": 0.001,
241
+ "loss": 1.1872,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 0.18,
246
+ "learning_rate": 0.001,
247
+ "loss": 1.1764,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 0.18,
252
+ "learning_rate": 0.001,
253
+ "loss": 1.1847,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 0.19,
258
+ "learning_rate": 0.001,
259
+ "loss": 1.1703,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 0.19,
264
+ "learning_rate": 0.001,
265
+ "loss": 1.1738,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 0.2,
270
+ "learning_rate": 0.001,
271
+ "loss": 1.1653,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 0.2,
276
+ "learning_rate": 0.001,
277
+ "loss": 1.1625,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 0.2,
282
+ "learning_rate": 0.001,
283
+ "loss": 1.188,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 0.21,
288
+ "learning_rate": 0.001,
289
+ "loss": 1.1871,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 0.21,
294
+ "learning_rate": 0.001,
295
+ "loss": 1.1614,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 0.22,
300
+ "learning_rate": 0.001,
301
+ "loss": 1.1665,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 0.22,
306
+ "learning_rate": 0.001,
307
+ "loss": 1.1696,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 0.23,
312
+ "learning_rate": 0.001,
313
+ "loss": 1.1465,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 0.23,
318
+ "learning_rate": 0.001,
319
+ "loss": 1.1538,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 0.24,
324
+ "learning_rate": 0.001,
325
+ "loss": 1.1523,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 0.24,
330
+ "learning_rate": 0.001,
331
+ "loss": 1.1376,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 0.24,
336
+ "learning_rate": 0.001,
337
+ "loss": 1.1364,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 0.25,
342
+ "learning_rate": 0.001,
343
+ "loss": 1.1567,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 0.25,
348
+ "learning_rate": 0.001,
349
+ "loss": 1.1393,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 0.26,
354
+ "learning_rate": 0.001,
355
+ "loss": 1.1399,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 0.26,
360
+ "learning_rate": 0.001,
361
+ "loss": 1.1328,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 0.27,
366
+ "learning_rate": 0.001,
367
+ "loss": 1.1456,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 0.27,
372
+ "learning_rate": 0.001,
373
+ "loss": 1.1311,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 0.28,
378
+ "learning_rate": 0.001,
379
+ "loss": 1.1246,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 0.28,
384
+ "learning_rate": 0.001,
385
+ "loss": 1.1358,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 0.28,
390
+ "learning_rate": 0.001,
391
+ "loss": 1.1271,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 0.29,
396
+ "learning_rate": 0.001,
397
+ "loss": 1.1082,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 0.29,
402
+ "learning_rate": 0.001,
403
+ "loss": 1.1064,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 0.3,
408
+ "learning_rate": 0.001,
409
+ "loss": 1.122,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 0.3,
414
+ "learning_rate": 0.001,
415
+ "loss": 1.1282,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 0.31,
420
+ "learning_rate": 0.001,
421
+ "loss": 1.1088,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 0.31,
426
+ "learning_rate": 0.001,
427
+ "loss": 1.1167,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 0.32,
432
+ "learning_rate": 0.001,
433
+ "loss": 1.0914,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 0.32,
438
+ "learning_rate": 0.001,
439
+ "loss": 1.1197,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 0.32,
444
+ "learning_rate": 0.001,
445
+ "loss": 1.1085,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 0.33,
450
+ "learning_rate": 0.001,
451
+ "loss": 1.0976,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 0.33,
456
+ "learning_rate": 0.001,
457
+ "loss": 1.1057,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 0.34,
462
+ "learning_rate": 0.001,
463
+ "loss": 1.0936,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 0.34,
468
+ "learning_rate": 0.001,
469
+ "loss": 1.0813,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 0.35,
474
+ "learning_rate": 0.001,
475
+ "loss": 1.0897,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 0.35,
480
+ "learning_rate": 0.001,
481
+ "loss": 1.1009,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 0.36,
486
+ "learning_rate": 0.001,
487
+ "loss": 1.0939,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 0.36,
492
+ "learning_rate": 0.001,
493
+ "loss": 1.0905,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 0.36,
498
+ "learning_rate": 0.001,
499
+ "loss": 1.0965,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 0.37,
504
+ "learning_rate": 0.001,
505
+ "loss": 1.0823,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 0.37,
510
+ "learning_rate": 0.001,
511
+ "loss": 1.0951,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 0.38,
516
+ "learning_rate": 0.001,
517
+ "loss": 1.0823,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 0.38,
522
+ "learning_rate": 0.001,
523
+ "loss": 1.0741,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 0.39,
528
+ "learning_rate": 0.001,
529
+ "loss": 1.0622,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 0.39,
534
+ "learning_rate": 0.001,
535
+ "loss": 1.0879,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 0.4,
540
+ "learning_rate": 0.001,
541
+ "loss": 1.0559,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 0.4,
546
+ "learning_rate": 0.001,
547
+ "loss": 1.0766,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 0.4,
552
+ "learning_rate": 0.001,
553
+ "loss": 1.0609,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 0.41,
558
+ "learning_rate": 0.001,
559
+ "loss": 1.063,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 0.41,
564
+ "learning_rate": 0.001,
565
+ "loss": 1.0618,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 0.42,
570
+ "learning_rate": 0.001,
571
+ "loss": 1.0618,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 0.42,
576
+ "learning_rate": 0.001,
577
+ "loss": 1.0488,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 0.43,
582
+ "learning_rate": 0.001,
583
+ "loss": 1.0686,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 0.43,
588
+ "learning_rate": 0.001,
589
+ "loss": 1.0457,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 0.44,
594
+ "learning_rate": 0.001,
595
+ "loss": 1.0494,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 0.44,
600
+ "learning_rate": 0.001,
601
+ "loss": 1.0498,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 0.44,
606
+ "learning_rate": 0.001,
607
+ "loss": 1.0677,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 0.45,
612
+ "learning_rate": 0.001,
613
+ "loss": 1.0548,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 0.45,
618
+ "learning_rate": 0.001,
619
+ "loss": 1.0721,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 0.46,
624
+ "learning_rate": 0.001,
625
+ "loss": 1.0442,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 0.46,
630
+ "learning_rate": 0.001,
631
+ "loss": 1.0502,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 0.47,
636
+ "learning_rate": 0.001,
637
+ "loss": 1.0457,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 0.47,
642
+ "learning_rate": 0.001,
643
+ "loss": 1.0459,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 0.48,
648
+ "learning_rate": 0.001,
649
+ "loss": 1.0419,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 0.48,
654
+ "learning_rate": 0.001,
655
+ "loss": 1.0533,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 0.48,
660
+ "learning_rate": 0.001,
661
+ "loss": 1.0498,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 0.49,
666
+ "learning_rate": 0.001,
667
+ "loss": 1.0347,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 0.49,
672
+ "learning_rate": 0.001,
673
+ "loss": 1.0422,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 0.5,
678
+ "learning_rate": 0.001,
679
+ "loss": 1.0431,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 0.5,
684
+ "learning_rate": 0.001,
685
+ "loss": 1.0365,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 0.51,
690
+ "learning_rate": 0.001,
691
+ "loss": 1.0383,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 0.51,
696
+ "learning_rate": 0.001,
697
+ "loss": 1.0291,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 0.52,
702
+ "learning_rate": 0.001,
703
+ "loss": 1.037,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 0.52,
708
+ "learning_rate": 0.001,
709
+ "loss": 1.0293,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 0.52,
714
+ "learning_rate": 0.001,
715
+ "loss": 1.0243,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 0.53,
720
+ "learning_rate": 0.001,
721
+ "loss": 1.0319,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 0.53,
726
+ "learning_rate": 0.001,
727
+ "loss": 1.0237,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 0.54,
732
+ "learning_rate": 0.001,
733
+ "loss": 1.0394,
734
+ "step": 60500
735
+ },
736
+ {
737
+ "epoch": 0.54,
738
+ "learning_rate": 0.001,
739
+ "loss": 1.0306,
740
+ "step": 61000
741
+ },
742
+ {
743
+ "epoch": 0.55,
744
+ "learning_rate": 0.001,
745
+ "loss": 1.0184,
746
+ "step": 61500
747
+ },
748
+ {
749
+ "epoch": 0.55,
750
+ "learning_rate": 0.001,
751
+ "loss": 1.0149,
752
+ "step": 62000
753
+ },
754
+ {
755
+ "epoch": 0.56,
756
+ "learning_rate": 0.001,
757
+ "loss": 1.024,
758
+ "step": 62500
759
+ },
760
+ {
761
+ "epoch": 0.56,
762
+ "learning_rate": 0.001,
763
+ "loss": 1.0095,
764
+ "step": 63000
765
+ },
766
+ {
767
+ "epoch": 0.56,
768
+ "learning_rate": 0.001,
769
+ "loss": 1.008,
770
+ "step": 63500
771
+ },
772
+ {
773
+ "epoch": 0.57,
774
+ "learning_rate": 0.001,
775
+ "loss": 1.0057,
776
+ "step": 64000
777
+ },
778
+ {
779
+ "epoch": 0.57,
780
+ "learning_rate": 0.001,
781
+ "loss": 1.0232,
782
+ "step": 64500
783
+ },
784
+ {
785
+ "epoch": 0.58,
786
+ "learning_rate": 0.001,
787
+ "loss": 1.02,
788
+ "step": 65000
789
+ },
790
+ {
791
+ "epoch": 0.58,
792
+ "learning_rate": 0.001,
793
+ "loss": 1.0141,
794
+ "step": 65500
795
+ },
796
+ {
797
+ "epoch": 0.59,
798
+ "learning_rate": 0.001,
799
+ "loss": 1.0132,
800
+ "step": 66000
801
+ },
802
+ {
803
+ "epoch": 0.59,
804
+ "learning_rate": 0.001,
805
+ "loss": 1.0229,
806
+ "step": 66500
807
+ },
808
+ {
809
+ "epoch": 0.6,
810
+ "learning_rate": 0.001,
811
+ "loss": 1.0211,
812
+ "step": 67000
813
+ },
814
+ {
815
+ "epoch": 0.6,
816
+ "learning_rate": 0.001,
817
+ "loss": 1.0062,
818
+ "step": 67500
819
+ },
820
+ {
821
+ "epoch": 0.6,
822
+ "learning_rate": 0.001,
823
+ "loss": 1.0062,
824
+ "step": 68000
825
+ },
826
+ {
827
+ "epoch": 0.61,
828
+ "learning_rate": 0.001,
829
+ "loss": 1.0044,
830
+ "step": 68500
831
+ },
832
+ {
833
+ "epoch": 0.61,
834
+ "learning_rate": 0.001,
835
+ "loss": 1.0211,
836
+ "step": 69000
837
+ },
838
+ {
839
+ "epoch": 0.62,
840
+ "learning_rate": 0.001,
841
+ "loss": 0.9982,
842
+ "step": 69500
843
+ },
844
+ {
845
+ "epoch": 0.62,
846
+ "learning_rate": 0.001,
847
+ "loss": 1.011,
848
+ "step": 70000
849
+ },
850
+ {
851
+ "epoch": 0.63,
852
+ "learning_rate": 0.001,
853
+ "loss": 0.9963,
854
+ "step": 70500
855
+ },
856
+ {
857
+ "epoch": 0.63,
858
+ "learning_rate": 0.001,
859
+ "loss": 1.0009,
860
+ "step": 71000
861
+ },
862
+ {
863
+ "epoch": 0.64,
864
+ "learning_rate": 0.001,
865
+ "loss": 0.9925,
866
+ "step": 71500
867
+ },
868
+ {
869
+ "epoch": 0.64,
870
+ "learning_rate": 0.001,
871
+ "loss": 0.9819,
872
+ "step": 72000
873
+ },
874
+ {
875
+ "epoch": 0.64,
876
+ "learning_rate": 0.001,
877
+ "loss": 1.012,
878
+ "step": 72500
879
+ },
880
+ {
881
+ "epoch": 0.65,
882
+ "learning_rate": 0.001,
883
+ "loss": 0.9856,
884
+ "step": 73000
885
+ },
886
+ {
887
+ "epoch": 0.65,
888
+ "learning_rate": 0.001,
889
+ "loss": 0.9991,
890
+ "step": 73500
891
+ },
892
+ {
893
+ "epoch": 0.66,
894
+ "learning_rate": 0.001,
895
+ "loss": 0.9929,
896
+ "step": 74000
897
+ },
898
+ {
899
+ "epoch": 0.66,
900
+ "learning_rate": 0.001,
901
+ "loss": 0.9949,
902
+ "step": 74500
903
+ },
904
+ {
905
+ "epoch": 0.67,
906
+ "learning_rate": 0.001,
907
+ "loss": 0.9874,
908
+ "step": 75000
909
+ },
910
+ {
911
+ "epoch": 0.67,
912
+ "learning_rate": 0.001,
913
+ "loss": 0.9843,
914
+ "step": 75500
915
+ },
916
+ {
917
+ "epoch": 0.68,
918
+ "learning_rate": 0.001,
919
+ "loss": 1.0074,
920
+ "step": 76000
921
+ },
922
+ {
923
+ "epoch": 0.68,
924
+ "learning_rate": 0.001,
925
+ "loss": 0.9937,
926
+ "step": 76500
927
+ },
928
+ {
929
+ "epoch": 0.68,
930
+ "learning_rate": 0.001,
931
+ "loss": 0.9898,
932
+ "step": 77000
933
+ },
934
+ {
935
+ "epoch": 0.69,
936
+ "learning_rate": 0.001,
937
+ "loss": 0.9799,
938
+ "step": 77500
939
+ },
940
+ {
941
+ "epoch": 0.69,
942
+ "learning_rate": 0.001,
943
+ "loss": 0.9855,
944
+ "step": 78000
945
+ },
946
+ {
947
+ "epoch": 0.7,
948
+ "learning_rate": 0.001,
949
+ "loss": 0.9797,
950
+ "step": 78500
951
+ },
952
+ {
953
+ "epoch": 0.7,
954
+ "learning_rate": 0.001,
955
+ "loss": 0.9774,
956
+ "step": 79000
957
+ },
958
+ {
959
+ "epoch": 0.71,
960
+ "learning_rate": 0.001,
961
+ "loss": 0.981,
962
+ "step": 79500
963
+ },
964
+ {
965
+ "epoch": 0.71,
966
+ "learning_rate": 0.001,
967
+ "loss": 0.9909,
968
+ "step": 80000
969
+ },
970
+ {
971
+ "epoch": 0.72,
972
+ "learning_rate": 0.001,
973
+ "loss": 0.9734,
974
+ "step": 80500
975
+ },
976
+ {
977
+ "epoch": 0.72,
978
+ "learning_rate": 0.001,
979
+ "loss": 0.9857,
980
+ "step": 81000
981
+ },
982
+ {
983
+ "epoch": 0.72,
984
+ "learning_rate": 0.001,
985
+ "loss": 0.976,
986
+ "step": 81500
987
+ },
988
+ {
989
+ "epoch": 0.73,
990
+ "learning_rate": 0.001,
991
+ "loss": 0.9842,
992
+ "step": 82000
993
+ },
994
+ {
995
+ "epoch": 0.73,
996
+ "learning_rate": 0.001,
997
+ "loss": 0.9798,
998
+ "step": 82500
999
+ },
1000
+ {
1001
+ "epoch": 0.74,
1002
+ "learning_rate": 0.001,
1003
+ "loss": 0.9816,
1004
+ "step": 83000
1005
+ },
1006
+ {
1007
+ "epoch": 0.74,
1008
+ "learning_rate": 0.001,
1009
+ "loss": 0.9686,
1010
+ "step": 83500
1011
+ },
1012
+ {
1013
+ "epoch": 0.75,
1014
+ "learning_rate": 0.001,
1015
+ "loss": 0.9741,
1016
+ "step": 84000
1017
+ },
1018
+ {
1019
+ "epoch": 0.75,
1020
+ "learning_rate": 0.001,
1021
+ "loss": 1.0041,
1022
+ "step": 84500
1023
+ },
1024
+ {
1025
+ "epoch": 0.76,
1026
+ "learning_rate": 0.001,
1027
+ "loss": 0.9844,
1028
+ "step": 85000
1029
+ },
1030
+ {
1031
+ "epoch": 0.76,
1032
+ "learning_rate": 0.001,
1033
+ "loss": 0.9761,
1034
+ "step": 85500
1035
+ },
1036
+ {
1037
+ "epoch": 0.76,
1038
+ "learning_rate": 0.001,
1039
+ "loss": 0.9775,
1040
+ "step": 86000
1041
+ },
1042
+ {
1043
+ "epoch": 0.77,
1044
+ "learning_rate": 0.001,
1045
+ "loss": 0.9667,
1046
+ "step": 86500
1047
+ },
1048
+ {
1049
+ "epoch": 0.77,
1050
+ "learning_rate": 0.001,
1051
+ "loss": 0.9434,
1052
+ "step": 87000
1053
+ },
1054
+ {
1055
+ "epoch": 0.78,
1056
+ "learning_rate": 0.001,
1057
+ "loss": 0.976,
1058
+ "step": 87500
1059
+ },
1060
+ {
1061
+ "epoch": 0.78,
1062
+ "learning_rate": 0.001,
1063
+ "loss": 0.962,
1064
+ "step": 88000
1065
+ },
1066
+ {
1067
+ "epoch": 0.79,
1068
+ "learning_rate": 0.001,
1069
+ "loss": 0.9782,
1070
+ "step": 88500
1071
+ },
1072
+ {
1073
+ "epoch": 0.79,
1074
+ "learning_rate": 0.001,
1075
+ "loss": 0.9739,
1076
+ "step": 89000
1077
+ },
1078
+ {
1079
+ "epoch": 0.8,
1080
+ "learning_rate": 0.001,
1081
+ "loss": 0.9858,
1082
+ "step": 89500
1083
+ },
1084
+ {
1085
+ "epoch": 0.8,
1086
+ "learning_rate": 0.001,
1087
+ "loss": 0.9573,
1088
+ "step": 90000
1089
+ },
1090
+ {
1091
+ "epoch": 0.8,
1092
+ "learning_rate": 0.001,
1093
+ "loss": 0.9776,
1094
+ "step": 90500
1095
+ },
1096
+ {
1097
+ "epoch": 0.81,
1098
+ "learning_rate": 0.001,
1099
+ "loss": 0.9729,
1100
+ "step": 91000
1101
+ },
1102
+ {
1103
+ "epoch": 0.81,
1104
+ "learning_rate": 0.001,
1105
+ "loss": 0.9878,
1106
+ "step": 91500
1107
+ },
1108
+ {
1109
+ "epoch": 0.82,
1110
+ "learning_rate": 0.001,
1111
+ "loss": 0.9735,
1112
+ "step": 92000
1113
+ },
1114
+ {
1115
+ "epoch": 0.82,
1116
+ "learning_rate": 0.001,
1117
+ "loss": 0.9581,
1118
+ "step": 92500
1119
+ },
1120
+ {
1121
+ "epoch": 0.83,
1122
+ "learning_rate": 0.001,
1123
+ "loss": 0.9593,
1124
+ "step": 93000
1125
+ },
1126
+ {
1127
+ "epoch": 0.83,
1128
+ "learning_rate": 0.001,
1129
+ "loss": 0.9611,
1130
+ "step": 93500
1131
+ },
1132
+ {
1133
+ "epoch": 0.84,
1134
+ "learning_rate": 0.001,
1135
+ "loss": 0.9602,
1136
+ "step": 94000
1137
+ },
1138
+ {
1139
+ "epoch": 0.84,
1140
+ "learning_rate": 0.001,
1141
+ "loss": 0.9516,
1142
+ "step": 94500
1143
+ },
1144
+ {
1145
+ "epoch": 0.84,
1146
+ "learning_rate": 0.001,
1147
+ "loss": 0.934,
1148
+ "step": 95000
1149
+ },
1150
+ {
1151
+ "epoch": 0.85,
1152
+ "learning_rate": 0.001,
1153
+ "loss": 0.9451,
1154
+ "step": 95500
1155
+ },
1156
+ {
1157
+ "epoch": 0.85,
1158
+ "learning_rate": 0.001,
1159
+ "loss": 0.956,
1160
+ "step": 96000
1161
+ },
1162
+ {
1163
+ "epoch": 0.86,
1164
+ "learning_rate": 0.001,
1165
+ "loss": 0.9612,
1166
+ "step": 96500
1167
+ },
1168
+ {
1169
+ "epoch": 0.86,
1170
+ "learning_rate": 0.001,
1171
+ "loss": 0.946,
1172
+ "step": 97000
1173
+ },
1174
+ {
1175
+ "epoch": 0.87,
1176
+ "learning_rate": 0.001,
1177
+ "loss": 0.9406,
1178
+ "step": 97500
1179
+ },
1180
+ {
1181
+ "epoch": 0.87,
1182
+ "learning_rate": 0.001,
1183
+ "loss": 0.9641,
1184
+ "step": 98000
1185
+ },
1186
+ {
1187
+ "epoch": 0.88,
1188
+ "learning_rate": 0.001,
1189
+ "loss": 0.9696,
1190
+ "step": 98500
1191
+ },
1192
+ {
1193
+ "epoch": 0.88,
1194
+ "learning_rate": 0.001,
1195
+ "loss": 0.9599,
1196
+ "step": 99000
1197
+ },
1198
+ {
1199
+ "epoch": 0.88,
1200
+ "learning_rate": 0.001,
1201
+ "loss": 0.949,
1202
+ "step": 99500
1203
+ },
1204
+ {
1205
+ "epoch": 0.89,
1206
+ "learning_rate": 0.001,
1207
+ "loss": 0.9491,
1208
+ "step": 100000
1209
+ },
1210
+ {
1211
+ "epoch": 0.89,
1212
+ "learning_rate": 0.001,
1213
+ "loss": 0.9468,
1214
+ "step": 100500
1215
+ },
1216
+ {
1217
+ "epoch": 0.9,
1218
+ "learning_rate": 0.001,
1219
+ "loss": 0.9301,
1220
+ "step": 101000
1221
+ },
1222
+ {
1223
+ "epoch": 0.9,
1224
+ "learning_rate": 0.001,
1225
+ "loss": 0.9481,
1226
+ "step": 101500
1227
+ },
1228
+ {
1229
+ "epoch": 0.91,
1230
+ "learning_rate": 0.001,
1231
+ "loss": 0.946,
1232
+ "step": 102000
1233
+ },
1234
+ {
1235
+ "epoch": 0.91,
1236
+ "learning_rate": 0.001,
1237
+ "loss": 0.94,
1238
+ "step": 102500
1239
+ },
1240
+ {
1241
+ "epoch": 0.92,
1242
+ "learning_rate": 0.001,
1243
+ "loss": 0.9288,
1244
+ "step": 103000
1245
+ },
1246
+ {
1247
+ "epoch": 0.92,
1248
+ "learning_rate": 0.001,
1249
+ "loss": 0.9514,
1250
+ "step": 103500
1251
+ },
1252
+ {
1253
+ "epoch": 0.92,
1254
+ "learning_rate": 0.001,
1255
+ "loss": 0.9371,
1256
+ "step": 104000
1257
+ },
1258
+ {
1259
+ "epoch": 0.93,
1260
+ "learning_rate": 0.001,
1261
+ "loss": 0.9269,
1262
+ "step": 104500
1263
+ },
1264
+ {
1265
+ "epoch": 0.93,
1266
+ "learning_rate": 0.001,
1267
+ "loss": 0.9405,
1268
+ "step": 105000
1269
+ },
1270
+ {
1271
+ "epoch": 0.94,
1272
+ "learning_rate": 0.001,
1273
+ "loss": 0.9306,
1274
+ "step": 105500
1275
+ },
1276
+ {
1277
+ "epoch": 0.94,
1278
+ "learning_rate": 0.001,
1279
+ "loss": 0.9397,
1280
+ "step": 106000
1281
+ },
1282
+ {
1283
+ "epoch": 0.95,
1284
+ "learning_rate": 0.001,
1285
+ "loss": 0.9247,
1286
+ "step": 106500
1287
+ },
1288
+ {
1289
+ "epoch": 0.95,
1290
+ "learning_rate": 0.001,
1291
+ "loss": 0.9122,
1292
+ "step": 107000
1293
+ },
1294
+ {
1295
+ "epoch": 0.96,
1296
+ "learning_rate": 0.001,
1297
+ "loss": 0.9361,
1298
+ "step": 107500
1299
+ },
1300
+ {
1301
+ "epoch": 0.96,
1302
+ "learning_rate": 0.001,
1303
+ "loss": 0.9255,
1304
+ "step": 108000
1305
+ },
1306
+ {
1307
+ "epoch": 0.96,
1308
+ "learning_rate": 0.001,
1309
+ "loss": 0.9434,
1310
+ "step": 108500
1311
+ },
1312
+ {
1313
+ "epoch": 0.97,
1314
+ "learning_rate": 0.001,
1315
+ "loss": 0.9383,
1316
+ "step": 109000
1317
+ },
1318
+ {
1319
+ "epoch": 0.97,
1320
+ "learning_rate": 0.001,
1321
+ "loss": 0.9371,
1322
+ "step": 109500
1323
+ },
1324
+ {
1325
+ "epoch": 0.98,
1326
+ "learning_rate": 0.001,
1327
+ "loss": 0.9272,
1328
+ "step": 110000
1329
+ },
1330
+ {
1331
+ "epoch": 0.98,
1332
+ "learning_rate": 0.001,
1333
+ "loss": 0.9288,
1334
+ "step": 110500
1335
+ },
1336
+ {
1337
+ "epoch": 0.99,
1338
+ "learning_rate": 0.001,
1339
+ "loss": 0.9248,
1340
+ "step": 111000
1341
+ },
1342
+ {
1343
+ "epoch": 0.99,
1344
+ "learning_rate": 0.001,
1345
+ "loss": 0.924,
1346
+ "step": 111500
1347
+ },
1348
+ {
1349
+ "epoch": 1.0,
1350
+ "learning_rate": 0.001,
1351
+ "loss": 0.9319,
1352
+ "step": 112000
1353
+ },
1354
+ {
1355
+ "epoch": 1.0,
1356
+ "learning_rate": 0.001,
1357
+ "loss": 0.9324,
1358
+ "step": 112500
1359
+ },
1360
+ {
1361
+ "epoch": 1.0,
1362
+ "eval_bleu": 25.3594,
1363
+ "eval_gen_len": 11.3081,
1364
+ "eval_loss": 1.1617918014526367,
1365
+ "eval_runtime": 419.1845,
1366
+ "eval_samples_per_second": 33.398,
1367
+ "step": 112500
1368
+ },
1369
+ {
1370
+ "epoch": 1.0,
1371
+ "learning_rate": 0.001,
1372
+ "loss": 0.8688,
1373
+ "step": 113000
1374
+ },
1375
+ {
1376
+ "epoch": 1.01,
1377
+ "learning_rate": 0.001,
1378
+ "loss": 0.8644,
1379
+ "step": 113500
1380
+ },
1381
+ {
1382
+ "epoch": 1.01,
1383
+ "learning_rate": 0.001,
1384
+ "loss": 0.8799,
1385
+ "step": 114000
1386
+ },
1387
+ {
1388
+ "epoch": 1.02,
1389
+ "learning_rate": 0.001,
1390
+ "loss": 0.8541,
1391
+ "step": 114500
1392
+ },
1393
+ {
1394
+ "epoch": 1.02,
1395
+ "learning_rate": 0.001,
1396
+ "loss": 0.8661,
1397
+ "step": 115000
1398
+ },
1399
+ {
1400
+ "epoch": 1.03,
1401
+ "learning_rate": 0.001,
1402
+ "loss": 0.8715,
1403
+ "step": 115500
1404
+ },
1405
+ {
1406
+ "epoch": 1.03,
1407
+ "learning_rate": 0.001,
1408
+ "loss": 0.8783,
1409
+ "step": 116000
1410
+ },
1411
+ {
1412
+ "epoch": 1.04,
1413
+ "learning_rate": 0.001,
1414
+ "loss": 0.8769,
1415
+ "step": 116500
1416
+ },
1417
+ {
1418
+ "epoch": 1.04,
1419
+ "learning_rate": 0.001,
1420
+ "loss": 0.8697,
1421
+ "step": 117000
1422
+ },
1423
+ {
1424
+ "epoch": 1.04,
1425
+ "learning_rate": 0.001,
1426
+ "loss": 0.8621,
1427
+ "step": 117500
1428
+ },
1429
+ {
1430
+ "epoch": 1.05,
1431
+ "learning_rate": 0.001,
1432
+ "loss": 0.86,
1433
+ "step": 118000
1434
+ },
1435
+ {
1436
+ "epoch": 1.05,
1437
+ "learning_rate": 0.001,
1438
+ "loss": 0.8785,
1439
+ "step": 118500
1440
+ },
1441
+ {
1442
+ "epoch": 1.06,
1443
+ "learning_rate": 0.001,
1444
+ "loss": 0.8818,
1445
+ "step": 119000
1446
+ },
1447
+ {
1448
+ "epoch": 1.06,
1449
+ "learning_rate": 0.001,
1450
+ "loss": 0.8571,
1451
+ "step": 119500
1452
+ },
1453
+ {
1454
+ "epoch": 1.07,
1455
+ "learning_rate": 0.001,
1456
+ "loss": 0.8614,
1457
+ "step": 120000
1458
+ },
1459
+ {
1460
+ "epoch": 1.07,
1461
+ "learning_rate": 0.001,
1462
+ "loss": 0.8627,
1463
+ "step": 120500
1464
+ },
1465
+ {
1466
+ "epoch": 1.08,
1467
+ "learning_rate": 0.001,
1468
+ "loss": 0.8623,
1469
+ "step": 121000
1470
+ },
1471
+ {
1472
+ "epoch": 1.08,
1473
+ "learning_rate": 0.001,
1474
+ "loss": 0.8851,
1475
+ "step": 121500
1476
+ },
1477
+ {
1478
+ "epoch": 1.08,
1479
+ "learning_rate": 0.001,
1480
+ "loss": 0.8638,
1481
+ "step": 122000
1482
+ },
1483
+ {
1484
+ "epoch": 1.09,
1485
+ "learning_rate": 0.001,
1486
+ "loss": 0.8637,
1487
+ "step": 122500
1488
+ },
1489
+ {
1490
+ "epoch": 1.09,
1491
+ "learning_rate": 0.001,
1492
+ "loss": 0.8661,
1493
+ "step": 123000
1494
+ },
1495
+ {
1496
+ "epoch": 1.1,
1497
+ "learning_rate": 0.001,
1498
+ "loss": 0.852,
1499
+ "step": 123500
1500
+ },
1501
+ {
1502
+ "epoch": 1.1,
1503
+ "learning_rate": 0.001,
1504
+ "loss": 0.8738,
1505
+ "step": 124000
1506
+ },
1507
+ {
1508
+ "epoch": 1.11,
1509
+ "learning_rate": 0.001,
1510
+ "loss": 0.871,
1511
+ "step": 124500
1512
+ },
1513
+ {
1514
+ "epoch": 1.11,
1515
+ "learning_rate": 0.001,
1516
+ "loss": 0.8642,
1517
+ "step": 125000
1518
+ },
1519
+ {
1520
+ "epoch": 1.12,
1521
+ "learning_rate": 0.001,
1522
+ "loss": 0.8721,
1523
+ "step": 125500
1524
+ },
1525
+ {
1526
+ "epoch": 1.12,
1527
+ "learning_rate": 0.001,
1528
+ "loss": 0.8672,
1529
+ "step": 126000
1530
+ },
1531
+ {
1532
+ "epoch": 1.12,
1533
+ "learning_rate": 0.001,
1534
+ "loss": 0.8628,
1535
+ "step": 126500
1536
+ },
1537
+ {
1538
+ "epoch": 1.13,
1539
+ "learning_rate": 0.001,
1540
+ "loss": 0.8731,
1541
+ "step": 127000
1542
+ },
1543
+ {
1544
+ "epoch": 1.13,
1545
+ "learning_rate": 0.001,
1546
+ "loss": 0.8753,
1547
+ "step": 127500
1548
+ },
1549
+ {
1550
+ "epoch": 1.14,
1551
+ "learning_rate": 0.001,
1552
+ "loss": 0.8566,
1553
+ "step": 128000
1554
+ },
1555
+ {
1556
+ "epoch": 1.14,
1557
+ "learning_rate": 0.001,
1558
+ "loss": 0.8601,
1559
+ "step": 128500
1560
+ },
1561
+ {
1562
+ "epoch": 1.15,
1563
+ "learning_rate": 0.001,
1564
+ "loss": 0.8622,
1565
+ "step": 129000
1566
+ },
1567
+ {
1568
+ "epoch": 1.15,
1569
+ "learning_rate": 0.001,
1570
+ "loss": 0.8706,
1571
+ "step": 129500
1572
+ },
1573
+ {
1574
+ "epoch": 1.16,
1575
+ "learning_rate": 0.001,
1576
+ "loss": 0.8695,
1577
+ "step": 130000
1578
+ },
1579
+ {
1580
+ "epoch": 1.16,
1581
+ "learning_rate": 0.001,
1582
+ "loss": 0.8633,
1583
+ "step": 130500
1584
+ },
1585
+ {
1586
+ "epoch": 1.16,
1587
+ "learning_rate": 0.001,
1588
+ "loss": 0.8651,
1589
+ "step": 131000
1590
+ },
1591
+ {
1592
+ "epoch": 1.17,
1593
+ "learning_rate": 0.001,
1594
+ "loss": 0.8501,
1595
+ "step": 131500
1596
+ },
1597
+ {
1598
+ "epoch": 1.17,
1599
+ "learning_rate": 0.001,
1600
+ "loss": 0.86,
1601
+ "step": 132000
1602
+ },
1603
+ {
1604
+ "epoch": 1.18,
1605
+ "learning_rate": 0.001,
1606
+ "loss": 0.8713,
1607
+ "step": 132500
1608
+ },
1609
+ {
1610
+ "epoch": 1.18,
1611
+ "learning_rate": 0.001,
1612
+ "loss": 0.8549,
1613
+ "step": 133000
1614
+ },
1615
+ {
1616
+ "epoch": 1.19,
1617
+ "learning_rate": 0.001,
1618
+ "loss": 0.8755,
1619
+ "step": 133500
1620
+ },
1621
+ {
1622
+ "epoch": 1.19,
1623
+ "learning_rate": 0.001,
1624
+ "loss": 0.8511,
1625
+ "step": 134000
1626
+ },
1627
+ {
1628
+ "epoch": 1.2,
1629
+ "learning_rate": 0.001,
1630
+ "loss": 0.8587,
1631
+ "step": 134500
1632
+ },
1633
+ {
1634
+ "epoch": 1.2,
1635
+ "learning_rate": 0.001,
1636
+ "loss": 0.8637,
1637
+ "step": 135000
1638
+ },
1639
+ {
1640
+ "epoch": 1.2,
1641
+ "learning_rate": 0.001,
1642
+ "loss": 0.8625,
1643
+ "step": 135500
1644
+ },
1645
+ {
1646
+ "epoch": 1.21,
1647
+ "learning_rate": 0.001,
1648
+ "loss": 0.8718,
1649
+ "step": 136000
1650
+ },
1651
+ {
1652
+ "epoch": 1.21,
1653
+ "learning_rate": 0.001,
1654
+ "loss": 0.8617,
1655
+ "step": 136500
1656
+ },
1657
+ {
1658
+ "epoch": 1.22,
1659
+ "learning_rate": 0.001,
1660
+ "loss": 0.8564,
1661
+ "step": 137000
1662
+ },
1663
+ {
1664
+ "epoch": 1.22,
1665
+ "learning_rate": 0.001,
1666
+ "loss": 0.859,
1667
+ "step": 137500
1668
+ },
1669
+ {
1670
+ "epoch": 1.23,
1671
+ "learning_rate": 0.001,
1672
+ "loss": 0.8669,
1673
+ "step": 138000
1674
+ },
1675
+ {
1676
+ "epoch": 1.23,
1677
+ "learning_rate": 0.001,
1678
+ "loss": 0.8575,
1679
+ "step": 138500
1680
+ },
1681
+ {
1682
+ "epoch": 1.24,
1683
+ "learning_rate": 0.001,
1684
+ "loss": 0.856,
1685
+ "step": 139000
1686
+ },
1687
+ {
1688
+ "epoch": 1.24,
1689
+ "learning_rate": 0.001,
1690
+ "loss": 0.847,
1691
+ "step": 139500
1692
+ },
1693
+ {
1694
+ "epoch": 1.24,
1695
+ "learning_rate": 0.001,
1696
+ "loss": 0.8527,
1697
+ "step": 140000
1698
+ },
1699
+ {
1700
+ "epoch": 1.25,
1701
+ "learning_rate": 0.001,
1702
+ "loss": 0.8722,
1703
+ "step": 140500
1704
+ },
1705
+ {
1706
+ "epoch": 1.25,
1707
+ "learning_rate": 0.001,
1708
+ "loss": 0.8562,
1709
+ "step": 141000
1710
+ },
1711
+ {
1712
+ "epoch": 1.26,
1713
+ "learning_rate": 0.001,
1714
+ "loss": 0.8549,
1715
+ "step": 141500
1716
+ },
1717
+ {
1718
+ "epoch": 1.26,
1719
+ "learning_rate": 0.001,
1720
+ "loss": 0.8518,
1721
+ "step": 142000
1722
+ },
1723
+ {
1724
+ "epoch": 1.27,
1725
+ "learning_rate": 0.001,
1726
+ "loss": 0.8492,
1727
+ "step": 142500
1728
+ },
1729
+ {
1730
+ "epoch": 1.27,
1731
+ "learning_rate": 0.001,
1732
+ "loss": 0.8715,
1733
+ "step": 143000
1734
+ },
1735
+ {
1736
+ "epoch": 1.28,
1737
+ "learning_rate": 0.001,
1738
+ "loss": 0.8469,
1739
+ "step": 143500
1740
+ },
1741
+ {
1742
+ "epoch": 1.28,
1743
+ "learning_rate": 0.001,
1744
+ "loss": 0.8494,
1745
+ "step": 144000
1746
+ },
1747
+ {
1748
+ "epoch": 1.28,
1749
+ "learning_rate": 0.001,
1750
+ "loss": 0.8518,
1751
+ "step": 144500
1752
+ },
1753
+ {
1754
+ "epoch": 1.29,
1755
+ "learning_rate": 0.001,
1756
+ "loss": 0.8491,
1757
+ "step": 145000
1758
+ },
1759
+ {
1760
+ "epoch": 1.29,
1761
+ "learning_rate": 0.001,
1762
+ "loss": 0.861,
1763
+ "step": 145500
1764
+ },
1765
+ {
1766
+ "epoch": 1.3,
1767
+ "learning_rate": 0.001,
1768
+ "loss": 0.8455,
1769
+ "step": 146000
1770
+ },
1771
+ {
1772
+ "epoch": 1.3,
1773
+ "learning_rate": 0.001,
1774
+ "loss": 0.8601,
1775
+ "step": 146500
1776
+ },
1777
+ {
1778
+ "epoch": 1.31,
1779
+ "learning_rate": 0.001,
1780
+ "loss": 0.8354,
1781
+ "step": 147000
1782
+ },
1783
+ {
1784
+ "epoch": 1.31,
1785
+ "learning_rate": 0.001,
1786
+ "loss": 0.8499,
1787
+ "step": 147500
1788
+ },
1789
+ {
1790
+ "epoch": 1.32,
1791
+ "learning_rate": 0.001,
1792
+ "loss": 0.8503,
1793
+ "step": 148000
1794
+ },
1795
+ {
1796
+ "epoch": 1.32,
1797
+ "learning_rate": 0.001,
1798
+ "loss": 0.8558,
1799
+ "step": 148500
1800
+ },
1801
+ {
1802
+ "epoch": 1.32,
1803
+ "learning_rate": 0.001,
1804
+ "loss": 0.8644,
1805
+ "step": 149000
1806
+ },
1807
+ {
1808
+ "epoch": 1.33,
1809
+ "learning_rate": 0.001,
1810
+ "loss": 0.8425,
1811
+ "step": 149500
1812
+ },
1813
+ {
1814
+ "epoch": 1.33,
1815
+ "learning_rate": 0.001,
1816
+ "loss": 0.8499,
1817
+ "step": 150000
1818
+ },
1819
+ {
1820
+ "epoch": 1.34,
1821
+ "learning_rate": 0.001,
1822
+ "loss": 0.8523,
1823
+ "step": 150500
1824
+ },
1825
+ {
1826
+ "epoch": 1.34,
1827
+ "learning_rate": 0.001,
1828
+ "loss": 0.8515,
1829
+ "step": 151000
1830
+ },
1831
+ {
1832
+ "epoch": 1.35,
1833
+ "learning_rate": 0.001,
1834
+ "loss": 0.8603,
1835
+ "step": 151500
1836
+ },
1837
+ {
1838
+ "epoch": 1.35,
1839
+ "learning_rate": 0.001,
1840
+ "loss": 0.8536,
1841
+ "step": 152000
1842
+ },
1843
+ {
1844
+ "epoch": 1.36,
1845
+ "learning_rate": 0.001,
1846
+ "loss": 0.8479,
1847
+ "step": 152500
1848
+ },
1849
+ {
1850
+ "epoch": 1.36,
1851
+ "learning_rate": 0.001,
1852
+ "loss": 0.8473,
1853
+ "step": 153000
1854
+ },
1855
+ {
1856
+ "epoch": 1.36,
1857
+ "learning_rate": 0.001,
1858
+ "loss": 0.8405,
1859
+ "step": 153500
1860
+ },
1861
+ {
1862
+ "epoch": 1.37,
1863
+ "learning_rate": 0.001,
1864
+ "loss": 0.8304,
1865
+ "step": 154000
1866
+ },
1867
+ {
1868
+ "epoch": 1.37,
1869
+ "learning_rate": 0.001,
1870
+ "loss": 0.8331,
1871
+ "step": 154500
1872
+ },
1873
+ {
1874
+ "epoch": 1.38,
1875
+ "learning_rate": 0.001,
1876
+ "loss": 0.8403,
1877
+ "step": 155000
1878
+ },
1879
+ {
1880
+ "epoch": 1.38,
1881
+ "learning_rate": 0.001,
1882
+ "loss": 0.8535,
1883
+ "step": 155500
1884
+ },
1885
+ {
1886
+ "epoch": 1.39,
1887
+ "learning_rate": 0.001,
1888
+ "loss": 0.8434,
1889
+ "step": 156000
1890
+ },
1891
+ {
1892
+ "epoch": 1.39,
1893
+ "learning_rate": 0.001,
1894
+ "loss": 0.8458,
1895
+ "step": 156500
1896
+ },
1897
+ {
1898
+ "epoch": 1.4,
1899
+ "learning_rate": 0.001,
1900
+ "loss": 0.8428,
1901
+ "step": 157000
1902
+ },
1903
+ {
1904
+ "epoch": 1.4,
1905
+ "learning_rate": 0.001,
1906
+ "loss": 0.8446,
1907
+ "step": 157500
1908
+ },
1909
+ {
1910
+ "epoch": 1.4,
1911
+ "learning_rate": 0.001,
1912
+ "loss": 0.8419,
1913
+ "step": 158000
1914
+ },
1915
+ {
1916
+ "epoch": 1.41,
1917
+ "learning_rate": 0.001,
1918
+ "loss": 0.8543,
1919
+ "step": 158500
1920
+ },
1921
+ {
1922
+ "epoch": 1.41,
1923
+ "learning_rate": 0.001,
1924
+ "loss": 0.8475,
1925
+ "step": 159000
1926
+ },
1927
+ {
1928
+ "epoch": 1.42,
1929
+ "learning_rate": 0.001,
1930
+ "loss": 0.8414,
1931
+ "step": 159500
1932
+ },
1933
+ {
1934
+ "epoch": 1.42,
1935
+ "learning_rate": 0.001,
1936
+ "loss": 0.8504,
1937
+ "step": 160000
1938
+ },
1939
+ {
1940
+ "epoch": 1.43,
1941
+ "learning_rate": 0.001,
1942
+ "loss": 0.8609,
1943
+ "step": 160500
1944
+ },
1945
+ {
1946
+ "epoch": 1.43,
1947
+ "learning_rate": 0.001,
1948
+ "loss": 0.82,
1949
+ "step": 161000
1950
+ },
1951
+ {
1952
+ "epoch": 1.44,
1953
+ "learning_rate": 0.001,
1954
+ "loss": 0.835,
1955
+ "step": 161500
1956
+ },
1957
+ {
1958
+ "epoch": 1.44,
1959
+ "learning_rate": 0.001,
1960
+ "loss": 0.8393,
1961
+ "step": 162000
1962
+ },
1963
+ {
1964
+ "epoch": 1.44,
1965
+ "learning_rate": 0.001,
1966
+ "loss": 0.8514,
1967
+ "step": 162500
1968
+ },
1969
+ {
1970
+ "epoch": 1.45,
1971
+ "learning_rate": 0.001,
1972
+ "loss": 0.8474,
1973
+ "step": 163000
1974
+ },
1975
+ {
1976
+ "epoch": 1.45,
1977
+ "learning_rate": 0.001,
1978
+ "loss": 0.8449,
1979
+ "step": 163500
1980
+ },
1981
+ {
1982
+ "epoch": 1.46,
1983
+ "learning_rate": 0.001,
1984
+ "loss": 0.8425,
1985
+ "step": 164000
1986
+ },
1987
+ {
1988
+ "epoch": 1.46,
1989
+ "learning_rate": 0.001,
1990
+ "loss": 0.8369,
1991
+ "step": 164500
1992
+ },
1993
+ {
1994
+ "epoch": 1.47,
1995
+ "learning_rate": 0.001,
1996
+ "loss": 0.8265,
1997
+ "step": 165000
1998
+ },
1999
+ {
2000
+ "epoch": 1.47,
2001
+ "learning_rate": 0.001,
2002
+ "loss": 0.8347,
2003
+ "step": 165500
2004
+ },
2005
+ {
2006
+ "epoch": 1.48,
2007
+ "learning_rate": 0.001,
2008
+ "loss": 0.8335,
2009
+ "step": 166000
2010
+ },
2011
+ {
2012
+ "epoch": 1.48,
2013
+ "learning_rate": 0.001,
2014
+ "loss": 0.8301,
2015
+ "step": 166500
2016
+ },
2017
+ {
2018
+ "epoch": 1.48,
2019
+ "learning_rate": 0.001,
2020
+ "loss": 0.8308,
2021
+ "step": 167000
2022
+ },
2023
+ {
2024
+ "epoch": 1.49,
2025
+ "learning_rate": 0.001,
2026
+ "loss": 0.8444,
2027
+ "step": 167500
2028
+ },
2029
+ {
2030
+ "epoch": 1.49,
2031
+ "learning_rate": 0.001,
2032
+ "loss": 0.831,
2033
+ "step": 168000
2034
+ },
2035
+ {
2036
+ "epoch": 1.5,
2037
+ "learning_rate": 0.001,
2038
+ "loss": 0.8535,
2039
+ "step": 168500
2040
+ },
2041
+ {
2042
+ "epoch": 1.5,
2043
+ "learning_rate": 0.001,
2044
+ "loss": 0.8472,
2045
+ "step": 169000
2046
+ },
2047
+ {
2048
+ "epoch": 1.51,
2049
+ "learning_rate": 0.001,
2050
+ "loss": 0.8475,
2051
+ "step": 169500
2052
+ },
2053
+ {
2054
+ "epoch": 1.51,
2055
+ "learning_rate": 0.001,
2056
+ "loss": 0.8209,
2057
+ "step": 170000
2058
+ },
2059
+ {
2060
+ "epoch": 1.52,
2061
+ "learning_rate": 0.001,
2062
+ "loss": 0.8422,
2063
+ "step": 170500
2064
+ },
2065
+ {
2066
+ "epoch": 1.52,
2067
+ "learning_rate": 0.001,
2068
+ "loss": 0.8297,
2069
+ "step": 171000
2070
+ },
2071
+ {
2072
+ "epoch": 1.52,
2073
+ "learning_rate": 0.001,
2074
+ "loss": 0.8287,
2075
+ "step": 171500
2076
+ },
2077
+ {
2078
+ "epoch": 1.53,
2079
+ "learning_rate": 0.001,
2080
+ "loss": 0.8265,
2081
+ "step": 172000
2082
+ },
2083
+ {
2084
+ "epoch": 1.53,
2085
+ "learning_rate": 0.001,
2086
+ "loss": 0.8265,
2087
+ "step": 172500
2088
+ },
2089
+ {
2090
+ "epoch": 1.54,
2091
+ "learning_rate": 0.001,
2092
+ "loss": 0.8394,
2093
+ "step": 173000
2094
+ },
2095
+ {
2096
+ "epoch": 1.54,
2097
+ "learning_rate": 0.001,
2098
+ "loss": 0.8364,
2099
+ "step": 173500
2100
+ },
2101
+ {
2102
+ "epoch": 1.55,
2103
+ "learning_rate": 0.001,
2104
+ "loss": 0.8242,
2105
+ "step": 174000
2106
+ },
2107
+ {
2108
+ "epoch": 1.55,
2109
+ "learning_rate": 0.001,
2110
+ "loss": 0.8311,
2111
+ "step": 174500
2112
+ },
2113
+ {
2114
+ "epoch": 1.56,
2115
+ "learning_rate": 0.001,
2116
+ "loss": 0.822,
2117
+ "step": 175000
2118
+ },
2119
+ {
2120
+ "epoch": 1.56,
2121
+ "learning_rate": 0.001,
2122
+ "loss": 0.8245,
2123
+ "step": 175500
2124
+ },
2125
+ {
2126
+ "epoch": 1.56,
2127
+ "learning_rate": 0.001,
2128
+ "loss": 0.8281,
2129
+ "step": 176000
2130
+ },
2131
+ {
2132
+ "epoch": 1.57,
2133
+ "learning_rate": 0.001,
2134
+ "loss": 0.8208,
2135
+ "step": 176500
2136
+ },
2137
+ {
2138
+ "epoch": 1.57,
2139
+ "learning_rate": 0.001,
2140
+ "loss": 0.8328,
2141
+ "step": 177000
2142
+ },
2143
+ {
2144
+ "epoch": 1.58,
2145
+ "learning_rate": 0.001,
2146
+ "loss": 0.83,
2147
+ "step": 177500
2148
+ },
2149
+ {
2150
+ "epoch": 1.58,
2151
+ "learning_rate": 0.001,
2152
+ "loss": 0.8314,
2153
+ "step": 178000
2154
+ },
2155
+ {
2156
+ "epoch": 1.59,
2157
+ "learning_rate": 0.001,
2158
+ "loss": 0.822,
2159
+ "step": 178500
2160
+ },
2161
+ {
2162
+ "epoch": 1.59,
2163
+ "learning_rate": 0.001,
2164
+ "loss": 0.8189,
2165
+ "step": 179000
2166
+ },
2167
+ {
2168
+ "epoch": 1.6,
2169
+ "learning_rate": 0.001,
2170
+ "loss": 0.8221,
2171
+ "step": 179500
2172
+ },
2173
+ {
2174
+ "epoch": 1.6,
2175
+ "learning_rate": 0.001,
2176
+ "loss": 0.8187,
2177
+ "step": 180000
2178
+ },
2179
+ {
2180
+ "epoch": 1.6,
2181
+ "learning_rate": 0.001,
2182
+ "loss": 0.8311,
2183
+ "step": 180500
2184
+ },
2185
+ {
2186
+ "epoch": 1.61,
2187
+ "learning_rate": 0.001,
2188
+ "loss": 0.8193,
2189
+ "step": 181000
2190
+ },
2191
+ {
2192
+ "epoch": 1.61,
2193
+ "learning_rate": 0.001,
2194
+ "loss": 0.8208,
2195
+ "step": 181500
2196
+ },
2197
+ {
2198
+ "epoch": 1.62,
2199
+ "learning_rate": 0.001,
2200
+ "loss": 0.8305,
2201
+ "step": 182000
2202
+ },
2203
+ {
2204
+ "epoch": 1.62,
2205
+ "learning_rate": 0.001,
2206
+ "loss": 0.8302,
2207
+ "step": 182500
2208
+ },
2209
+ {
2210
+ "epoch": 1.63,
2211
+ "learning_rate": 0.001,
2212
+ "loss": 0.8354,
2213
+ "step": 183000
2214
+ },
2215
+ {
2216
+ "epoch": 1.63,
2217
+ "learning_rate": 0.001,
2218
+ "loss": 0.8353,
2219
+ "step": 183500
2220
+ },
2221
+ {
2222
+ "epoch": 1.64,
2223
+ "learning_rate": 0.001,
2224
+ "loss": 0.8353,
2225
+ "step": 184000
2226
+ },
2227
+ {
2228
+ "epoch": 1.64,
2229
+ "learning_rate": 0.001,
2230
+ "loss": 0.8325,
2231
+ "step": 184500
2232
+ },
2233
+ {
2234
+ "epoch": 1.64,
2235
+ "learning_rate": 0.001,
2236
+ "loss": 0.8132,
2237
+ "step": 185000
2238
+ },
2239
+ {
2240
+ "epoch": 1.65,
2241
+ "learning_rate": 0.001,
2242
+ "loss": 0.8275,
2243
+ "step": 185500
2244
+ },
2245
+ {
2246
+ "epoch": 1.65,
2247
+ "learning_rate": 0.001,
2248
+ "loss": 0.8165,
2249
+ "step": 186000
2250
+ },
2251
+ {
2252
+ "epoch": 1.66,
2253
+ "learning_rate": 0.001,
2254
+ "loss": 0.8165,
2255
+ "step": 186500
2256
+ },
2257
+ {
2258
+ "epoch": 1.66,
2259
+ "learning_rate": 0.001,
2260
+ "loss": 0.8246,
2261
+ "step": 187000
2262
+ },
2263
+ {
2264
+ "epoch": 1.67,
2265
+ "learning_rate": 0.001,
2266
+ "loss": 0.8358,
2267
+ "step": 187500
2268
+ },
2269
+ {
2270
+ "epoch": 1.67,
2271
+ "learning_rate": 0.001,
2272
+ "loss": 0.8265,
2273
+ "step": 188000
2274
+ },
2275
+ {
2276
+ "epoch": 1.68,
2277
+ "learning_rate": 0.001,
2278
+ "loss": 0.8096,
2279
+ "step": 188500
2280
+ },
2281
+ {
2282
+ "epoch": 1.68,
2283
+ "learning_rate": 0.001,
2284
+ "loss": 0.813,
2285
+ "step": 189000
2286
+ },
2287
+ {
2288
+ "epoch": 1.68,
2289
+ "learning_rate": 0.001,
2290
+ "loss": 0.8203,
2291
+ "step": 189500
2292
+ },
2293
+ {
2294
+ "epoch": 1.69,
2295
+ "learning_rate": 0.001,
2296
+ "loss": 0.8272,
2297
+ "step": 190000
2298
+ },
2299
+ {
2300
+ "epoch": 1.69,
2301
+ "learning_rate": 0.001,
2302
+ "loss": 0.8114,
2303
+ "step": 190500
2304
+ },
2305
+ {
2306
+ "epoch": 1.7,
2307
+ "learning_rate": 0.001,
2308
+ "loss": 0.8045,
2309
+ "step": 191000
2310
+ },
2311
+ {
2312
+ "epoch": 1.7,
2313
+ "learning_rate": 0.001,
2314
+ "loss": 0.8295,
2315
+ "step": 191500
2316
+ },
2317
+ {
2318
+ "epoch": 1.71,
2319
+ "learning_rate": 0.001,
2320
+ "loss": 0.8153,
2321
+ "step": 192000
2322
+ },
2323
+ {
2324
+ "epoch": 1.71,
2325
+ "learning_rate": 0.001,
2326
+ "loss": 0.8181,
2327
+ "step": 192500
2328
+ },
2329
+ {
2330
+ "epoch": 1.72,
2331
+ "learning_rate": 0.001,
2332
+ "loss": 0.8101,
2333
+ "step": 193000
2334
+ },
2335
+ {
2336
+ "epoch": 1.72,
2337
+ "learning_rate": 0.001,
2338
+ "loss": 0.816,
2339
+ "step": 193500
2340
+ },
2341
+ {
2342
+ "epoch": 1.72,
2343
+ "learning_rate": 0.001,
2344
+ "loss": 0.8139,
2345
+ "step": 194000
2346
+ },
2347
+ {
2348
+ "epoch": 1.73,
2349
+ "learning_rate": 0.001,
2350
+ "loss": 0.8126,
2351
+ "step": 194500
2352
+ },
2353
+ {
2354
+ "epoch": 1.73,
2355
+ "learning_rate": 0.001,
2356
+ "loss": 0.8168,
2357
+ "step": 195000
2358
+ },
2359
+ {
2360
+ "epoch": 1.74,
2361
+ "learning_rate": 0.001,
2362
+ "loss": 0.8348,
2363
+ "step": 195500
2364
+ },
2365
+ {
2366
+ "epoch": 1.74,
2367
+ "learning_rate": 0.001,
2368
+ "loss": 0.8248,
2369
+ "step": 196000
2370
+ },
2371
+ {
2372
+ "epoch": 1.75,
2373
+ "learning_rate": 0.001,
2374
+ "loss": 0.8201,
2375
+ "step": 196500
2376
+ },
2377
+ {
2378
+ "epoch": 1.75,
2379
+ "learning_rate": 0.001,
2380
+ "loss": 0.8142,
2381
+ "step": 197000
2382
+ },
2383
+ {
2384
+ "epoch": 1.76,
2385
+ "learning_rate": 0.001,
2386
+ "loss": 0.8046,
2387
+ "step": 197500
2388
+ },
2389
+ {
2390
+ "epoch": 1.76,
2391
+ "learning_rate": 0.001,
2392
+ "loss": 0.8027,
2393
+ "step": 198000
2394
+ },
2395
+ {
2396
+ "epoch": 1.76,
2397
+ "learning_rate": 0.001,
2398
+ "loss": 0.8127,
2399
+ "step": 198500
2400
+ },
2401
+ {
2402
+ "epoch": 1.77,
2403
+ "learning_rate": 0.001,
2404
+ "loss": 0.8314,
2405
+ "step": 199000
2406
+ },
2407
+ {
2408
+ "epoch": 1.77,
2409
+ "learning_rate": 0.001,
2410
+ "loss": 0.8192,
2411
+ "step": 199500
2412
+ },
2413
+ {
2414
+ "epoch": 1.78,
2415
+ "learning_rate": 0.001,
2416
+ "loss": 0.8005,
2417
+ "step": 200000
2418
+ },
2419
+ {
2420
+ "epoch": 1.78,
2421
+ "learning_rate": 0.001,
2422
+ "loss": 0.8173,
2423
+ "step": 200500
2424
+ },
2425
+ {
2426
+ "epoch": 1.79,
2427
+ "learning_rate": 0.001,
2428
+ "loss": 0.8133,
2429
+ "step": 201000
2430
+ },
2431
+ {
2432
+ "epoch": 1.79,
2433
+ "learning_rate": 0.001,
2434
+ "loss": 0.8144,
2435
+ "step": 201500
2436
+ },
2437
+ {
2438
+ "epoch": 1.8,
2439
+ "learning_rate": 0.001,
2440
+ "loss": 0.8199,
2441
+ "step": 202000
2442
+ },
2443
+ {
2444
+ "epoch": 1.8,
2445
+ "learning_rate": 0.001,
2446
+ "loss": 0.7955,
2447
+ "step": 202500
2448
+ },
2449
+ {
2450
+ "epoch": 1.8,
2451
+ "learning_rate": 0.001,
2452
+ "loss": 0.8012,
2453
+ "step": 203000
2454
+ },
2455
+ {
2456
+ "epoch": 1.81,
2457
+ "learning_rate": 0.001,
2458
+ "loss": 0.807,
2459
+ "step": 203500
2460
+ },
2461
+ {
2462
+ "epoch": 1.81,
2463
+ "learning_rate": 0.001,
2464
+ "loss": 0.8053,
2465
+ "step": 204000
2466
+ },
2467
+ {
2468
+ "epoch": 1.82,
2469
+ "learning_rate": 0.001,
2470
+ "loss": 0.8077,
2471
+ "step": 204500
2472
+ },
2473
+ {
2474
+ "epoch": 1.82,
2475
+ "learning_rate": 0.001,
2476
+ "loss": 0.8047,
2477
+ "step": 205000
2478
+ },
2479
+ {
2480
+ "epoch": 1.83,
2481
+ "learning_rate": 0.001,
2482
+ "loss": 0.7873,
2483
+ "step": 205500
2484
+ },
2485
+ {
2486
+ "epoch": 1.83,
2487
+ "learning_rate": 0.001,
2488
+ "loss": 0.8178,
2489
+ "step": 206000
2490
+ },
2491
+ {
2492
+ "epoch": 1.84,
2493
+ "learning_rate": 0.001,
2494
+ "loss": 0.8041,
2495
+ "step": 206500
2496
+ },
2497
+ {
2498
+ "epoch": 1.84,
2499
+ "learning_rate": 0.001,
2500
+ "loss": 0.8003,
2501
+ "step": 207000
2502
+ },
2503
+ {
2504
+ "epoch": 1.84,
2505
+ "learning_rate": 0.001,
2506
+ "loss": 0.799,
2507
+ "step": 207500
2508
+ },
2509
+ {
2510
+ "epoch": 1.85,
2511
+ "learning_rate": 0.001,
2512
+ "loss": 0.8095,
2513
+ "step": 208000
2514
+ },
2515
+ {
2516
+ "epoch": 1.85,
2517
+ "learning_rate": 0.001,
2518
+ "loss": 0.8133,
2519
+ "step": 208500
2520
+ },
2521
+ {
2522
+ "epoch": 1.86,
2523
+ "learning_rate": 0.001,
2524
+ "loss": 0.8091,
2525
+ "step": 209000
2526
+ },
2527
+ {
2528
+ "epoch": 1.86,
2529
+ "learning_rate": 0.001,
2530
+ "loss": 0.8016,
2531
+ "step": 209500
2532
+ },
2533
+ {
2534
+ "epoch": 1.87,
2535
+ "learning_rate": 0.001,
2536
+ "loss": 0.796,
2537
+ "step": 210000
2538
+ },
2539
+ {
2540
+ "epoch": 1.87,
2541
+ "learning_rate": 0.001,
2542
+ "loss": 0.7999,
2543
+ "step": 210500
2544
+ },
2545
+ {
2546
+ "epoch": 1.88,
2547
+ "learning_rate": 0.001,
2548
+ "loss": 0.8037,
2549
+ "step": 211000
2550
+ },
2551
+ {
2552
+ "epoch": 1.88,
2553
+ "learning_rate": 0.001,
2554
+ "loss": 0.8096,
2555
+ "step": 211500
2556
+ },
2557
+ {
2558
+ "epoch": 1.88,
2559
+ "learning_rate": 0.001,
2560
+ "loss": 0.7965,
2561
+ "step": 212000
2562
+ },
2563
+ {
2564
+ "epoch": 1.89,
2565
+ "learning_rate": 0.001,
2566
+ "loss": 0.8053,
2567
+ "step": 212500
2568
+ },
2569
+ {
2570
+ "epoch": 1.89,
2571
+ "learning_rate": 0.001,
2572
+ "loss": 0.7845,
2573
+ "step": 213000
2574
+ },
2575
+ {
2576
+ "epoch": 1.9,
2577
+ "learning_rate": 0.001,
2578
+ "loss": 0.812,
2579
+ "step": 213500
2580
+ },
2581
+ {
2582
+ "epoch": 1.9,
2583
+ "learning_rate": 0.001,
2584
+ "loss": 0.7981,
2585
+ "step": 214000
2586
+ },
2587
+ {
2588
+ "epoch": 1.91,
2589
+ "learning_rate": 0.001,
2590
+ "loss": 0.7889,
2591
+ "step": 214500
2592
+ },
2593
+ {
2594
+ "epoch": 1.91,
2595
+ "learning_rate": 0.001,
2596
+ "loss": 0.7928,
2597
+ "step": 215000
2598
+ },
2599
+ {
2600
+ "epoch": 1.92,
2601
+ "learning_rate": 0.001,
2602
+ "loss": 0.7977,
2603
+ "step": 215500
2604
+ },
2605
+ {
2606
+ "epoch": 1.92,
2607
+ "learning_rate": 0.001,
2608
+ "loss": 0.8108,
2609
+ "step": 216000
2610
+ },
2611
+ {
2612
+ "epoch": 1.92,
2613
+ "learning_rate": 0.001,
2614
+ "loss": 0.7999,
2615
+ "step": 216500
2616
+ },
2617
+ {
2618
+ "epoch": 1.93,
2619
+ "learning_rate": 0.001,
2620
+ "loss": 0.8031,
2621
+ "step": 217000
2622
+ },
2623
+ {
2624
+ "epoch": 1.93,
2625
+ "learning_rate": 0.001,
2626
+ "loss": 0.8014,
2627
+ "step": 217500
2628
+ },
2629
+ {
2630
+ "epoch": 1.94,
2631
+ "learning_rate": 0.001,
2632
+ "loss": 0.7943,
2633
+ "step": 218000
2634
+ },
2635
+ {
2636
+ "epoch": 1.94,
2637
+ "learning_rate": 0.001,
2638
+ "loss": 0.7998,
2639
+ "step": 218500
2640
+ },
2641
+ {
2642
+ "epoch": 1.95,
2643
+ "learning_rate": 0.001,
2644
+ "loss": 0.8035,
2645
+ "step": 219000
2646
+ },
2647
+ {
2648
+ "epoch": 1.95,
2649
+ "learning_rate": 0.001,
2650
+ "loss": 0.8102,
2651
+ "step": 219500
2652
+ },
2653
+ {
2654
+ "epoch": 1.96,
2655
+ "learning_rate": 0.001,
2656
+ "loss": 0.7945,
2657
+ "step": 220000
2658
+ },
2659
+ {
2660
+ "epoch": 1.96,
2661
+ "learning_rate": 0.001,
2662
+ "loss": 0.7827,
2663
+ "step": 220500
2664
+ },
2665
+ {
2666
+ "epoch": 1.96,
2667
+ "learning_rate": 0.001,
2668
+ "loss": 0.8058,
2669
+ "step": 221000
2670
+ },
2671
+ {
2672
+ "epoch": 1.97,
2673
+ "learning_rate": 0.001,
2674
+ "loss": 0.7898,
2675
+ "step": 221500
2676
+ },
2677
+ {
2678
+ "epoch": 1.97,
2679
+ "learning_rate": 0.001,
2680
+ "loss": 0.8023,
2681
+ "step": 222000
2682
+ },
2683
+ {
2684
+ "epoch": 1.98,
2685
+ "learning_rate": 0.001,
2686
+ "loss": 0.7975,
2687
+ "step": 222500
2688
+ },
2689
+ {
2690
+ "epoch": 1.98,
2691
+ "learning_rate": 0.001,
2692
+ "loss": 0.7971,
2693
+ "step": 223000
2694
+ },
2695
+ {
2696
+ "epoch": 1.99,
2697
+ "learning_rate": 0.001,
2698
+ "loss": 0.7937,
2699
+ "step": 223500
2700
+ },
2701
+ {
2702
+ "epoch": 1.99,
2703
+ "learning_rate": 0.001,
2704
+ "loss": 0.8034,
2705
+ "step": 224000
2706
+ },
2707
+ {
2708
+ "epoch": 2.0,
2709
+ "learning_rate": 0.001,
2710
+ "loss": 0.7894,
2711
+ "step": 224500
2712
+ },
2713
+ {
2714
+ "epoch": 2.0,
2715
+ "learning_rate": 0.001,
2716
+ "loss": 0.7931,
2717
+ "step": 225000
2718
+ },
2719
+ {
2720
+ "epoch": 2.0,
2721
+ "eval_bleu": 25.3566,
2722
+ "eval_gen_len": 11.3335,
2723
+ "eval_loss": 1.1507222652435303,
2724
+ "eval_runtime": 423.4735,
2725
+ "eval_samples_per_second": 33.06,
2726
+ "step": 225000
2727
+ },
2728
+ {
2729
+ "epoch": 2.0,
2730
+ "learning_rate": 0.001,
2731
+ "loss": 0.7253,
2732
+ "step": 225500
2733
+ },
2734
+ {
2735
+ "epoch": 2.01,
2736
+ "learning_rate": 0.001,
2737
+ "loss": 0.7256,
2738
+ "step": 226000
2739
+ },
2740
+ {
2741
+ "epoch": 2.01,
2742
+ "learning_rate": 0.001,
2743
+ "loss": 0.737,
2744
+ "step": 226500
2745
+ },
2746
+ {
2747
+ "epoch": 2.02,
2748
+ "learning_rate": 0.001,
2749
+ "loss": 0.7269,
2750
+ "step": 227000
2751
+ },
2752
+ {
2753
+ "epoch": 2.02,
2754
+ "learning_rate": 0.001,
2755
+ "loss": 0.7416,
2756
+ "step": 227500
2757
+ },
2758
+ {
2759
+ "epoch": 2.03,
2760
+ "learning_rate": 0.001,
2761
+ "loss": 0.7401,
2762
+ "step": 228000
2763
+ },
2764
+ {
2765
+ "epoch": 2.03,
2766
+ "learning_rate": 0.001,
2767
+ "loss": 0.7468,
2768
+ "step": 228500
2769
+ },
2770
+ {
2771
+ "epoch": 2.04,
2772
+ "learning_rate": 0.001,
2773
+ "loss": 0.7392,
2774
+ "step": 229000
2775
+ },
2776
+ {
2777
+ "epoch": 2.04,
2778
+ "learning_rate": 0.001,
2779
+ "loss": 0.728,
2780
+ "step": 229500
2781
+ },
2782
+ {
2783
+ "epoch": 2.04,
2784
+ "learning_rate": 0.001,
2785
+ "loss": 0.741,
2786
+ "step": 230000
2787
+ },
2788
+ {
2789
+ "epoch": 2.05,
2790
+ "learning_rate": 0.001,
2791
+ "loss": 0.7415,
2792
+ "step": 230500
2793
+ },
2794
+ {
2795
+ "epoch": 2.05,
2796
+ "learning_rate": 0.001,
2797
+ "loss": 0.721,
2798
+ "step": 231000
2799
+ },
2800
+ {
2801
+ "epoch": 2.06,
2802
+ "learning_rate": 0.001,
2803
+ "loss": 0.7308,
2804
+ "step": 231500
2805
+ },
2806
+ {
2807
+ "epoch": 2.06,
2808
+ "learning_rate": 0.001,
2809
+ "loss": 0.7378,
2810
+ "step": 232000
2811
+ },
2812
+ {
2813
+ "epoch": 2.07,
2814
+ "learning_rate": 0.001,
2815
+ "loss": 0.7433,
2816
+ "step": 232500
2817
+ },
2818
+ {
2819
+ "epoch": 2.07,
2820
+ "learning_rate": 0.001,
2821
+ "loss": 0.7267,
2822
+ "step": 233000
2823
+ },
2824
+ {
2825
+ "epoch": 2.08,
2826
+ "learning_rate": 0.001,
2827
+ "loss": 0.7335,
2828
+ "step": 233500
2829
+ },
2830
+ {
2831
+ "epoch": 2.08,
2832
+ "learning_rate": 0.001,
2833
+ "loss": 0.7513,
2834
+ "step": 234000
2835
+ },
2836
+ {
2837
+ "epoch": 2.08,
2838
+ "learning_rate": 0.001,
2839
+ "loss": 0.7453,
2840
+ "step": 234500
2841
+ },
2842
+ {
2843
+ "epoch": 2.09,
2844
+ "learning_rate": 0.001,
2845
+ "loss": 0.7346,
2846
+ "step": 235000
2847
+ },
2848
+ {
2849
+ "epoch": 2.09,
2850
+ "learning_rate": 0.001,
2851
+ "loss": 0.7454,
2852
+ "step": 235500
2853
+ },
2854
+ {
2855
+ "epoch": 2.1,
2856
+ "learning_rate": 0.001,
2857
+ "loss": 0.7376,
2858
+ "step": 236000
2859
+ },
2860
+ {
2861
+ "epoch": 2.1,
2862
+ "learning_rate": 0.001,
2863
+ "loss": 0.7269,
2864
+ "step": 236500
2865
+ },
2866
+ {
2867
+ "epoch": 2.11,
2868
+ "learning_rate": 0.001,
2869
+ "loss": 0.7299,
2870
+ "step": 237000
2871
+ },
2872
+ {
2873
+ "epoch": 2.11,
2874
+ "learning_rate": 0.001,
2875
+ "loss": 0.7482,
2876
+ "step": 237500
2877
+ },
2878
+ {
2879
+ "epoch": 2.12,
2880
+ "learning_rate": 0.001,
2881
+ "loss": 0.7489,
2882
+ "step": 238000
2883
+ },
2884
+ {
2885
+ "epoch": 2.12,
2886
+ "learning_rate": 0.001,
2887
+ "loss": 0.7369,
2888
+ "step": 238500
2889
+ },
2890
+ {
2891
+ "epoch": 2.12,
2892
+ "learning_rate": 0.001,
2893
+ "loss": 0.737,
2894
+ "step": 239000
2895
+ },
2896
+ {
2897
+ "epoch": 2.13,
2898
+ "learning_rate": 0.001,
2899
+ "loss": 0.738,
2900
+ "step": 239500
2901
+ },
2902
+ {
2903
+ "epoch": 2.13,
2904
+ "learning_rate": 0.001,
2905
+ "loss": 0.7427,
2906
+ "step": 240000
2907
+ },
2908
+ {
2909
+ "epoch": 2.14,
2910
+ "learning_rate": 0.001,
2911
+ "loss": 0.7427,
2912
+ "step": 240500
2913
+ },
2914
+ {
2915
+ "epoch": 2.14,
2916
+ "learning_rate": 0.001,
2917
+ "loss": 0.7381,
2918
+ "step": 241000
2919
+ },
2920
+ {
2921
+ "epoch": 2.15,
2922
+ "learning_rate": 0.001,
2923
+ "loss": 0.7344,
2924
+ "step": 241500
2925
+ },
2926
+ {
2927
+ "epoch": 2.15,
2928
+ "learning_rate": 0.001,
2929
+ "loss": 0.7378,
2930
+ "step": 242000
2931
+ },
2932
+ {
2933
+ "epoch": 2.16,
2934
+ "learning_rate": 0.001,
2935
+ "loss": 0.7354,
2936
+ "step": 242500
2937
+ },
2938
+ {
2939
+ "epoch": 2.16,
2940
+ "learning_rate": 0.001,
2941
+ "loss": 0.7284,
2942
+ "step": 243000
2943
+ },
2944
+ {
2945
+ "epoch": 2.16,
2946
+ "learning_rate": 0.001,
2947
+ "loss": 0.7332,
2948
+ "step": 243500
2949
+ },
2950
+ {
2951
+ "epoch": 2.17,
2952
+ "learning_rate": 0.001,
2953
+ "loss": 0.7295,
2954
+ "step": 244000
2955
+ },
2956
+ {
2957
+ "epoch": 2.17,
2958
+ "learning_rate": 0.001,
2959
+ "loss": 0.7441,
2960
+ "step": 244500
2961
+ },
2962
+ {
2963
+ "epoch": 2.18,
2964
+ "learning_rate": 0.001,
2965
+ "loss": 0.7482,
2966
+ "step": 245000
2967
+ },
2968
+ {
2969
+ "epoch": 2.18,
2970
+ "learning_rate": 0.001,
2971
+ "loss": 0.7387,
2972
+ "step": 245500
2973
+ },
2974
+ {
2975
+ "epoch": 2.19,
2976
+ "learning_rate": 0.001,
2977
+ "loss": 0.7443,
2978
+ "step": 246000
2979
+ },
2980
+ {
2981
+ "epoch": 2.19,
2982
+ "learning_rate": 0.001,
2983
+ "loss": 0.7458,
2984
+ "step": 246500
2985
+ },
2986
+ {
2987
+ "epoch": 2.2,
2988
+ "learning_rate": 0.001,
2989
+ "loss": 0.7347,
2990
+ "step": 247000
2991
+ },
2992
+ {
2993
+ "epoch": 2.2,
2994
+ "learning_rate": 0.001,
2995
+ "loss": 0.7474,
2996
+ "step": 247500
2997
+ },
2998
+ {
2999
+ "epoch": 2.2,
3000
+ "learning_rate": 0.001,
3001
+ "loss": 0.7404,
3002
+ "step": 248000
3003
+ },
3004
+ {
3005
+ "epoch": 2.21,
3006
+ "learning_rate": 0.001,
3007
+ "loss": 0.7364,
3008
+ "step": 248500
3009
+ },
3010
+ {
3011
+ "epoch": 2.21,
3012
+ "learning_rate": 0.001,
3013
+ "loss": 0.7472,
3014
+ "step": 249000
3015
+ },
3016
+ {
3017
+ "epoch": 2.22,
3018
+ "learning_rate": 0.001,
3019
+ "loss": 0.7474,
3020
+ "step": 249500
3021
+ },
3022
+ {
3023
+ "epoch": 2.22,
3024
+ "learning_rate": 0.001,
3025
+ "loss": 0.7337,
3026
+ "step": 250000
3027
+ },
3028
+ {
3029
+ "epoch": 2.23,
3030
+ "learning_rate": 0.001,
3031
+ "loss": 0.7369,
3032
+ "step": 250500
3033
+ },
3034
+ {
3035
+ "epoch": 2.23,
3036
+ "learning_rate": 0.001,
3037
+ "loss": 0.7431,
3038
+ "step": 251000
3039
+ },
3040
+ {
3041
+ "epoch": 2.24,
3042
+ "learning_rate": 0.001,
3043
+ "loss": 0.7295,
3044
+ "step": 251500
3045
+ },
3046
+ {
3047
+ "epoch": 2.24,
3048
+ "learning_rate": 0.001,
3049
+ "loss": 0.7479,
3050
+ "step": 252000
3051
+ },
3052
+ {
3053
+ "epoch": 2.24,
3054
+ "learning_rate": 0.001,
3055
+ "loss": 0.7457,
3056
+ "step": 252500
3057
+ },
3058
+ {
3059
+ "epoch": 2.25,
3060
+ "learning_rate": 0.001,
3061
+ "loss": 0.741,
3062
+ "step": 253000
3063
+ },
3064
+ {
3065
+ "epoch": 2.25,
3066
+ "learning_rate": 0.001,
3067
+ "loss": 0.7333,
3068
+ "step": 253500
3069
+ },
3070
+ {
3071
+ "epoch": 2.26,
3072
+ "learning_rate": 0.001,
3073
+ "loss": 0.7445,
3074
+ "step": 254000
3075
+ },
3076
+ {
3077
+ "epoch": 2.26,
3078
+ "learning_rate": 0.001,
3079
+ "loss": 0.7334,
3080
+ "step": 254500
3081
+ },
3082
+ {
3083
+ "epoch": 2.27,
3084
+ "learning_rate": 0.001,
3085
+ "loss": 0.7405,
3086
+ "step": 255000
3087
+ },
3088
+ {
3089
+ "epoch": 2.27,
3090
+ "learning_rate": 0.001,
3091
+ "loss": 0.7382,
3092
+ "step": 255500
3093
+ },
3094
+ {
3095
+ "epoch": 2.28,
3096
+ "learning_rate": 0.001,
3097
+ "loss": 0.7421,
3098
+ "step": 256000
3099
+ },
3100
+ {
3101
+ "epoch": 2.28,
3102
+ "learning_rate": 0.001,
3103
+ "loss": 0.7422,
3104
+ "step": 256500
3105
+ },
3106
+ {
3107
+ "epoch": 2.28,
3108
+ "learning_rate": 0.001,
3109
+ "loss": 0.7331,
3110
+ "step": 257000
3111
+ },
3112
+ {
3113
+ "epoch": 2.29,
3114
+ "learning_rate": 0.001,
3115
+ "loss": 0.7303,
3116
+ "step": 257500
3117
+ },
3118
+ {
3119
+ "epoch": 2.29,
3120
+ "learning_rate": 0.001,
3121
+ "loss": 0.7348,
3122
+ "step": 258000
3123
+ },
3124
+ {
3125
+ "epoch": 2.3,
3126
+ "learning_rate": 0.001,
3127
+ "loss": 0.7532,
3128
+ "step": 258500
3129
+ },
3130
+ {
3131
+ "epoch": 2.3,
3132
+ "learning_rate": 0.001,
3133
+ "loss": 0.7396,
3134
+ "step": 259000
3135
+ },
3136
+ {
3137
+ "epoch": 2.31,
3138
+ "learning_rate": 0.001,
3139
+ "loss": 0.7383,
3140
+ "step": 259500
3141
+ },
3142
+ {
3143
+ "epoch": 2.31,
3144
+ "learning_rate": 0.001,
3145
+ "loss": 0.7474,
3146
+ "step": 260000
3147
+ },
3148
+ {
3149
+ "epoch": 2.32,
3150
+ "learning_rate": 0.001,
3151
+ "loss": 0.733,
3152
+ "step": 260500
3153
+ },
3154
+ {
3155
+ "epoch": 2.32,
3156
+ "learning_rate": 0.001,
3157
+ "loss": 0.7368,
3158
+ "step": 261000
3159
+ },
3160
+ {
3161
+ "epoch": 2.32,
3162
+ "learning_rate": 0.001,
3163
+ "loss": 0.7301,
3164
+ "step": 261500
3165
+ },
3166
+ {
3167
+ "epoch": 2.33,
3168
+ "learning_rate": 0.001,
3169
+ "loss": 0.7414,
3170
+ "step": 262000
3171
+ },
3172
+ {
3173
+ "epoch": 2.33,
3174
+ "learning_rate": 0.001,
3175
+ "loss": 0.7374,
3176
+ "step": 262500
3177
+ },
3178
+ {
3179
+ "epoch": 2.34,
3180
+ "learning_rate": 0.001,
3181
+ "loss": 0.7529,
3182
+ "step": 263000
3183
+ },
3184
+ {
3185
+ "epoch": 2.34,
3186
+ "learning_rate": 0.001,
3187
+ "loss": 0.7328,
3188
+ "step": 263500
3189
+ },
3190
+ {
3191
+ "epoch": 2.35,
3192
+ "learning_rate": 0.001,
3193
+ "loss": 0.7442,
3194
+ "step": 264000
3195
+ },
3196
+ {
3197
+ "epoch": 2.35,
3198
+ "learning_rate": 0.001,
3199
+ "loss": 0.7357,
3200
+ "step": 264500
3201
+ },
3202
+ {
3203
+ "epoch": 2.36,
3204
+ "learning_rate": 0.001,
3205
+ "loss": 0.7313,
3206
+ "step": 265000
3207
+ },
3208
+ {
3209
+ "epoch": 2.36,
3210
+ "learning_rate": 0.001,
3211
+ "loss": 0.7404,
3212
+ "step": 265500
3213
+ },
3214
+ {
3215
+ "epoch": 2.36,
3216
+ "learning_rate": 0.001,
3217
+ "loss": 0.7372,
3218
+ "step": 266000
3219
+ },
3220
+ {
3221
+ "epoch": 2.37,
3222
+ "learning_rate": 0.001,
3223
+ "loss": 0.7419,
3224
+ "step": 266500
3225
+ },
3226
+ {
3227
+ "epoch": 2.37,
3228
+ "learning_rate": 0.001,
3229
+ "loss": 0.7305,
3230
+ "step": 267000
3231
+ },
3232
+ {
3233
+ "epoch": 2.38,
3234
+ "learning_rate": 0.001,
3235
+ "loss": 0.7376,
3236
+ "step": 267500
3237
+ },
3238
+ {
3239
+ "epoch": 2.38,
3240
+ "learning_rate": 0.001,
3241
+ "loss": 0.722,
3242
+ "step": 268000
3243
+ },
3244
+ {
3245
+ "epoch": 2.39,
3246
+ "learning_rate": 0.001,
3247
+ "loss": 0.7442,
3248
+ "step": 268500
3249
+ },
3250
+ {
3251
+ "epoch": 2.39,
3252
+ "learning_rate": 0.001,
3253
+ "loss": 0.7323,
3254
+ "step": 269000
3255
+ },
3256
+ {
3257
+ "epoch": 2.4,
3258
+ "learning_rate": 0.001,
3259
+ "loss": 0.7262,
3260
+ "step": 269500
3261
+ },
3262
+ {
3263
+ "epoch": 2.4,
3264
+ "learning_rate": 0.001,
3265
+ "loss": 0.7313,
3266
+ "step": 270000
3267
+ },
3268
+ {
3269
+ "epoch": 2.4,
3270
+ "learning_rate": 0.001,
3271
+ "loss": 0.7305,
3272
+ "step": 270500
3273
+ },
3274
+ {
3275
+ "epoch": 2.41,
3276
+ "learning_rate": 0.001,
3277
+ "loss": 0.7271,
3278
+ "step": 271000
3279
+ },
3280
+ {
3281
+ "epoch": 2.41,
3282
+ "learning_rate": 0.001,
3283
+ "loss": 0.7391,
3284
+ "step": 271500
3285
+ },
3286
+ {
3287
+ "epoch": 2.42,
3288
+ "learning_rate": 0.001,
3289
+ "loss": 0.7361,
3290
+ "step": 272000
3291
+ },
3292
+ {
3293
+ "epoch": 2.42,
3294
+ "learning_rate": 0.001,
3295
+ "loss": 0.7339,
3296
+ "step": 272500
3297
+ },
3298
+ {
3299
+ "epoch": 2.43,
3300
+ "learning_rate": 0.001,
3301
+ "loss": 0.7318,
3302
+ "step": 273000
3303
+ },
3304
+ {
3305
+ "epoch": 2.43,
3306
+ "learning_rate": 0.001,
3307
+ "loss": 0.7278,
3308
+ "step": 273500
3309
+ },
3310
+ {
3311
+ "epoch": 2.44,
3312
+ "learning_rate": 0.001,
3313
+ "loss": 0.7438,
3314
+ "step": 274000
3315
+ },
3316
+ {
3317
+ "epoch": 2.44,
3318
+ "learning_rate": 0.001,
3319
+ "loss": 0.7411,
3320
+ "step": 274500
3321
+ },
3322
+ {
3323
+ "epoch": 2.44,
3324
+ "learning_rate": 0.001,
3325
+ "loss": 0.7447,
3326
+ "step": 275000
3327
+ },
3328
+ {
3329
+ "epoch": 2.45,
3330
+ "learning_rate": 0.001,
3331
+ "loss": 0.7326,
3332
+ "step": 275500
3333
+ },
3334
+ {
3335
+ "epoch": 2.45,
3336
+ "learning_rate": 0.001,
3337
+ "loss": 0.7282,
3338
+ "step": 276000
3339
+ },
3340
+ {
3341
+ "epoch": 2.46,
3342
+ "learning_rate": 0.001,
3343
+ "loss": 0.7312,
3344
+ "step": 276500
3345
+ },
3346
+ {
3347
+ "epoch": 2.46,
3348
+ "learning_rate": 0.001,
3349
+ "loss": 0.7387,
3350
+ "step": 277000
3351
+ },
3352
+ {
3353
+ "epoch": 2.47,
3354
+ "learning_rate": 0.001,
3355
+ "loss": 0.7263,
3356
+ "step": 277500
3357
+ },
3358
+ {
3359
+ "epoch": 2.47,
3360
+ "learning_rate": 0.001,
3361
+ "loss": 0.715,
3362
+ "step": 278000
3363
+ },
3364
+ {
3365
+ "epoch": 2.48,
3366
+ "learning_rate": 0.001,
3367
+ "loss": 0.722,
3368
+ "step": 278500
3369
+ },
3370
+ {
3371
+ "epoch": 2.48,
3372
+ "learning_rate": 0.001,
3373
+ "loss": 0.7268,
3374
+ "step": 279000
3375
+ },
3376
+ {
3377
+ "epoch": 2.48,
3378
+ "learning_rate": 0.001,
3379
+ "loss": 0.7386,
3380
+ "step": 279500
3381
+ },
3382
+ {
3383
+ "epoch": 2.49,
3384
+ "learning_rate": 0.001,
3385
+ "loss": 0.7435,
3386
+ "step": 280000
3387
+ },
3388
+ {
3389
+ "epoch": 2.49,
3390
+ "learning_rate": 0.001,
3391
+ "loss": 0.7273,
3392
+ "step": 280500
3393
+ },
3394
+ {
3395
+ "epoch": 2.5,
3396
+ "learning_rate": 0.001,
3397
+ "loss": 0.7378,
3398
+ "step": 281000
3399
+ },
3400
+ {
3401
+ "epoch": 2.5,
3402
+ "learning_rate": 0.001,
3403
+ "loss": 0.7376,
3404
+ "step": 281500
3405
+ },
3406
+ {
3407
+ "epoch": 2.51,
3408
+ "learning_rate": 0.001,
3409
+ "loss": 0.7397,
3410
+ "step": 282000
3411
+ },
3412
+ {
3413
+ "epoch": 2.51,
3414
+ "learning_rate": 0.001,
3415
+ "loss": 0.7364,
3416
+ "step": 282500
3417
+ },
3418
+ {
3419
+ "epoch": 2.52,
3420
+ "learning_rate": 0.001,
3421
+ "loss": 0.7197,
3422
+ "step": 283000
3423
+ },
3424
+ {
3425
+ "epoch": 2.52,
3426
+ "learning_rate": 0.001,
3427
+ "loss": 0.7421,
3428
+ "step": 283500
3429
+ },
3430
+ {
3431
+ "epoch": 2.52,
3432
+ "learning_rate": 0.001,
3433
+ "loss": 0.7335,
3434
+ "step": 284000
3435
+ },
3436
+ {
3437
+ "epoch": 2.53,
3438
+ "learning_rate": 0.001,
3439
+ "loss": 0.7281,
3440
+ "step": 284500
3441
+ },
3442
+ {
3443
+ "epoch": 2.53,
3444
+ "learning_rate": 0.001,
3445
+ "loss": 0.7263,
3446
+ "step": 285000
3447
+ },
3448
+ {
3449
+ "epoch": 2.54,
3450
+ "learning_rate": 0.001,
3451
+ "loss": 0.7195,
3452
+ "step": 285500
3453
+ },
3454
+ {
3455
+ "epoch": 2.54,
3456
+ "learning_rate": 0.001,
3457
+ "loss": 0.7336,
3458
+ "step": 286000
3459
+ },
3460
+ {
3461
+ "epoch": 2.55,
3462
+ "learning_rate": 0.001,
3463
+ "loss": 0.7307,
3464
+ "step": 286500
3465
+ },
3466
+ {
3467
+ "epoch": 2.55,
3468
+ "learning_rate": 0.001,
3469
+ "loss": 0.7255,
3470
+ "step": 287000
3471
+ },
3472
+ {
3473
+ "epoch": 2.56,
3474
+ "learning_rate": 0.001,
3475
+ "loss": 0.7471,
3476
+ "step": 287500
3477
+ },
3478
+ {
3479
+ "epoch": 2.56,
3480
+ "learning_rate": 0.001,
3481
+ "loss": 0.7307,
3482
+ "step": 288000
3483
+ },
3484
+ {
3485
+ "epoch": 2.56,
3486
+ "learning_rate": 0.001,
3487
+ "loss": 0.7251,
3488
+ "step": 288500
3489
+ },
3490
+ {
3491
+ "epoch": 2.57,
3492
+ "learning_rate": 0.001,
3493
+ "loss": 0.7269,
3494
+ "step": 289000
3495
+ },
3496
+ {
3497
+ "epoch": 2.57,
3498
+ "learning_rate": 0.001,
3499
+ "loss": 0.7324,
3500
+ "step": 289500
3501
+ },
3502
+ {
3503
+ "epoch": 2.58,
3504
+ "learning_rate": 0.001,
3505
+ "loss": 0.7142,
3506
+ "step": 290000
3507
+ },
3508
+ {
3509
+ "epoch": 2.58,
3510
+ "learning_rate": 0.001,
3511
+ "loss": 0.7318,
3512
+ "step": 290500
3513
+ },
3514
+ {
3515
+ "epoch": 2.59,
3516
+ "learning_rate": 0.001,
3517
+ "loss": 0.7236,
3518
+ "step": 291000
3519
+ },
3520
+ {
3521
+ "epoch": 2.59,
3522
+ "learning_rate": 0.001,
3523
+ "loss": 0.7267,
3524
+ "step": 291500
3525
+ },
3526
+ {
3527
+ "epoch": 2.6,
3528
+ "learning_rate": 0.001,
3529
+ "loss": 0.7387,
3530
+ "step": 292000
3531
+ },
3532
+ {
3533
+ "epoch": 2.6,
3534
+ "learning_rate": 0.001,
3535
+ "loss": 0.7299,
3536
+ "step": 292500
3537
+ },
3538
+ {
3539
+ "epoch": 2.6,
3540
+ "learning_rate": 0.001,
3541
+ "loss": 0.7518,
3542
+ "step": 293000
3543
+ },
3544
+ {
3545
+ "epoch": 2.61,
3546
+ "learning_rate": 0.001,
3547
+ "loss": 0.7227,
3548
+ "step": 293500
3549
+ },
3550
+ {
3551
+ "epoch": 2.61,
3552
+ "learning_rate": 0.001,
3553
+ "loss": 0.7261,
3554
+ "step": 294000
3555
+ },
3556
+ {
3557
+ "epoch": 2.62,
3558
+ "learning_rate": 0.001,
3559
+ "loss": 0.7325,
3560
+ "step": 294500
3561
+ },
3562
+ {
3563
+ "epoch": 2.62,
3564
+ "learning_rate": 0.001,
3565
+ "loss": 0.7262,
3566
+ "step": 295000
3567
+ },
3568
+ {
3569
+ "epoch": 2.63,
3570
+ "learning_rate": 0.001,
3571
+ "loss": 0.7226,
3572
+ "step": 295500
3573
+ },
3574
+ {
3575
+ "epoch": 2.63,
3576
+ "learning_rate": 0.001,
3577
+ "loss": 0.7218,
3578
+ "step": 296000
3579
+ },
3580
+ {
3581
+ "epoch": 2.64,
3582
+ "learning_rate": 0.001,
3583
+ "loss": 0.7392,
3584
+ "step": 296500
3585
+ },
3586
+ {
3587
+ "epoch": 2.64,
3588
+ "learning_rate": 0.001,
3589
+ "loss": 0.7262,
3590
+ "step": 297000
3591
+ },
3592
+ {
3593
+ "epoch": 2.64,
3594
+ "learning_rate": 0.001,
3595
+ "loss": 0.7234,
3596
+ "step": 297500
3597
+ },
3598
+ {
3599
+ "epoch": 2.65,
3600
+ "learning_rate": 0.001,
3601
+ "loss": 0.7265,
3602
+ "step": 298000
3603
+ },
3604
+ {
3605
+ "epoch": 2.65,
3606
+ "learning_rate": 0.001,
3607
+ "loss": 0.721,
3608
+ "step": 298500
3609
+ },
3610
+ {
3611
+ "epoch": 2.66,
3612
+ "learning_rate": 0.001,
3613
+ "loss": 0.7163,
3614
+ "step": 299000
3615
+ },
3616
+ {
3617
+ "epoch": 2.66,
3618
+ "learning_rate": 0.001,
3619
+ "loss": 0.7361,
3620
+ "step": 299500
3621
+ },
3622
+ {
3623
+ "epoch": 2.67,
3624
+ "learning_rate": 0.001,
3625
+ "loss": 0.7102,
3626
+ "step": 300000
3627
+ },
3628
+ {
3629
+ "epoch": 2.67,
3630
+ "learning_rate": 0.001,
3631
+ "loss": 0.727,
3632
+ "step": 300500
3633
+ },
3634
+ {
3635
+ "epoch": 2.68,
3636
+ "learning_rate": 0.001,
3637
+ "loss": 0.7309,
3638
+ "step": 301000
3639
+ },
3640
+ {
3641
+ "epoch": 2.68,
3642
+ "learning_rate": 0.001,
3643
+ "loss": 0.7185,
3644
+ "step": 301500
3645
+ },
3646
+ {
3647
+ "epoch": 2.68,
3648
+ "learning_rate": 0.001,
3649
+ "loss": 0.7242,
3650
+ "step": 302000
3651
+ },
3652
+ {
3653
+ "epoch": 2.69,
3654
+ "learning_rate": 0.001,
3655
+ "loss": 0.7308,
3656
+ "step": 302500
3657
+ },
3658
+ {
3659
+ "epoch": 2.69,
3660
+ "learning_rate": 0.001,
3661
+ "loss": 0.7323,
3662
+ "step": 303000
3663
+ },
3664
+ {
3665
+ "epoch": 2.7,
3666
+ "learning_rate": 0.001,
3667
+ "loss": 0.7293,
3668
+ "step": 303500
3669
+ },
3670
+ {
3671
+ "epoch": 2.7,
3672
+ "learning_rate": 0.001,
3673
+ "loss": 0.7078,
3674
+ "step": 304000
3675
+ },
3676
+ {
3677
+ "epoch": 2.71,
3678
+ "learning_rate": 0.001,
3679
+ "loss": 0.719,
3680
+ "step": 304500
3681
+ },
3682
+ {
3683
+ "epoch": 2.71,
3684
+ "learning_rate": 0.001,
3685
+ "loss": 0.7252,
3686
+ "step": 305000
3687
+ },
3688
+ {
3689
+ "epoch": 2.72,
3690
+ "learning_rate": 0.001,
3691
+ "loss": 0.7252,
3692
+ "step": 305500
3693
+ },
3694
+ {
3695
+ "epoch": 2.72,
3696
+ "learning_rate": 0.001,
3697
+ "loss": 0.7243,
3698
+ "step": 306000
3699
+ },
3700
+ {
3701
+ "epoch": 2.72,
3702
+ "learning_rate": 0.001,
3703
+ "loss": 0.7154,
3704
+ "step": 306500
3705
+ },
3706
+ {
3707
+ "epoch": 2.73,
3708
+ "learning_rate": 0.001,
3709
+ "loss": 0.718,
3710
+ "step": 307000
3711
+ },
3712
+ {
3713
+ "epoch": 2.73,
3714
+ "learning_rate": 0.001,
3715
+ "loss": 0.7238,
3716
+ "step": 307500
3717
+ },
3718
+ {
3719
+ "epoch": 2.74,
3720
+ "learning_rate": 0.001,
3721
+ "loss": 0.7181,
3722
+ "step": 308000
3723
+ },
3724
+ {
3725
+ "epoch": 2.74,
3726
+ "learning_rate": 0.001,
3727
+ "loss": 0.7288,
3728
+ "step": 308500
3729
+ },
3730
+ {
3731
+ "epoch": 2.75,
3732
+ "learning_rate": 0.001,
3733
+ "loss": 0.717,
3734
+ "step": 309000
3735
+ },
3736
+ {
3737
+ "epoch": 2.75,
3738
+ "learning_rate": 0.001,
3739
+ "loss": 0.7242,
3740
+ "step": 309500
3741
+ },
3742
+ {
3743
+ "epoch": 2.76,
3744
+ "learning_rate": 0.001,
3745
+ "loss": 0.7265,
3746
+ "step": 310000
3747
+ },
3748
+ {
3749
+ "epoch": 2.76,
3750
+ "learning_rate": 0.001,
3751
+ "loss": 0.7091,
3752
+ "step": 310500
3753
+ },
3754
+ {
3755
+ "epoch": 2.76,
3756
+ "learning_rate": 0.001,
3757
+ "loss": 0.7156,
3758
+ "step": 311000
3759
+ },
3760
+ {
3761
+ "epoch": 2.77,
3762
+ "learning_rate": 0.001,
3763
+ "loss": 0.7215,
3764
+ "step": 311500
3765
+ },
3766
+ {
3767
+ "epoch": 2.77,
3768
+ "learning_rate": 0.001,
3769
+ "loss": 0.719,
3770
+ "step": 312000
3771
+ },
3772
+ {
3773
+ "epoch": 2.78,
3774
+ "learning_rate": 0.001,
3775
+ "loss": 0.7224,
3776
+ "step": 312500
3777
+ },
3778
+ {
3779
+ "epoch": 2.78,
3780
+ "learning_rate": 0.001,
3781
+ "loss": 0.726,
3782
+ "step": 313000
3783
+ },
3784
+ {
3785
+ "epoch": 2.79,
3786
+ "learning_rate": 0.001,
3787
+ "loss": 0.7081,
3788
+ "step": 313500
3789
+ },
3790
+ {
3791
+ "epoch": 2.79,
3792
+ "learning_rate": 0.001,
3793
+ "loss": 0.714,
3794
+ "step": 314000
3795
+ },
3796
+ {
3797
+ "epoch": 2.8,
3798
+ "learning_rate": 0.001,
3799
+ "loss": 0.7191,
3800
+ "step": 314500
3801
+ },
3802
+ {
3803
+ "epoch": 2.8,
3804
+ "learning_rate": 0.001,
3805
+ "loss": 0.7115,
3806
+ "step": 315000
3807
+ },
3808
+ {
3809
+ "epoch": 2.8,
3810
+ "learning_rate": 0.001,
3811
+ "loss": 0.711,
3812
+ "step": 315500
3813
+ },
3814
+ {
3815
+ "epoch": 2.81,
3816
+ "learning_rate": 0.001,
3817
+ "loss": 0.7064,
3818
+ "step": 316000
3819
+ },
3820
+ {
3821
+ "epoch": 2.81,
3822
+ "learning_rate": 0.001,
3823
+ "loss": 0.7096,
3824
+ "step": 316500
3825
+ },
3826
+ {
3827
+ "epoch": 2.82,
3828
+ "learning_rate": 0.001,
3829
+ "loss": 0.7159,
3830
+ "step": 317000
3831
+ },
3832
+ {
3833
+ "epoch": 2.82,
3834
+ "learning_rate": 0.001,
3835
+ "loss": 0.7251,
3836
+ "step": 317500
3837
+ },
3838
+ {
3839
+ "epoch": 2.83,
3840
+ "learning_rate": 0.001,
3841
+ "loss": 0.7117,
3842
+ "step": 318000
3843
+ },
3844
+ {
3845
+ "epoch": 2.83,
3846
+ "learning_rate": 0.001,
3847
+ "loss": 0.7108,
3848
+ "step": 318500
3849
+ },
3850
+ {
3851
+ "epoch": 2.84,
3852
+ "learning_rate": 0.001,
3853
+ "loss": 0.7104,
3854
+ "step": 319000
3855
+ },
3856
+ {
3857
+ "epoch": 2.84,
3858
+ "learning_rate": 0.001,
3859
+ "loss": 0.7087,
3860
+ "step": 319500
3861
+ },
3862
+ {
3863
+ "epoch": 2.84,
3864
+ "learning_rate": 0.001,
3865
+ "loss": 0.7163,
3866
+ "step": 320000
3867
+ },
3868
+ {
3869
+ "epoch": 2.85,
3870
+ "learning_rate": 0.001,
3871
+ "loss": 0.7341,
3872
+ "step": 320500
3873
+ },
3874
+ {
3875
+ "epoch": 2.85,
3876
+ "learning_rate": 0.001,
3877
+ "loss": 0.7153,
3878
+ "step": 321000
3879
+ },
3880
+ {
3881
+ "epoch": 2.86,
3882
+ "learning_rate": 0.001,
3883
+ "loss": 0.717,
3884
+ "step": 321500
3885
+ },
3886
+ {
3887
+ "epoch": 2.86,
3888
+ "learning_rate": 0.001,
3889
+ "loss": 0.7155,
3890
+ "step": 322000
3891
+ },
3892
+ {
3893
+ "epoch": 2.87,
3894
+ "learning_rate": 0.001,
3895
+ "loss": 0.7144,
3896
+ "step": 322500
3897
+ },
3898
+ {
3899
+ "epoch": 2.87,
3900
+ "learning_rate": 0.001,
3901
+ "loss": 0.7093,
3902
+ "step": 323000
3903
+ },
3904
+ {
3905
+ "epoch": 2.88,
3906
+ "learning_rate": 0.001,
3907
+ "loss": 0.7071,
3908
+ "step": 323500
3909
+ },
3910
+ {
3911
+ "epoch": 2.88,
3912
+ "learning_rate": 0.001,
3913
+ "loss": 0.7087,
3914
+ "step": 324000
3915
+ },
3916
+ {
3917
+ "epoch": 2.88,
3918
+ "learning_rate": 0.001,
3919
+ "loss": 0.7374,
3920
+ "step": 324500
3921
+ },
3922
+ {
3923
+ "epoch": 2.89,
3924
+ "learning_rate": 0.001,
3925
+ "loss": 0.7276,
3926
+ "step": 325000
3927
+ },
3928
+ {
3929
+ "epoch": 2.89,
3930
+ "learning_rate": 0.001,
3931
+ "loss": 0.7269,
3932
+ "step": 325500
3933
+ },
3934
+ {
3935
+ "epoch": 2.9,
3936
+ "learning_rate": 0.001,
3937
+ "loss": 0.7252,
3938
+ "step": 326000
3939
+ },
3940
+ {
3941
+ "epoch": 2.9,
3942
+ "learning_rate": 0.001,
3943
+ "loss": 0.7061,
3944
+ "step": 326500
3945
+ },
3946
+ {
3947
+ "epoch": 2.91,
3948
+ "learning_rate": 0.001,
3949
+ "loss": 0.7138,
3950
+ "step": 327000
3951
+ },
3952
+ {
3953
+ "epoch": 2.91,
3954
+ "learning_rate": 0.001,
3955
+ "loss": 0.7033,
3956
+ "step": 327500
3957
+ },
3958
+ {
3959
+ "epoch": 2.92,
3960
+ "learning_rate": 0.001,
3961
+ "loss": 0.7057,
3962
+ "step": 328000
3963
+ },
3964
+ {
3965
+ "epoch": 2.92,
3966
+ "learning_rate": 0.001,
3967
+ "loss": 0.7082,
3968
+ "step": 328500
3969
+ },
3970
+ {
3971
+ "epoch": 2.92,
3972
+ "learning_rate": 0.001,
3973
+ "loss": 0.7056,
3974
+ "step": 329000
3975
+ },
3976
+ {
3977
+ "epoch": 2.93,
3978
+ "learning_rate": 0.001,
3979
+ "loss": 0.7125,
3980
+ "step": 329500
3981
+ },
3982
+ {
3983
+ "epoch": 2.93,
3984
+ "learning_rate": 0.001,
3985
+ "loss": 0.7022,
3986
+ "step": 330000
3987
+ },
3988
+ {
3989
+ "epoch": 2.94,
3990
+ "learning_rate": 0.001,
3991
+ "loss": 0.7132,
3992
+ "step": 330500
3993
+ },
3994
+ {
3995
+ "epoch": 2.94,
3996
+ "learning_rate": 0.001,
3997
+ "loss": 0.7175,
3998
+ "step": 331000
3999
+ },
4000
+ {
4001
+ "epoch": 2.95,
4002
+ "learning_rate": 0.001,
4003
+ "loss": 0.7104,
4004
+ "step": 331500
4005
+ },
4006
+ {
4007
+ "epoch": 2.95,
4008
+ "learning_rate": 0.001,
4009
+ "loss": 0.7158,
4010
+ "step": 332000
4011
+ },
4012
+ {
4013
+ "epoch": 2.96,
4014
+ "learning_rate": 0.001,
4015
+ "loss": 0.7102,
4016
+ "step": 332500
4017
+ }
4018
+ ],
4019
+ "max_steps": 337500,
4020
+ "num_train_epochs": 3,
4021
+ "total_flos": 369594455408640.0,
4022
+ "trial_name": null,
4023
+ "trial_params": null
4024
+ }