Hugging Face's logo

zhiyuanyou
/
DeQA-Score-LoRA-Mix3

English
mplug_owl2
Model card Files Community
zhiyuanyou commited on
Commit
3c1258b
·
1 Parent(s): bbe5e81

init commit

Browse files
Files changed (5) hide show
  1. adapter_config.json +83 -0
  2. adapter_model.bin +3 -0
  3. config.json +209 -0
  4. non_lora_trainables.bin +3 -0
  5. trainer_state.json +3900 -0
adapter_config.json ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "MAGAer13/mplug-owl2-llama2-7b",
4
+ "bias": "none",
5
+ "fan_in_fan_out": false,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 256,
11
+ "lora_dropout": 0.05,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 128,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "model.layers.16.self_attn.v_proj.multiway.1",
18
+ "model.layers.10.self_attn.v_proj.multiway.1",
19
+ "model.layers.26.self_attn.q_proj",
20
+ "model.layers.0.self_attn.v_proj.multiway.1",
21
+ "model.layers.30.self_attn.q_proj",
22
+ "model.layers.24.self_attn.v_proj.multiway.1",
23
+ "model.layers.28.self_attn.v_proj.multiway.1",
24
+ "model.layers.11.self_attn.q_proj",
25
+ "model.layers.2.self_attn.v_proj.multiway.1",
26
+ "model.layers.13.self_attn.q_proj",
27
+ "model.layers.14.self_attn.q_proj",
28
+ "model.layers.19.self_attn.v_proj.multiway.1",
29
+ "model.layers.24.self_attn.q_proj",
30
+ "model.layers.16.self_attn.q_proj",
31
+ "model.layers.25.self_attn.v_proj.multiway.1",
32
+ "model.layers.4.self_attn.q_proj",
33
+ "model.layers.27.self_attn.v_proj.multiway.1",
34
+ "model.layers.9.self_attn.v_proj.multiway.1",
35
+ "model.layers.12.self_attn.q_proj",
36
+ "model.layers.22.self_attn.v_proj.multiway.1",
37
+ "model.layers.31.self_attn.v_proj.multiway.1",
38
+ "model.layers.7.self_attn.q_proj",
39
+ "model.layers.15.self_attn.v_proj.multiway.1",
40
+ "model.layers.22.self_attn.q_proj",
41
+ "model.layers.1.self_attn.q_proj",
42
+ "model.layers.19.self_attn.q_proj",
43
+ "model.layers.23.self_attn.v_proj.multiway.1",
44
+ "model.layers.25.self_attn.q_proj",
45
+ "model.layers.28.self_attn.q_proj",
46
+ "model.layers.21.self_attn.v_proj.multiway.1",
47
+ "model.layers.29.self_attn.q_proj",
48
+ "model.layers.6.self_attn.q_proj",
49
+ "model.layers.17.self_attn.v_proj.multiway.1",
50
+ "model.layers.5.self_attn.v_proj.multiway.1",
51
+ "model.layers.20.self_attn.q_proj",
52
+ "model.layers.31.self_attn.q_proj",
53
+ "model.layers.13.self_attn.v_proj.multiway.1",
54
+ "model.layers.1.self_attn.v_proj.multiway.1",
55
+ "model.layers.10.self_attn.q_proj",
56
+ "model.layers.8.self_attn.q_proj",
57
+ "model.layers.0.self_attn.q_proj",
58
+ "model.layers.2.self_attn.q_proj",
59
+ "model.layers.4.self_attn.v_proj.multiway.1",
60
+ "model.layers.20.self_attn.v_proj.multiway.1",
61
+ "model.layers.23.self_attn.q_proj",
62
+ "model.layers.30.self_attn.v_proj.multiway.1",
63
+ "model.layers.18.self_attn.q_proj",
64
+ "model.layers.9.self_attn.q_proj",
65
+ "model.layers.8.self_attn.v_proj.multiway.1",
66
+ "model.layers.18.self_attn.v_proj.multiway.1",
67
+ "model.layers.3.self_attn.q_proj",
68
+ "model.layers.3.self_attn.v_proj.multiway.1",
69
+ "model.layers.7.self_attn.v_proj.multiway.1",
70
+ "model.layers.12.self_attn.v_proj.multiway.1",
71
+ "model.layers.5.self_attn.q_proj",
72
+ "model.layers.11.self_attn.v_proj.multiway.1",
73
+ "model.layers.14.self_attn.v_proj.multiway.1",
74
+ "model.layers.17.self_attn.q_proj",
75
+ "model.layers.27.self_attn.q_proj",
76
+ "model.layers.21.self_attn.q_proj",
77
+ "model.layers.29.self_attn.v_proj.multiway.1",
78
+ "model.layers.26.self_attn.v_proj.multiway.1",
79
+ "model.layers.6.self_attn.v_proj.multiway.1",
80
+ "model.layers.15.self_attn.q_proj"
81
+ ],
82
+ "task_type": "CAUSAL_LM"
83
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d93ac613f6ae78565e0a08c3b4b517540d022b14741e76bd7f8ef87ef865c85
3
+ size 134262669
config.json ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "zhiyuanyou/DeQA-Score-LoRA-Mix3",
3
+ "attention_bias": false,
4
+ "attention_dropout": 0.0,
5
+ "binary_rating_loss": "fidelity",
6
+ "bos_token_id": 1,
7
+ "closeset_rating_loss": true,
8
+ "continuous_rating_loss": true,
9
+ "detach_pred_std": true,
10
+ "eos_token_id": 2,
11
+ "freeze_vision_model": false,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 4096,
14
+ "image_aspect_ratio": "pad",
15
+ "image_grid_pinpoints": null,
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 11008,
18
+ "level_ids": [
19
+ 15129,
20
+ 1781,
21
+ 6534,
22
+ 6460,
23
+ 4319
24
+ ],
25
+ "level_prefix": [
26
+ 450,
27
+ 11029,
28
+ 310,
29
+ 278,
30
+ 1967,
31
+ 338
32
+ ],
33
+ "max_position_embeddings": 2048,
34
+ "model_type": "mplug_owl2",
35
+ "num_attention_heads": 32,
36
+ "num_hidden_layers": 32,
37
+ "num_key_value_heads": 32,
38
+ "pretraining_tp": 1,
39
+ "rms_norm_eps": 1e-06,
40
+ "rope_scaling": null,
41
+ "rope_theta": 10000.0,
42
+ "softkl_loss": true,
43
+ "tie_word_embeddings": false,
44
+ "transformers_version": "4.36.1",
45
+ "tune_visual_abstractor": true,
46
+ "use_cache": true,
47
+ "use_fix_std": true,
48
+ "visual_abstractor_lr": null,
49
+ "visual_config": {
50
+ "visual_abstractor": {
51
+ "_name_or_path": "",
52
+ "add_cross_attention": false,
53
+ "architectures": null,
54
+ "attention_probs_dropout_prob": 0.0,
55
+ "bad_words_ids": null,
56
+ "begin_suppress_tokens": null,
57
+ "bos_token_id": null,
58
+ "chunk_size_feed_forward": 0,
59
+ "cross_attention_hidden_size": null,
60
+ "decoder_start_token_id": null,
61
+ "diversity_penalty": 0.0,
62
+ "do_sample": false,
63
+ "early_stopping": false,
64
+ "encoder_hidden_size": 1024,
65
+ "encoder_no_repeat_ngram_size": 0,
66
+ "eos_token_id": null,
67
+ "exponential_decay_length_penalty": null,
68
+ "finetuning_task": null,
69
+ "forced_bos_token_id": null,
70
+ "forced_eos_token_id": null,
71
+ "grid_size": 32,
72
+ "hidden_size": 1024,
73
+ "id2label": {
74
+ "0": "LABEL_0",
75
+ "1": "LABEL_1"
76
+ },
77
+ "initializer_range": 0.02,
78
+ "intermediate_size": 2816,
79
+ "is_decoder": false,
80
+ "is_encoder_decoder": false,
81
+ "label2id": {
82
+ "LABEL_0": 0,
83
+ "LABEL_1": 1
84
+ },
85
+ "layer_norm_eps": 1e-06,
86
+ "length_penalty": 1.0,
87
+ "max_length": 20,
88
+ "min_length": 0,
89
+ "model_type": "mplug_owl_visual_abstract",
90
+ "no_repeat_ngram_size": 0,
91
+ "num_attention_heads": 16,
92
+ "num_beam_groups": 1,
93
+ "num_beams": 1,
94
+ "num_hidden_layers": 6,
95
+ "num_learnable_queries": 64,
96
+ "num_return_sequences": 1,
97
+ "output_attentions": false,
98
+ "output_hidden_states": false,
99
+ "output_scores": false,
100
+ "pad_token_id": null,
101
+ "prefix": null,
102
+ "problem_type": null,
103
+ "pruned_heads": {},
104
+ "remove_invalid_values": false,
105
+ "repetition_penalty": 1.0,
106
+ "return_dict": true,
107
+ "return_dict_in_generate": false,
108
+ "sep_token_id": null,
109
+ "suppress_tokens": null,
110
+ "task_specific_params": null,
111
+ "temperature": 1.0,
112
+ "tf_legacy_loss": false,
113
+ "tie_encoder_decoder": false,
114
+ "tie_word_embeddings": true,
115
+ "tokenizer_class": null,
116
+ "top_k": 50,
117
+ "top_p": 1.0,
118
+ "torch_dtype": null,
119
+ "torchscript": false,
120
+ "transformers_version": "4.28.1",
121
+ "typical_p": 1.0,
122
+ "use_bfloat16": false
123
+ },
124
+ "visual_model": {
125
+ "_name_or_path": "",
126
+ "add_cross_attention": false,
127
+ "architectures": null,
128
+ "attention_dropout": 0.0,
129
+ "bad_words_ids": null,
130
+ "begin_suppress_tokens": null,
131
+ "bos_token_id": null,
132
+ "chunk_size_feed_forward": 0,
133
+ "cross_attention_hidden_size": null,
134
+ "decoder_start_token_id": null,
135
+ "diversity_penalty": 0.0,
136
+ "do_sample": false,
137
+ "early_stopping": false,
138
+ "encoder_no_repeat_ngram_size": 0,
139
+ "eos_token_id": null,
140
+ "exponential_decay_length_penalty": null,
141
+ "finetuning_task": null,
142
+ "forced_bos_token_id": null,
143
+ "forced_eos_token_id": null,
144
+ "hidden_act": "quick_gelu",
145
+ "hidden_size": 1024,
146
+ "id2label": {
147
+ "0": "LABEL_0",
148
+ "1": "LABEL_1"
149
+ },
150
+ "image_size": 448,
151
+ "initializer_factor": 1.0,
152
+ "initializer_range": 0.02,
153
+ "intermediate_size": 4096,
154
+ "is_decoder": false,
155
+ "is_encoder_decoder": false,
156
+ "label2id": {
157
+ "LABEL_0": 0,
158
+ "LABEL_1": 1
159
+ },
160
+ "layer_norm_eps": 1e-06,
161
+ "length_penalty": 1.0,
162
+ "max_length": 20,
163
+ "min_length": 0,
164
+ "model_type": "mplug_owl_vision_model",
165
+ "no_repeat_ngram_size": 0,
166
+ "num_attention_heads": 16,
167
+ "num_beam_groups": 1,
168
+ "num_beams": 1,
169
+ "num_channels": 3,
170
+ "num_hidden_layers": 24,
171
+ "num_return_sequences": 1,
172
+ "output_attentions": false,
173
+ "output_hidden_states": false,
174
+ "output_scores": false,
175
+ "pad_token_id": null,
176
+ "patch_size": 14,
177
+ "prefix": null,
178
+ "problem_type": null,
179
+ "projection_dim": 768,
180
+ "pruned_heads": {},
181
+ "remove_invalid_values": false,
182
+ "repetition_penalty": 1.0,
183
+ "return_dict": true,
184
+ "return_dict_in_generate": false,
185
+ "sep_token_id": null,
186
+ "suppress_tokens": null,
187
+ "task_specific_params": null,
188
+ "temperature": 1.0,
189
+ "tf_legacy_loss": false,
190
+ "tie_encoder_decoder": false,
191
+ "tie_word_embeddings": true,
192
+ "tokenizer_class": null,
193
+ "top_k": 50,
194
+ "top_p": 1.0,
195
+ "torch_dtype": null,
196
+ "torchscript": false,
197
+ "transformers_version": "4.28.1",
198
+ "typical_p": 1.0,
199
+ "use_bfloat16": false,
200
+ "use_flash_attn": false
201
+ }
202
+ },
203
+ "vocab_size": 32000,
204
+ "weight_desp": 1.0,
205
+ "weight_in_level": null,
206
+ "weight_next_token": 0.05,
207
+ "weight_rank": 1.0,
208
+ "weight_softkl": 1.0
209
+ }
non_lora_trainables.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:615d6bc146025d551ffe9ef54cc1429acfb0b5030cfd4bda5cb1704eaef9d2dc
3
+ size 771037807
trainer_state.json ADDED
@@ -0,0 +1,3900 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "eval_steps": 500,
6
+ "global_step": 645,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 1.0000000000000002e-06,
14
+ "loss": 0.4037,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 2.0000000000000003e-06,
20
+ "loss": 0.4032,
21
+ "step": 2
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 3e-06,
26
+ "loss": 0.4071,
27
+ "step": 3
28
+ },
29
+ {
30
+ "epoch": 0.02,
31
+ "learning_rate": 4.000000000000001e-06,
32
+ "loss": 0.406,
33
+ "step": 4
34
+ },
35
+ {
36
+ "epoch": 0.02,
37
+ "learning_rate": 5e-06,
38
+ "loss": 0.4054,
39
+ "step": 5
40
+ },
41
+ {
42
+ "epoch": 0.03,
43
+ "learning_rate": 6e-06,
44
+ "loss": 0.4099,
45
+ "step": 6
46
+ },
47
+ {
48
+ "epoch": 0.03,
49
+ "learning_rate": 7e-06,
50
+ "loss": 0.385,
51
+ "step": 7
52
+ },
53
+ {
54
+ "epoch": 0.04,
55
+ "learning_rate": 8.000000000000001e-06,
56
+ "loss": 0.3404,
57
+ "step": 8
58
+ },
59
+ {
60
+ "epoch": 0.04,
61
+ "learning_rate": 9e-06,
62
+ "loss": 0.3211,
63
+ "step": 9
64
+ },
65
+ {
66
+ "epoch": 0.05,
67
+ "learning_rate": 1e-05,
68
+ "loss": 0.2757,
69
+ "step": 10
70
+ },
71
+ {
72
+ "epoch": 0.05,
73
+ "learning_rate": 1.1000000000000001e-05,
74
+ "loss": 0.2633,
75
+ "step": 11
76
+ },
77
+ {
78
+ "epoch": 0.06,
79
+ "learning_rate": 1.2e-05,
80
+ "loss": 0.2397,
81
+ "step": 12
82
+ },
83
+ {
84
+ "epoch": 0.06,
85
+ "learning_rate": 1.3000000000000001e-05,
86
+ "loss": 0.178,
87
+ "step": 13
88
+ },
89
+ {
90
+ "epoch": 0.07,
91
+ "learning_rate": 1.4e-05,
92
+ "loss": 0.1417,
93
+ "step": 14
94
+ },
95
+ {
96
+ "epoch": 0.07,
97
+ "learning_rate": 1.5000000000000002e-05,
98
+ "loss": 0.1309,
99
+ "step": 15
100
+ },
101
+ {
102
+ "epoch": 0.07,
103
+ "learning_rate": 1.6000000000000003e-05,
104
+ "loss": 0.109,
105
+ "step": 16
106
+ },
107
+ {
108
+ "epoch": 0.08,
109
+ "learning_rate": 1.7e-05,
110
+ "loss": 0.1101,
111
+ "step": 17
112
+ },
113
+ {
114
+ "epoch": 0.08,
115
+ "learning_rate": 1.8e-05,
116
+ "loss": 0.0791,
117
+ "step": 18
118
+ },
119
+ {
120
+ "epoch": 0.09,
121
+ "learning_rate": 1.9e-05,
122
+ "loss": 0.0903,
123
+ "step": 19
124
+ },
125
+ {
126
+ "epoch": 0.09,
127
+ "learning_rate": 2e-05,
128
+ "loss": 0.0738,
129
+ "step": 20
130
+ },
131
+ {
132
+ "epoch": 0.1,
133
+ "learning_rate": 1.999987366932966e-05,
134
+ "loss": 0.066,
135
+ "step": 21
136
+ },
137
+ {
138
+ "epoch": 0.1,
139
+ "learning_rate": 1.999949468051052e-05,
140
+ "loss": 0.0583,
141
+ "step": 22
142
+ },
143
+ {
144
+ "epoch": 0.11,
145
+ "learning_rate": 1.9998863043118163e-05,
146
+ "loss": 0.0541,
147
+ "step": 23
148
+ },
149
+ {
150
+ "epoch": 0.11,
151
+ "learning_rate": 1.999797877311163e-05,
152
+ "loss": 0.0733,
153
+ "step": 24
154
+ },
155
+ {
156
+ "epoch": 0.12,
157
+ "learning_rate": 1.9996841892833e-05,
158
+ "loss": 0.0517,
159
+ "step": 25
160
+ },
161
+ {
162
+ "epoch": 0.12,
163
+ "learning_rate": 1.9995452431006844e-05,
164
+ "loss": 0.0579,
165
+ "step": 26
166
+ },
167
+ {
168
+ "epoch": 0.13,
169
+ "learning_rate": 1.9993810422739496e-05,
170
+ "loss": 0.065,
171
+ "step": 27
172
+ },
173
+ {
174
+ "epoch": 0.13,
175
+ "learning_rate": 1.9991915909518146e-05,
176
+ "loss": 0.0462,
177
+ "step": 28
178
+ },
179
+ {
180
+ "epoch": 0.13,
181
+ "learning_rate": 1.9989768939209826e-05,
182
+ "loss": 0.0468,
183
+ "step": 29
184
+ },
185
+ {
186
+ "epoch": 0.14,
187
+ "learning_rate": 1.998736956606018e-05,
188
+ "loss": 0.0531,
189
+ "step": 30
190
+ },
191
+ {
192
+ "epoch": 0.14,
193
+ "learning_rate": 1.998471785069208e-05,
194
+ "loss": 0.053,
195
+ "step": 31
196
+ },
197
+ {
198
+ "epoch": 0.15,
199
+ "learning_rate": 1.998181386010413e-05,
200
+ "loss": 0.052,
201
+ "step": 32
202
+ },
203
+ {
204
+ "epoch": 0.15,
205
+ "learning_rate": 1.9978657667668945e-05,
206
+ "loss": 0.046,
207
+ "step": 33
208
+ },
209
+ {
210
+ "epoch": 0.16,
211
+ "learning_rate": 1.9975249353131304e-05,
212
+ "loss": 0.0524,
213
+ "step": 34
214
+ },
215
+ {
216
+ "epoch": 0.16,
217
+ "learning_rate": 1.997158900260614e-05,
218
+ "loss": 0.055,
219
+ "step": 35
220
+ },
221
+ {
222
+ "epoch": 0.17,
223
+ "learning_rate": 1.9967676708576362e-05,
224
+ "loss": 0.0445,
225
+ "step": 36
226
+ },
227
+ {
228
+ "epoch": 0.17,
229
+ "learning_rate": 1.9963512569890512e-05,
230
+ "loss": 0.0405,
231
+ "step": 37
232
+ },
233
+ {
234
+ "epoch": 0.18,
235
+ "learning_rate": 1.9959096691760284e-05,
236
+ "loss": 0.0434,
237
+ "step": 38
238
+ },
239
+ {
240
+ "epoch": 0.18,
241
+ "learning_rate": 1.9954429185757835e-05,
242
+ "loss": 0.0559,
243
+ "step": 39
244
+ },
245
+ {
246
+ "epoch": 0.19,
247
+ "learning_rate": 1.9949510169813006e-05,
248
+ "loss": 0.0446,
249
+ "step": 40
250
+ },
251
+ {
252
+ "epoch": 0.19,
253
+ "learning_rate": 1.994433976821031e-05,
254
+ "loss": 0.0374,
255
+ "step": 41
256
+ },
257
+ {
258
+ "epoch": 0.2,
259
+ "learning_rate": 1.9938918111585805e-05,
260
+ "loss": 0.037,
261
+ "step": 42
262
+ },
263
+ {
264
+ "epoch": 0.2,
265
+ "learning_rate": 1.9933245336923798e-05,
266
+ "loss": 0.0359,
267
+ "step": 43
268
+ },
269
+ {
270
+ "epoch": 0.2,
271
+ "learning_rate": 1.9927321587553378e-05,
272
+ "loss": 0.0359,
273
+ "step": 44
274
+ },
275
+ {
276
+ "epoch": 0.21,
277
+ "learning_rate": 1.9921147013144782e-05,
278
+ "loss": 0.0359,
279
+ "step": 45
280
+ },
281
+ {
282
+ "epoch": 0.21,
283
+ "learning_rate": 1.9914721769705637e-05,
284
+ "loss": 0.0396,
285
+ "step": 46
286
+ },
287
+ {
288
+ "epoch": 0.22,
289
+ "learning_rate": 1.990804601957701e-05,
290
+ "loss": 0.0348,
291
+ "step": 47
292
+ },
293
+ {
294
+ "epoch": 0.22,
295
+ "learning_rate": 1.9901119931429294e-05,
296
+ "loss": 0.0345,
297
+ "step": 48
298
+ },
299
+ {
300
+ "epoch": 0.23,
301
+ "learning_rate": 1.9893943680257964e-05,
302
+ "loss": 0.0331,
303
+ "step": 49
304
+ },
305
+ {
306
+ "epoch": 0.23,
307
+ "learning_rate": 1.988651744737914e-05,
308
+ "loss": 0.04,
309
+ "step": 50
310
+ },
311
+ {
312
+ "epoch": 0.24,
313
+ "learning_rate": 1.9878841420425023e-05,
314
+ "loss": 0.0335,
315
+ "step": 51
316
+ },
317
+ {
318
+ "epoch": 0.24,
319
+ "learning_rate": 1.9870915793339137e-05,
320
+ "loss": 0.034,
321
+ "step": 52
322
+ },
323
+ {
324
+ "epoch": 0.25,
325
+ "learning_rate": 1.9862740766371434e-05,
326
+ "loss": 0.0345,
327
+ "step": 53
328
+ },
329
+ {
330
+ "epoch": 0.25,
331
+ "learning_rate": 1.985431654607325e-05,
332
+ "loss": 0.0329,
333
+ "step": 54
334
+ },
335
+ {
336
+ "epoch": 0.26,
337
+ "learning_rate": 1.9845643345292055e-05,
338
+ "loss": 0.0334,
339
+ "step": 55
340
+ },
341
+ {
342
+ "epoch": 0.26,
343
+ "learning_rate": 1.983672138316611e-05,
344
+ "loss": 0.037,
345
+ "step": 56
346
+ },
347
+ {
348
+ "epoch": 0.27,
349
+ "learning_rate": 1.9827550885118902e-05,
350
+ "loss": 0.0337,
351
+ "step": 57
352
+ },
353
+ {
354
+ "epoch": 0.27,
355
+ "learning_rate": 1.9818132082853466e-05,
356
+ "loss": 0.0354,
357
+ "step": 58
358
+ },
359
+ {
360
+ "epoch": 0.27,
361
+ "learning_rate": 1.9808465214346525e-05,
362
+ "loss": 0.0333,
363
+ "step": 59
364
+ },
365
+ {
366
+ "epoch": 0.28,
367
+ "learning_rate": 1.979855052384247e-05,
368
+ "loss": 0.0375,
369
+ "step": 60
370
+ },
371
+ {
372
+ "epoch": 0.28,
373
+ "learning_rate": 1.9788388261847204e-05,
374
+ "loss": 0.0417,
375
+ "step": 61
376
+ },
377
+ {
378
+ "epoch": 0.29,
379
+ "learning_rate": 1.97779786851218e-05,
380
+ "loss": 0.0378,
381
+ "step": 62
382
+ },
383
+ {
384
+ "epoch": 0.29,
385
+ "learning_rate": 1.9767322056676018e-05,
386
+ "loss": 0.0371,
387
+ "step": 63
388
+ },
389
+ {
390
+ "epoch": 0.3,
391
+ "learning_rate": 1.975641864576166e-05,
392
+ "loss": 0.0352,
393
+ "step": 64
394
+ },
395
+ {
396
+ "epoch": 0.3,
397
+ "learning_rate": 1.9745268727865774e-05,
398
+ "loss": 0.0326,
399
+ "step": 65
400
+ },
401
+ {
402
+ "epoch": 0.31,
403
+ "learning_rate": 1.9733872584703673e-05,
404
+ "loss": 0.0333,
405
+ "step": 66
406
+ },
407
+ {
408
+ "epoch": 0.31,
409
+ "learning_rate": 1.9722230504211843e-05,
410
+ "loss": 0.0279,
411
+ "step": 67
412
+ },
413
+ {
414
+ "epoch": 0.32,
415
+ "learning_rate": 1.971034278054065e-05,
416
+ "loss": 0.0345,
417
+ "step": 68
418
+ },
419
+ {
420
+ "epoch": 0.32,
421
+ "learning_rate": 1.969820971404691e-05,
422
+ "loss": 0.0321,
423
+ "step": 69
424
+ },
425
+ {
426
+ "epoch": 0.33,
427
+ "learning_rate": 1.9685831611286312e-05,
428
+ "loss": 0.0341,
429
+ "step": 70
430
+ },
431
+ {
432
+ "epoch": 0.33,
433
+ "learning_rate": 1.9673208785005658e-05,
434
+ "loss": 0.0335,
435
+ "step": 71
436
+ },
437
+ {
438
+ "epoch": 0.33,
439
+ "learning_rate": 1.9660341554134972e-05,
440
+ "loss": 0.0372,
441
+ "step": 72
442
+ },
443
+ {
444
+ "epoch": 0.34,
445
+ "learning_rate": 1.9647230243779432e-05,
446
+ "loss": 0.0333,
447
+ "step": 73
448
+ },
449
+ {
450
+ "epoch": 0.34,
451
+ "learning_rate": 1.963387518521116e-05,
452
+ "loss": 0.0422,
453
+ "step": 74
454
+ },
455
+ {
456
+ "epoch": 0.35,
457
+ "learning_rate": 1.962027671586086e-05,
458
+ "loss": 0.0412,
459
+ "step": 75
460
+ },
461
+ {
462
+ "epoch": 0.35,
463
+ "learning_rate": 1.9606435179309284e-05,
464
+ "loss": 0.0314,
465
+ "step": 76
466
+ },
467
+ {
468
+ "epoch": 0.36,
469
+ "learning_rate": 1.9592350925278546e-05,
470
+ "loss": 0.0475,
471
+ "step": 77
472
+ },
473
+ {
474
+ "epoch": 0.36,
475
+ "learning_rate": 1.9578024309623296e-05,
476
+ "loss": 0.0524,
477
+ "step": 78
478
+ },
479
+ {
480
+ "epoch": 0.37,
481
+ "learning_rate": 1.956345569432173e-05,
482
+ "loss": 0.031,
483
+ "step": 79
484
+ },
485
+ {
486
+ "epoch": 0.37,
487
+ "learning_rate": 1.9548645447466433e-05,
488
+ "loss": 0.037,
489
+ "step": 80
490
+ },
491
+ {
492
+ "epoch": 0.38,
493
+ "learning_rate": 1.9533593943255087e-05,
494
+ "loss": 0.0382,
495
+ "step": 81
496
+ },
497
+ {
498
+ "epoch": 0.38,
499
+ "learning_rate": 1.9518301561981016e-05,
500
+ "loss": 0.0323,
501
+ "step": 82
502
+ },
503
+ {
504
+ "epoch": 0.39,
505
+ "learning_rate": 1.9502768690023574e-05,
506
+ "loss": 0.0349,
507
+ "step": 83
508
+ },
509
+ {
510
+ "epoch": 0.39,
511
+ "learning_rate": 1.9486995719838392e-05,
512
+ "loss": 0.0417,
513
+ "step": 84
514
+ },
515
+ {
516
+ "epoch": 0.4,
517
+ "learning_rate": 1.9470983049947446e-05,
518
+ "loss": 0.0389,
519
+ "step": 85
520
+ },
521
+ {
522
+ "epoch": 0.4,
523
+ "learning_rate": 1.9454731084928995e-05,
524
+ "loss": 0.0354,
525
+ "step": 86
526
+ },
527
+ {
528
+ "epoch": 0.4,
529
+ "learning_rate": 1.9438240235407375e-05,
530
+ "loss": 0.0366,
531
+ "step": 87
532
+ },
533
+ {
534
+ "epoch": 0.41,
535
+ "learning_rate": 1.9421510918042593e-05,
536
+ "loss": 0.0388,
537
+ "step": 88
538
+ },
539
+ {
540
+ "epoch": 0.41,
541
+ "learning_rate": 1.940454355551983e-05,
542
+ "loss": 0.0312,
543
+ "step": 89
544
+ },
545
+ {
546
+ "epoch": 0.42,
547
+ "learning_rate": 1.9387338576538743e-05,
548
+ "loss": 0.0354,
549
+ "step": 90
550
+ },
551
+ {
552
+ "epoch": 0.42,
553
+ "learning_rate": 1.936989641580263e-05,
554
+ "loss": 0.0383,
555
+ "step": 91
556
+ },
557
+ {
558
+ "epoch": 0.43,
559
+ "learning_rate": 1.935221751400747e-05,
560
+ "loss": 0.0331,
561
+ "step": 92
562
+ },
563
+ {
564
+ "epoch": 0.43,
565
+ "learning_rate": 1.9334302317830764e-05,
566
+ "loss": 0.0297,
567
+ "step": 93
568
+ },
569
+ {
570
+ "epoch": 0.44,
571
+ "learning_rate": 1.931615127992026e-05,
572
+ "loss": 0.0312,
573
+ "step": 94
574
+ },
575
+ {
576
+ "epoch": 0.44,
577
+ "learning_rate": 1.9297764858882516e-05,
578
+ "loss": 0.0378,
579
+ "step": 95
580
+ },
581
+ {
582
+ "epoch": 0.45,
583
+ "learning_rate": 1.927914351927131e-05,
584
+ "loss": 0.0266,
585
+ "step": 96
586
+ },
587
+ {
588
+ "epoch": 0.45,
589
+ "learning_rate": 1.9260287731575902e-05,
590
+ "loss": 0.0295,
591
+ "step": 97
592
+ },
593
+ {
594
+ "epoch": 0.46,
595
+ "learning_rate": 1.9241197972209157e-05,
596
+ "loss": 0.0371,
597
+ "step": 98
598
+ },
599
+ {
600
+ "epoch": 0.46,
601
+ "learning_rate": 1.9221874723495494e-05,
602
+ "loss": 0.0307,
603
+ "step": 99
604
+ },
605
+ {
606
+ "epoch": 0.47,
607
+ "learning_rate": 1.9202318473658707e-05,
608
+ "loss": 0.0294,
609
+ "step": 100
610
+ },
611
+ {
612
+ "epoch": 0.47,
613
+ "learning_rate": 1.9182529716809618e-05,
614
+ "loss": 0.0311,
615
+ "step": 101
616
+ },
617
+ {
618
+ "epoch": 0.47,
619
+ "learning_rate": 1.916250895293362e-05,
620
+ "loss": 0.025,
621
+ "step": 102
622
+ },
623
+ {
624
+ "epoch": 0.48,
625
+ "learning_rate": 1.9142256687878012e-05,
626
+ "loss": 0.0294,
627
+ "step": 103
628
+ },
629
+ {
630
+ "epoch": 0.48,
631
+ "learning_rate": 1.912177343333924e-05,
632
+ "loss": 0.0336,
633
+ "step": 104
634
+ },
635
+ {
636
+ "epoch": 0.49,
637
+ "learning_rate": 1.9101059706849957e-05,
638
+ "loss": 0.035,
639
+ "step": 105
640
+ },
641
+ {
642
+ "epoch": 0.49,
643
+ "learning_rate": 1.908011603176596e-05,
644
+ "loss": 0.0303,
645
+ "step": 106
646
+ },
647
+ {
648
+ "epoch": 0.5,
649
+ "learning_rate": 1.9058942937252943e-05,
650
+ "loss": 0.0324,
651
+ "step": 107
652
+ },
653
+ {
654
+ "epoch": 0.5,
655
+ "learning_rate": 1.903754095827316e-05,
656
+ "loss": 0.0387,
657
+ "step": 108
658
+ },
659
+ {
660
+ "epoch": 0.51,
661
+ "learning_rate": 1.901591063557187e-05,
662
+ "loss": 0.0323,
663
+ "step": 109
664
+ },
665
+ {
666
+ "epoch": 0.51,
667
+ "learning_rate": 1.899405251566371e-05,
668
+ "loss": 0.0289,
669
+ "step": 110
670
+ },
671
+ {
672
+ "epoch": 0.52,
673
+ "learning_rate": 1.897196715081888e-05,
674
+ "loss": 0.0286,
675
+ "step": 111
676
+ },
677
+ {
678
+ "epoch": 0.52,
679
+ "learning_rate": 1.894965509904915e-05,
680
+ "loss": 0.0318,
681
+ "step": 112
682
+ },
683
+ {
684
+ "epoch": 0.53,
685
+ "learning_rate": 1.8927116924093824e-05,
686
+ "loss": 0.0261,
687
+ "step": 113
688
+ },
689
+ {
690
+ "epoch": 0.53,
691
+ "learning_rate": 1.890435319540545e-05,
692
+ "loss": 0.0307,
693
+ "step": 114
694
+ },
695
+ {
696
+ "epoch": 0.53,
697
+ "learning_rate": 1.8881364488135448e-05,
698
+ "loss": 0.0287,
699
+ "step": 115
700
+ },
701
+ {
702
+ "epoch": 0.54,
703
+ "learning_rate": 1.8858151383119576e-05,
704
+ "loss": 0.0274,
705
+ "step": 116
706
+ },
707
+ {
708
+ "epoch": 0.54,
709
+ "learning_rate": 1.883471446686326e-05,
710
+ "loss": 0.0274,
711
+ "step": 117
712
+ },
713
+ {
714
+ "epoch": 0.55,
715
+ "learning_rate": 1.881105433152677e-05,
716
+ "loss": 0.0286,
717
+ "step": 118
718
+ },
719
+ {
720
+ "epoch": 0.55,
721
+ "learning_rate": 1.878717157491025e-05,
722
+ "loss": 0.0274,
723
+ "step": 119
724
+ },
725
+ {
726
+ "epoch": 0.56,
727
+ "learning_rate": 1.8763066800438638e-05,
728
+ "loss": 0.0214,
729
+ "step": 120
730
+ },
731
+ {
732
+ "epoch": 0.56,
733
+ "learning_rate": 1.8738740617146396e-05,
734
+ "loss": 0.0217,
735
+ "step": 121
736
+ },
737
+ {
738
+ "epoch": 0.57,
739
+ "learning_rate": 1.871419363966213e-05,
740
+ "loss": 0.0335,
741
+ "step": 122
742
+ },
743
+ {
744
+ "epoch": 0.57,
745
+ "learning_rate": 1.8689426488193066e-05,
746
+ "loss": 0.0233,
747
+ "step": 123
748
+ },
749
+ {
750
+ "epoch": 0.58,
751
+ "learning_rate": 1.866443978850937e-05,
752
+ "loss": 0.0325,
753
+ "step": 124
754
+ },
755
+ {
756
+ "epoch": 0.58,
757
+ "learning_rate": 1.8639234171928355e-05,
758
+ "loss": 0.0242,
759
+ "step": 125
760
+ },
761
+ {
762
+ "epoch": 0.59,
763
+ "learning_rate": 1.86138102752985e-05,
764
+ "loss": 0.0296,
765
+ "step": 126
766
+ },
767
+ {
768
+ "epoch": 0.59,
769
+ "learning_rate": 1.858816874098339e-05,
770
+ "loss": 0.0275,
771
+ "step": 127
772
+ },
773
+ {
774
+ "epoch": 0.6,
775
+ "learning_rate": 1.8562310216845463e-05,
776
+ "loss": 0.0294,
777
+ "step": 128
778
+ },
779
+ {
780
+ "epoch": 0.6,
781
+ "learning_rate": 1.8536235356229667e-05,
782
+ "loss": 0.0295,
783
+ "step": 129
784
+ },
785
+ {
786
+ "epoch": 0.6,
787
+ "learning_rate": 1.850994481794692e-05,
788
+ "loss": 0.0259,
789
+ "step": 130
790
+ },
791
+ {
792
+ "epoch": 0.61,
793
+ "learning_rate": 1.8483439266257485e-05,
794
+ "loss": 0.0308,
795
+ "step": 131
796
+ },
797
+ {
798
+ "epoch": 0.61,
799
+ "learning_rate": 1.845671937085419e-05,
800
+ "loss": 0.0299,
801
+ "step": 132
802
+ },
803
+ {
804
+ "epoch": 0.62,
805
+ "learning_rate": 1.842978580684549e-05,
806
+ "loss": 0.024,
807
+ "step": 133
808
+ },
809
+ {
810
+ "epoch": 0.62,
811
+ "learning_rate": 1.8402639254738422e-05,
812
+ "loss": 0.0266,
813
+ "step": 134
814
+ },
815
+ {
816
+ "epoch": 0.63,
817
+ "learning_rate": 1.837528040042142e-05,
818
+ "loss": 0.0241,
819
+ "step": 135
820
+ },
821
+ {
822
+ "epoch": 0.63,
823
+ "learning_rate": 1.8347709935146958e-05,
824
+ "loss": 0.0273,
825
+ "step": 136
826
+ },
827
+ {
828
+ "epoch": 0.64,
829
+ "learning_rate": 1.8319928555514108e-05,
830
+ "loss": 0.028,
831
+ "step": 137
832
+ },
833
+ {
834
+ "epoch": 0.64,
835
+ "learning_rate": 1.8291936963450933e-05,
836
+ "loss": 0.0277,
837
+ "step": 138
838
+ },
839
+ {
840
+ "epoch": 0.65,
841
+ "learning_rate": 1.8263735866196758e-05,
842
+ "loss": 0.0314,
843
+ "step": 139
844
+ },
845
+ {
846
+ "epoch": 0.65,
847
+ "learning_rate": 1.8235325976284276e-05,
848
+ "loss": 0.0273,
849
+ "step": 140
850
+ },
851
+ {
852
+ "epoch": 0.66,
853
+ "learning_rate": 1.820670801152158e-05,
854
+ "loss": 0.0242,
855
+ "step": 141
856
+ },
857
+ {
858
+ "epoch": 0.66,
859
+ "learning_rate": 1.8177882694974008e-05,
860
+ "loss": 0.0265,
861
+ "step": 142
862
+ },
863
+ {
864
+ "epoch": 0.67,
865
+ "learning_rate": 1.8148850754945865e-05,
866
+ "loss": 0.0282,
867
+ "step": 143
868
+ },
869
+ {
870
+ "epoch": 0.67,
871
+ "learning_rate": 1.8119612924962043e-05,
872
+ "loss": 0.0244,
873
+ "step": 144
874
+ },
875
+ {
876
+ "epoch": 0.67,
877
+ "learning_rate": 1.8090169943749477e-05,
878
+ "loss": 0.0249,
879
+ "step": 145
880
+ },
881
+ {
882
+ "epoch": 0.68,
883
+ "learning_rate": 1.806052255521847e-05,
884
+ "loss": 0.0246,
885
+ "step": 146
886
+ },
887
+ {
888
+ "epoch": 0.68,
889
+ "learning_rate": 1.8030671508443928e-05,
890
+ "loss": 0.028,
891
+ "step": 147
892
+ },
893
+ {
894
+ "epoch": 0.69,
895
+ "learning_rate": 1.8000617557646392e-05,
896
+ "loss": 0.0225,
897
+ "step": 148
898
+ },
899
+ {
900
+ "epoch": 0.69,
901
+ "learning_rate": 1.797036146217301e-05,
902
+ "loss": 0.0257,
903
+ "step": 149
904
+ },
905
+ {
906
+ "epoch": 0.7,
907
+ "learning_rate": 1.7939903986478354e-05,
908
+ "loss": 0.0283,
909
+ "step": 150
910
+ },
911
+ {
912
+ "epoch": 0.7,
913
+ "learning_rate": 1.7909245900105085e-05,
914
+ "loss": 0.0264,
915
+ "step": 151
916
+ },
917
+ {
918
+ "epoch": 0.71,
919
+ "learning_rate": 1.7878387977664522e-05,
920
+ "loss": 0.0277,
921
+ "step": 152
922
+ },
923
+ {
924
+ "epoch": 0.71,
925
+ "learning_rate": 1.784733099881707e-05,
926
+ "loss": 0.0244,
927
+ "step": 153
928
+ },
929
+ {
930
+ "epoch": 0.72,
931
+ "learning_rate": 1.7816075748252526e-05,
932
+ "loss": 0.0246,
933
+ "step": 154
934
+ },
935
+ {
936
+ "epoch": 0.72,
937
+ "learning_rate": 1.7784623015670237e-05,
938
+ "loss": 0.0306,
939
+ "step": 155
940
+ },
941
+ {
942
+ "epoch": 0.73,
943
+ "learning_rate": 1.775297359575916e-05,
944
+ "loss": 0.0234,
945
+ "step": 156
946
+ },
947
+ {
948
+ "epoch": 0.73,
949
+ "learning_rate": 1.7721128288177782e-05,
950
+ "loss": 0.0267,
951
+ "step": 157
952
+ },
953
+ {
954
+ "epoch": 0.73,
955
+ "learning_rate": 1.7689087897533916e-05,
956
+ "loss": 0.0272,
957
+ "step": 158
958
+ },
959
+ {
960
+ "epoch": 0.74,
961
+ "learning_rate": 1.765685323336437e-05,
962
+ "loss": 0.0271,
963
+ "step": 159
964
+ },
965
+ {
966
+ "epoch": 0.74,
967
+ "learning_rate": 1.762442511011448e-05,
968
+ "loss": 0.0269,
969
+ "step": 160
970
+ },
971
+ {
972
+ "epoch": 0.75,
973
+ "learning_rate": 1.759180434711757e-05,
974
+ "loss": 0.0258,
975
+ "step": 161
976
+ },
977
+ {
978
+ "epoch": 0.75,
979
+ "learning_rate": 1.7558991768574197e-05,
980
+ "loss": 0.0254,
981
+ "step": 162
982
+ },
983
+ {
984
+ "epoch": 0.76,
985
+ "learning_rate": 1.752598820353138e-05,
986
+ "loss": 0.029,
987
+ "step": 163
988
+ },
989
+ {
990
+ "epoch": 0.76,
991
+ "learning_rate": 1.749279448586162e-05,
992
+ "loss": 0.0262,
993
+ "step": 164
994
+ },
995
+ {
996
+ "epoch": 0.77,
997
+ "learning_rate": 1.7459411454241822e-05,
998
+ "loss": 0.0328,
999
+ "step": 165
1000
+ },
1001
+ {
1002
+ "epoch": 0.77,
1003
+ "learning_rate": 1.7425839952132157e-05,
1004
+ "loss": 0.0284,
1005
+ "step": 166
1006
+ },
1007
+ {
1008
+ "epoch": 0.78,
1009
+ "learning_rate": 1.739208082775469e-05,
1010
+ "loss": 0.0306,
1011
+ "step": 167
1012
+ },
1013
+ {
1014
+ "epoch": 0.78,
1015
+ "learning_rate": 1.7358134934071978e-05,
1016
+ "loss": 0.0297,
1017
+ "step": 168
1018
+ },
1019
+ {
1020
+ "epoch": 0.79,
1021
+ "learning_rate": 1.7324003128765536e-05,
1022
+ "loss": 0.0253,
1023
+ "step": 169
1024
+ },
1025
+ {
1026
+ "epoch": 0.79,
1027
+ "learning_rate": 1.7289686274214116e-05,
1028
+ "loss": 0.0273,
1029
+ "step": 170
1030
+ },
1031
+ {
1032
+ "epoch": 0.8,
1033
+ "learning_rate": 1.7255185237471978e-05,
1034
+ "loss": 0.0268,
1035
+ "step": 171
1036
+ },
1037
+ {
1038
+ "epoch": 0.8,
1039
+ "learning_rate": 1.7220500890246944e-05,
1040
+ "loss": 0.0313,
1041
+ "step": 172
1042
+ },
1043
+ {
1044
+ "epoch": 0.8,
1045
+ "learning_rate": 1.7185634108878367e-05,
1046
+ "loss": 0.0319,
1047
+ "step": 173
1048
+ },
1049
+ {
1050
+ "epoch": 0.81,
1051
+ "learning_rate": 1.715058577431503e-05,
1052
+ "loss": 0.0274,
1053
+ "step": 174
1054
+ },
1055
+ {
1056
+ "epoch": 0.81,
1057
+ "learning_rate": 1.7115356772092858e-05,
1058
+ "loss": 0.0268,
1059
+ "step": 175
1060
+ },
1061
+ {
1062
+ "epoch": 0.82,
1063
+ "learning_rate": 1.707994799231253e-05,
1064
+ "loss": 0.0255,
1065
+ "step": 176
1066
+ },
1067
+ {
1068
+ "epoch": 0.82,
1069
+ "learning_rate": 1.704436032961703e-05,
1070
+ "loss": 0.0273,
1071
+ "step": 177
1072
+ },
1073
+ {
1074
+ "epoch": 0.83,
1075
+ "learning_rate": 1.7008594683169018e-05,
1076
+ "loss": 0.0261,
1077
+ "step": 178
1078
+ },
1079
+ {
1080
+ "epoch": 0.83,
1081
+ "learning_rate": 1.6972651956628108e-05,
1082
+ "loss": 0.0269,
1083
+ "step": 179
1084
+ },
1085
+ {
1086
+ "epoch": 0.84,
1087
+ "learning_rate": 1.693653305812805e-05,
1088
+ "loss": 0.0259,
1089
+ "step": 180
1090
+ },
1091
+ {
1092
+ "epoch": 0.84,
1093
+ "learning_rate": 1.6900238900253777e-05,
1094
+ "loss": 0.0248,
1095
+ "step": 181
1096
+ },
1097
+ {
1098
+ "epoch": 0.85,
1099
+ "learning_rate": 1.6863770400018344e-05,
1100
+ "loss": 0.0256,
1101
+ "step": 182
1102
+ },
1103
+ {
1104
+ "epoch": 0.85,
1105
+ "learning_rate": 1.6827128478839767e-05,
1106
+ "loss": 0.0242,
1107
+ "step": 183
1108
+ },
1109
+ {
1110
+ "epoch": 0.86,
1111
+ "learning_rate": 1.679031406251774e-05,
1112
+ "loss": 0.0262,
1113
+ "step": 184
1114
+ },
1115
+ {
1116
+ "epoch": 0.86,
1117
+ "learning_rate": 1.6753328081210244e-05,
1118
+ "loss": 0.0285,
1119
+ "step": 185
1120
+ },
1121
+ {
1122
+ "epoch": 0.87,
1123
+ "learning_rate": 1.6716171469410042e-05,
1124
+ "loss": 0.0324,
1125
+ "step": 186
1126
+ },
1127
+ {
1128
+ "epoch": 0.87,
1129
+ "learning_rate": 1.6678845165921066e-05,
1130
+ "loss": 0.0236,
1131
+ "step": 187
1132
+ },
1133
+ {
1134
+ "epoch": 0.87,
1135
+ "learning_rate": 1.6641350113834705e-05,
1136
+ "loss": 0.0247,
1137
+ "step": 188
1138
+ },
1139
+ {
1140
+ "epoch": 0.88,
1141
+ "learning_rate": 1.660368726050597e-05,
1142
+ "loss": 0.0247,
1143
+ "step": 189
1144
+ },
1145
+ {
1146
+ "epoch": 0.88,
1147
+ "learning_rate": 1.6565857557529567e-05,
1148
+ "loss": 0.0273,
1149
+ "step": 190
1150
+ },
1151
+ {
1152
+ "epoch": 0.89,
1153
+ "learning_rate": 1.652786196071584e-05,
1154
+ "loss": 0.0275,
1155
+ "step": 191
1156
+ },
1157
+ {
1158
+ "epoch": 0.89,
1159
+ "learning_rate": 1.6489701430066632e-05,
1160
+ "loss": 0.0244,
1161
+ "step": 192
1162
+ },
1163
+ {
1164
+ "epoch": 0.9,
1165
+ "learning_rate": 1.6451376929751028e-05,
1166
+ "loss": 0.0259,
1167
+ "step": 193
1168
+ },
1169
+ {
1170
+ "epoch": 0.9,
1171
+ "learning_rate": 1.6412889428080992e-05,
1172
+ "loss": 0.0281,
1173
+ "step": 194
1174
+ },
1175
+ {
1176
+ "epoch": 0.91,
1177
+ "learning_rate": 1.63742398974869e-05,
1178
+ "loss": 0.0255,
1179
+ "step": 195
1180
+ },
1181
+ {
1182
+ "epoch": 0.91,
1183
+ "learning_rate": 1.633542931449297e-05,
1184
+ "loss": 0.0259,
1185
+ "step": 196
1186
+ },
1187
+ {
1188
+ "epoch": 0.92,
1189
+ "learning_rate": 1.62964586596926e-05,
1190
+ "loss": 0.034,
1191
+ "step": 197
1192
+ },
1193
+ {
1194
+ "epoch": 0.92,
1195
+ "learning_rate": 1.625732891772358e-05,
1196
+ "loss": 0.0243,
1197
+ "step": 198
1198
+ },
1199
+ {
1200
+ "epoch": 0.93,
1201
+ "learning_rate": 1.6218041077243213e-05,
1202
+ "loss": 0.0229,
1203
+ "step": 199
1204
+ },
1205
+ {
1206
+ "epoch": 0.93,
1207
+ "learning_rate": 1.6178596130903345e-05,
1208
+ "loss": 0.024,
1209
+ "step": 200
1210
+ },
1211
+ {
1212
+ "epoch": 0.93,
1213
+ "learning_rate": 1.6138995075325277e-05,
1214
+ "loss": 0.0233,
1215
+ "step": 201
1216
+ },
1217
+ {
1218
+ "epoch": 0.94,
1219
+ "learning_rate": 1.609923891107459e-05,
1220
+ "loss": 0.0277,
1221
+ "step": 202
1222
+ },
1223
+ {
1224
+ "epoch": 0.94,
1225
+ "learning_rate": 1.6059328642635864e-05,
1226
+ "loss": 0.0315,
1227
+ "step": 203
1228
+ },
1229
+ {
1230
+ "epoch": 0.95,
1231
+ "learning_rate": 1.6019265278387287e-05,
1232
+ "loss": 0.027,
1233
+ "step": 204
1234
+ },
1235
+ {
1236
+ "epoch": 0.95,
1237
+ "learning_rate": 1.597904983057519e-05,
1238
+ "loss": 0.0285,
1239
+ "step": 205
1240
+ },
1241
+ {
1242
+ "epoch": 0.96,
1243
+ "learning_rate": 1.5938683315288472e-05,
1244
+ "loss": 0.0237,
1245
+ "step": 206
1246
+ },
1247
+ {
1248
+ "epoch": 0.96,
1249
+ "learning_rate": 1.589816675243292e-05,
1250
+ "loss": 0.0254,
1251
+ "step": 207
1252
+ },
1253
+ {
1254
+ "epoch": 0.97,
1255
+ "learning_rate": 1.5857501165705443e-05,
1256
+ "loss": 0.0281,
1257
+ "step": 208
1258
+ },
1259
+ {
1260
+ "epoch": 0.97,
1261
+ "learning_rate": 1.581668758256821e-05,
1262
+ "loss": 0.03,
1263
+ "step": 209
1264
+ },
1265
+ {
1266
+ "epoch": 0.98,
1267
+ "learning_rate": 1.5775727034222675e-05,
1268
+ "loss": 0.0257,
1269
+ "step": 210
1270
+ },
1271
+ {
1272
+ "epoch": 0.98,
1273
+ "learning_rate": 1.5734620555583555e-05,
1274
+ "loss": 0.0262,
1275
+ "step": 211
1276
+ },
1277
+ {
1278
+ "epoch": 0.99,
1279
+ "learning_rate": 1.5693369185252648e-05,
1280
+ "loss": 0.0238,
1281
+ "step": 212
1282
+ },
1283
+ {
1284
+ "epoch": 0.99,
1285
+ "learning_rate": 1.56519739654926e-05,
1286
+ "loss": 0.0275,
1287
+ "step": 213
1288
+ },
1289
+ {
1290
+ "epoch": 1.0,
1291
+ "learning_rate": 1.561043594220059e-05,
1292
+ "loss": 0.0221,
1293
+ "step": 214
1294
+ },
1295
+ {
1296
+ "epoch": 1.0,
1297
+ "learning_rate": 1.556875616488188e-05,
1298
+ "loss": 0.0291,
1299
+ "step": 215
1300
+ },
1301
+ {
1302
+ "epoch": 1.0,
1303
+ "learning_rate": 1.5526935686623316e-05,
1304
+ "loss": 0.0223,
1305
+ "step": 216
1306
+ },
1307
+ {
1308
+ "epoch": 1.01,
1309
+ "learning_rate": 1.5484975564066704e-05,
1310
+ "loss": 0.0208,
1311
+ "step": 217
1312
+ },
1313
+ {
1314
+ "epoch": 1.01,
1315
+ "learning_rate": 1.544287685738213e-05,
1316
+ "loss": 0.0292,
1317
+ "step": 218
1318
+ },
1319
+ {
1320
+ "epoch": 1.02,
1321
+ "learning_rate": 1.540064063024116e-05,
1322
+ "loss": 0.023,
1323
+ "step": 219
1324
+ },
1325
+ {
1326
+ "epoch": 1.02,
1327
+ "learning_rate": 1.5358267949789968e-05,
1328
+ "loss": 0.0252,
1329
+ "step": 220
1330
+ },
1331
+ {
1332
+ "epoch": 1.03,
1333
+ "learning_rate": 1.531575988662238e-05,
1334
+ "loss": 0.0253,
1335
+ "step": 221
1336
+ },
1337
+ {
1338
+ "epoch": 1.03,
1339
+ "learning_rate": 1.5273117514752826e-05,
1340
+ "loss": 0.029,
1341
+ "step": 222
1342
+ },
1343
+ {
1344
+ "epoch": 1.04,
1345
+ "learning_rate": 1.5230341911589183e-05,
1346
+ "loss": 0.02,
1347
+ "step": 223
1348
+ },
1349
+ {
1350
+ "epoch": 1.04,
1351
+ "learning_rate": 1.5187434157905575e-05,
1352
+ "loss": 0.0212,
1353
+ "step": 224
1354
+ },
1355
+ {
1356
+ "epoch": 1.05,
1357
+ "learning_rate": 1.5144395337815066e-05,
1358
+ "loss": 0.0204,
1359
+ "step": 225
1360
+ },
1361
+ {
1362
+ "epoch": 1.05,
1363
+ "learning_rate": 1.5101226538742248e-05,
1364
+ "loss": 0.0211,
1365
+ "step": 226
1366
+ },
1367
+ {
1368
+ "epoch": 1.06,
1369
+ "learning_rate": 1.505792885139579e-05,
1370
+ "loss": 0.0252,
1371
+ "step": 227
1372
+ },
1373
+ {
1374
+ "epoch": 1.06,
1375
+ "learning_rate": 1.5014503369740866e-05,
1376
+ "loss": 0.0241,
1377
+ "step": 228
1378
+ },
1379
+ {
1380
+ "epoch": 1.07,
1381
+ "learning_rate": 1.4970951190971512e-05,
1382
+ "loss": 0.0271,
1383
+ "step": 229
1384
+ },
1385
+ {
1386
+ "epoch": 1.07,
1387
+ "learning_rate": 1.4927273415482916e-05,
1388
+ "loss": 0.0291,
1389
+ "step": 230
1390
+ },
1391
+ {
1392
+ "epoch": 1.07,
1393
+ "learning_rate": 1.4883471146843617e-05,
1394
+ "loss": 0.0285,
1395
+ "step": 231
1396
+ },
1397
+ {
1398
+ "epoch": 1.08,
1399
+ "learning_rate": 1.4839545491767599e-05,
1400
+ "loss": 0.0256,
1401
+ "step": 232
1402
+ },
1403
+ {
1404
+ "epoch": 1.08,
1405
+ "learning_rate": 1.4795497560086358e-05,
1406
+ "loss": 0.0265,
1407
+ "step": 233
1408
+ },
1409
+ {
1410
+ "epoch": 1.09,
1411
+ "learning_rate": 1.4751328464720842e-05,
1412
+ "loss": 0.0229,
1413
+ "step": 234
1414
+ },
1415
+ {
1416
+ "epoch": 1.09,
1417
+ "learning_rate": 1.470703932165333e-05,
1418
+ "loss": 0.0257,
1419
+ "step": 235
1420
+ },
1421
+ {
1422
+ "epoch": 1.1,
1423
+ "learning_rate": 1.4662631249899248e-05,
1424
+ "loss": 0.0257,
1425
+ "step": 236
1426
+ },
1427
+ {
1428
+ "epoch": 1.1,
1429
+ "learning_rate": 1.4618105371478896e-05,
1430
+ "loss": 0.0235,
1431
+ "step": 237
1432
+ },
1433
+ {
1434
+ "epoch": 1.11,
1435
+ "learning_rate": 1.4573462811389087e-05,
1436
+ "loss": 0.0199,
1437
+ "step": 238
1438
+ },
1439
+ {
1440
+ "epoch": 1.11,
1441
+ "learning_rate": 1.4528704697574729e-05,
1442
+ "loss": 0.0201,
1443
+ "step": 239
1444
+ },
1445
+ {
1446
+ "epoch": 1.12,
1447
+ "learning_rate": 1.4483832160900326e-05,
1448
+ "loss": 0.0229,
1449
+ "step": 240
1450
+ },
1451
+ {
1452
+ "epoch": 1.12,
1453
+ "learning_rate": 1.4438846335121402e-05,
1454
+ "loss": 0.0275,
1455
+ "step": 241
1456
+ },
1457
+ {
1458
+ "epoch": 1.13,
1459
+ "learning_rate": 1.4393748356855865e-05,
1460
+ "loss": 0.0195,
1461
+ "step": 242
1462
+ },
1463
+ {
1464
+ "epoch": 1.13,
1465
+ "learning_rate": 1.4348539365555283e-05,
1466
+ "loss": 0.0273,
1467
+ "step": 243
1468
+ },
1469
+ {
1470
+ "epoch": 1.13,
1471
+ "learning_rate": 1.430322050347609e-05,
1472
+ "loss": 0.0304,
1473
+ "step": 244
1474
+ },
1475
+ {
1476
+ "epoch": 1.14,
1477
+ "learning_rate": 1.4257792915650728e-05,
1478
+ "loss": 0.0228,
1479
+ "step": 245
1480
+ },
1481
+ {
1482
+ "epoch": 1.14,
1483
+ "learning_rate": 1.4212257749858727e-05,
1484
+ "loss": 0.022,
1485
+ "step": 246
1486
+ },
1487
+ {
1488
+ "epoch": 1.15,
1489
+ "learning_rate": 1.416661615659768e-05,
1490
+ "loss": 0.0266,
1491
+ "step": 247
1492
+ },
1493
+ {
1494
+ "epoch": 1.15,
1495
+ "learning_rate": 1.412086928905421e-05,
1496
+ "loss": 0.0254,
1497
+ "step": 248
1498
+ },
1499
+ {
1500
+ "epoch": 1.16,
1501
+ "learning_rate": 1.4075018303074808e-05,
1502
+ "loss": 0.0229,
1503
+ "step": 249
1504
+ },
1505
+ {
1506
+ "epoch": 1.16,
1507
+ "learning_rate": 1.4029064357136628e-05,
1508
+ "loss": 0.0246,
1509
+ "step": 250
1510
+ },
1511
+ {
1512
+ "epoch": 1.17,
1513
+ "learning_rate": 1.398300861231823e-05,
1514
+ "loss": 0.0221,
1515
+ "step": 251
1516
+ },
1517
+ {
1518
+ "epoch": 1.17,
1519
+ "learning_rate": 1.3936852232270236e-05,
1520
+ "loss": 0.0213,
1521
+ "step": 252
1522
+ },
1523
+ {
1524
+ "epoch": 1.18,
1525
+ "learning_rate": 1.3890596383185934e-05,
1526
+ "loss": 0.0266,
1527
+ "step": 253
1528
+ },
1529
+ {
1530
+ "epoch": 1.18,
1531
+ "learning_rate": 1.384424223377181e-05,
1532
+ "loss": 0.0241,
1533
+ "step": 254
1534
+ },
1535
+ {
1536
+ "epoch": 1.19,
1537
+ "learning_rate": 1.3797790955218014e-05,
1538
+ "loss": 0.0231,
1539
+ "step": 255
1540
+ },
1541
+ {
1542
+ "epoch": 1.19,
1543
+ "learning_rate": 1.3751243721168778e-05,
1544
+ "loss": 0.0272,
1545
+ "step": 256
1546
+ },
1547
+ {
1548
+ "epoch": 1.2,
1549
+ "learning_rate": 1.3704601707692762e-05,
1550
+ "loss": 0.023,
1551
+ "step": 257
1552
+ },
1553
+ {
1554
+ "epoch": 1.2,
1555
+ "learning_rate": 1.3657866093253327e-05,
1556
+ "loss": 0.0194,
1557
+ "step": 258
1558
+ },
1559
+ {
1560
+ "epoch": 1.2,
1561
+ "learning_rate": 1.3611038058678776e-05,
1562
+ "loss": 0.0205,
1563
+ "step": 259
1564
+ },
1565
+ {
1566
+ "epoch": 1.21,
1567
+ "learning_rate": 1.3564118787132507e-05,
1568
+ "loss": 0.0233,
1569
+ "step": 260
1570
+ },
1571
+ {
1572
+ "epoch": 1.21,
1573
+ "learning_rate": 1.3517109464083129e-05,
1574
+ "loss": 0.0238,
1575
+ "step": 261
1576
+ },
1577
+ {
1578
+ "epoch": 1.22,
1579
+ "learning_rate": 1.3470011277274497e-05,
1580
+ "loss": 0.017,
1581
+ "step": 262
1582
+ },
1583
+ {
1584
+ "epoch": 1.22,
1585
+ "learning_rate": 1.3422825416695713e-05,
1586
+ "loss": 0.0233,
1587
+ "step": 263
1588
+ },
1589
+ {
1590
+ "epoch": 1.23,
1591
+ "learning_rate": 1.337555307455106e-05,
1592
+ "loss": 0.0201,
1593
+ "step": 264
1594
+ },
1595
+ {
1596
+ "epoch": 1.23,
1597
+ "learning_rate": 1.3328195445229869e-05,
1598
+ "loss": 0.0242,
1599
+ "step": 265
1600
+ },
1601
+ {
1602
+ "epoch": 1.24,
1603
+ "learning_rate": 1.3280753725276352e-05,
1604
+ "loss": 0.0255,
1605
+ "step": 266
1606
+ },
1607
+ {
1608
+ "epoch": 1.24,
1609
+ "learning_rate": 1.3233229113359368e-05,
1610
+ "loss": 0.0239,
1611
+ "step": 267
1612
+ },
1613
+ {
1614
+ "epoch": 1.25,
1615
+ "learning_rate": 1.3185622810242129e-05,
1616
+ "loss": 0.0215,
1617
+ "step": 268
1618
+ },
1619
+ {
1620
+ "epoch": 1.25,
1621
+ "learning_rate": 1.3137936018751876e-05,
1622
+ "loss": 0.0225,
1623
+ "step": 269
1624
+ },
1625
+ {
1626
+ "epoch": 1.26,
1627
+ "learning_rate": 1.3090169943749475e-05,
1628
+ "loss": 0.0212,
1629
+ "step": 270
1630
+ },
1631
+ {
1632
+ "epoch": 1.26,
1633
+ "learning_rate": 1.3042325792098982e-05,
1634
+ "loss": 0.027,
1635
+ "step": 271
1636
+ },
1637
+ {
1638
+ "epoch": 1.27,
1639
+ "learning_rate": 1.2994404772637145e-05,
1640
+ "loss": 0.0278,
1641
+ "step": 272
1642
+ },
1643
+ {
1644
+ "epoch": 1.27,
1645
+ "learning_rate": 1.2946408096142866e-05,
1646
+ "loss": 0.0243,
1647
+ "step": 273
1648
+ },
1649
+ {
1650
+ "epoch": 1.27,
1651
+ "learning_rate": 1.289833697530661e-05,
1652
+ "loss": 0.0221,
1653
+ "step": 274
1654
+ },
1655
+ {
1656
+ "epoch": 1.28,
1657
+ "learning_rate": 1.2850192624699762e-05,
1658
+ "loss": 0.0212,
1659
+ "step": 275
1660
+ },
1661
+ {
1662
+ "epoch": 1.28,
1663
+ "learning_rate": 1.2801976260743937e-05,
1664
+ "loss": 0.0241,
1665
+ "step": 276
1666
+ },
1667
+ {
1668
+ "epoch": 1.29,
1669
+ "learning_rate": 1.2753689101680252e-05,
1670
+ "loss": 0.0207,
1671
+ "step": 277
1672
+ },
1673
+ {
1674
+ "epoch": 1.29,
1675
+ "learning_rate": 1.2705332367538539e-05,
1676
+ "loss": 0.0191,
1677
+ "step": 278
1678
+ },
1679
+ {
1680
+ "epoch": 1.3,
1681
+ "learning_rate": 1.2656907280106528e-05,
1682
+ "loss": 0.0193,
1683
+ "step": 279
1684
+ },
1685
+ {
1686
+ "epoch": 1.3,
1687
+ "learning_rate": 1.2608415062898971e-05,
1688
+ "loss": 0.0237,
1689
+ "step": 280
1690
+ },
1691
+ {
1692
+ "epoch": 1.31,
1693
+ "learning_rate": 1.255985694112673e-05,
1694
+ "loss": 0.0272,
1695
+ "step": 281
1696
+ },
1697
+ {
1698
+ "epoch": 1.31,
1699
+ "learning_rate": 1.2511234141665816e-05,
1700
+ "loss": 0.0201,
1701
+ "step": 282
1702
+ },
1703
+ {
1704
+ "epoch": 1.32,
1705
+ "learning_rate": 1.2462547893026403e-05,
1706
+ "loss": 0.0216,
1707
+ "step": 283
1708
+ },
1709
+ {
1710
+ "epoch": 1.32,
1711
+ "learning_rate": 1.2413799425321774e-05,
1712
+ "loss": 0.0225,
1713
+ "step": 284
1714
+ },
1715
+ {
1716
+ "epoch": 1.33,
1717
+ "learning_rate": 1.236498997023725e-05,
1718
+ "loss": 0.0259,
1719
+ "step": 285
1720
+ },
1721
+ {
1722
+ "epoch": 1.33,
1723
+ "learning_rate": 1.2316120760999066e-05,
1724
+ "loss": 0.0283,
1725
+ "step": 286
1726
+ },
1727
+ {
1728
+ "epoch": 1.33,
1729
+ "learning_rate": 1.2267193032343219e-05,
1730
+ "loss": 0.0201,
1731
+ "step": 287
1732
+ },
1733
+ {
1734
+ "epoch": 1.34,
1735
+ "learning_rate": 1.2218208020484255e-05,
1736
+ "loss": 0.0199,
1737
+ "step": 288
1738
+ },
1739
+ {
1740
+ "epoch": 1.34,
1741
+ "learning_rate": 1.2169166963084056e-05,
1742
+ "loss": 0.0221,
1743
+ "step": 289
1744
+ },
1745
+ {
1746
+ "epoch": 1.35,
1747
+ "learning_rate": 1.212007109922055e-05,
1748
+ "loss": 0.0274,
1749
+ "step": 290
1750
+ },
1751
+ {
1752
+ "epoch": 1.35,
1753
+ "learning_rate": 1.2070921669356415e-05,
1754
+ "loss": 0.0299,
1755
+ "step": 291
1756
+ },
1757
+ {
1758
+ "epoch": 1.36,
1759
+ "learning_rate": 1.2021719915307737e-05,
1760
+ "loss": 0.0263,
1761
+ "step": 292
1762
+ },
1763
+ {
1764
+ "epoch": 1.36,
1765
+ "learning_rate": 1.197246708021263e-05,
1766
+ "loss": 0.023,
1767
+ "step": 293
1768
+ },
1769
+ {
1770
+ "epoch": 1.37,
1771
+ "learning_rate": 1.192316440849983e-05,
1772
+ "loss": 0.0228,
1773
+ "step": 294
1774
+ },
1775
+ {
1776
+ "epoch": 1.37,
1777
+ "learning_rate": 1.187381314585725e-05,
1778
+ "loss": 0.0232,
1779
+ "step": 295
1780
+ },
1781
+ {
1782
+ "epoch": 1.38,
1783
+ "learning_rate": 1.1824414539200505e-05,
1784
+ "loss": 0.0179,
1785
+ "step": 296
1786
+ },
1787
+ {
1788
+ "epoch": 1.38,
1789
+ "learning_rate": 1.1774969836641417e-05,
1790
+ "loss": 0.0189,
1791
+ "step": 297
1792
+ },
1793
+ {
1794
+ "epoch": 1.39,
1795
+ "learning_rate": 1.1725480287456467e-05,
1796
+ "loss": 0.0222,
1797
+ "step": 298
1798
+ },
1799
+ {
1800
+ "epoch": 1.39,
1801
+ "learning_rate": 1.1675947142055241e-05,
1802
+ "loss": 0.0204,
1803
+ "step": 299
1804
+ },
1805
+ {
1806
+ "epoch": 1.4,
1807
+ "learning_rate": 1.1626371651948839e-05,
1808
+ "loss": 0.0228,
1809
+ "step": 300
1810
+ },
1811
+ {
1812
+ "epoch": 1.4,
1813
+ "learning_rate": 1.1576755069718229e-05,
1814
+ "loss": 0.0253,
1815
+ "step": 301
1816
+ },
1817
+ {
1818
+ "epoch": 1.4,
1819
+ "learning_rate": 1.1527098648982634e-05,
1820
+ "loss": 0.0224,
1821
+ "step": 302
1822
+ },
1823
+ {
1824
+ "epoch": 1.41,
1825
+ "learning_rate": 1.1477403644367839e-05,
1826
+ "loss": 0.0199,
1827
+ "step": 303
1828
+ },
1829
+ {
1830
+ "epoch": 1.41,
1831
+ "learning_rate": 1.1427671311474489e-05,
1832
+ "loss": 0.0217,
1833
+ "step": 304
1834
+ },
1835
+ {
1836
+ "epoch": 1.42,
1837
+ "learning_rate": 1.137790290684638e-05,
1838
+ "loss": 0.0226,
1839
+ "step": 305
1840
+ },
1841
+ {
1842
+ "epoch": 1.42,
1843
+ "learning_rate": 1.1328099687938696e-05,
1844
+ "loss": 0.0212,
1845
+ "step": 306
1846
+ },
1847
+ {
1848
+ "epoch": 1.43,
1849
+ "learning_rate": 1.1278262913086238e-05,
1850
+ "loss": 0.0224,
1851
+ "step": 307
1852
+ },
1853
+ {
1854
+ "epoch": 1.43,
1855
+ "learning_rate": 1.1228393841471644e-05,
1856
+ "loss": 0.021,
1857
+ "step": 308
1858
+ },
1859
+ {
1860
+ "epoch": 1.44,
1861
+ "learning_rate": 1.117849373309356e-05,
1862
+ "loss": 0.0266,
1863
+ "step": 309
1864
+ },
1865
+ {
1866
+ "epoch": 1.44,
1867
+ "learning_rate": 1.1128563848734817e-05,
1868
+ "loss": 0.0221,
1869
+ "step": 310
1870
+ },
1871
+ {
1872
+ "epoch": 1.45,
1873
+ "learning_rate": 1.1078605449930569e-05,
1874
+ "loss": 0.0231,
1875
+ "step": 311
1876
+ },
1877
+ {
1878
+ "epoch": 1.45,
1879
+ "learning_rate": 1.1028619798936418e-05,
1880
+ "loss": 0.0226,
1881
+ "step": 312
1882
+ },
1883
+ {
1884
+ "epoch": 1.46,
1885
+ "learning_rate": 1.0978608158696517e-05,
1886
+ "loss": 0.0164,
1887
+ "step": 313
1888
+ },
1889
+ {
1890
+ "epoch": 1.46,
1891
+ "learning_rate": 1.092857179281168e-05,
1892
+ "loss": 0.0204,
1893
+ "step": 314
1894
+ },
1895
+ {
1896
+ "epoch": 1.47,
1897
+ "learning_rate": 1.0878511965507435e-05,
1898
+ "loss": 0.0228,
1899
+ "step": 315
1900
+ },
1901
+ {
1902
+ "epoch": 1.47,
1903
+ "learning_rate": 1.0828429941602082e-05,
1904
+ "loss": 0.023,
1905
+ "step": 316
1906
+ },
1907
+ {
1908
+ "epoch": 1.47,
1909
+ "learning_rate": 1.0778326986474765e-05,
1910
+ "loss": 0.0208,
1911
+ "step": 317
1912
+ },
1913
+ {
1914
+ "epoch": 1.48,
1915
+ "learning_rate": 1.072820436603346e-05,
1916
+ "loss": 0.018,
1917
+ "step": 318
1918
+ },
1919
+ {
1920
+ "epoch": 1.48,
1921
+ "learning_rate": 1.067806334668301e-05,
1922
+ "loss": 0.0199,
1923
+ "step": 319
1924
+ },
1925
+ {
1926
+ "epoch": 1.49,
1927
+ "learning_rate": 1.0627905195293135e-05,
1928
+ "loss": 0.0218,
1929
+ "step": 320
1930
+ },
1931
+ {
1932
+ "epoch": 1.49,
1933
+ "learning_rate": 1.0577731179166415e-05,
1934
+ "loss": 0.0275,
1935
+ "step": 321
1936
+ },
1937
+ {
1938
+ "epoch": 1.5,
1939
+ "learning_rate": 1.0527542566006266e-05,
1940
+ "loss": 0.0201,
1941
+ "step": 322
1942
+ },
1943
+ {
1944
+ "epoch": 1.5,
1945
+ "learning_rate": 1.0477340623884917e-05,
1946
+ "loss": 0.0199,
1947
+ "step": 323
1948
+ },
1949
+ {
1950
+ "epoch": 1.51,
1951
+ "learning_rate": 1.0427126621211372e-05,
1952
+ "loss": 0.0243,
1953
+ "step": 324
1954
+ },
1955
+ {
1956
+ "epoch": 1.51,
1957
+ "learning_rate": 1.0376901826699349e-05,
1958
+ "loss": 0.0207,
1959
+ "step": 325
1960
+ },
1961
+ {
1962
+ "epoch": 1.52,
1963
+ "learning_rate": 1.032666750933524e-05,
1964
+ "loss": 0.0202,
1965
+ "step": 326
1966
+ },
1967
+ {
1968
+ "epoch": 1.52,
1969
+ "learning_rate": 1.0276424938346044e-05,
1970
+ "loss": 0.0209,
1971
+ "step": 327
1972
+ },
1973
+ {
1974
+ "epoch": 1.53,
1975
+ "learning_rate": 1.0226175383167299e-05,
1976
+ "loss": 0.0222,
1977
+ "step": 328
1978
+ },
1979
+ {
1980
+ "epoch": 1.53,
1981
+ "learning_rate": 1.0175920113410998e-05,
1982
+ "loss": 0.0191,
1983
+ "step": 329
1984
+ },
1985
+ {
1986
+ "epoch": 1.53,
1987
+ "learning_rate": 1.0125660398833528e-05,
1988
+ "loss": 0.0197,
1989
+ "step": 330
1990
+ },
1991
+ {
1992
+ "epoch": 1.54,
1993
+ "learning_rate": 1.0075397509303573e-05,
1994
+ "loss": 0.0258,
1995
+ "step": 331
1996
+ },
1997
+ {
1998
+ "epoch": 1.54,
1999
+ "learning_rate": 1.0025132714770041e-05,
2000
+ "loss": 0.02,
2001
+ "step": 332
2002
+ },
2003
+ {
2004
+ "epoch": 1.55,
2005
+ "learning_rate": 9.974867285229962e-06,
2006
+ "loss": 0.0224,
2007
+ "step": 333
2008
+ },
2009
+ {
2010
+ "epoch": 1.55,
2011
+ "learning_rate": 9.92460249069643e-06,
2012
+ "loss": 0.0236,
2013
+ "step": 334
2014
+ },
2015
+ {
2016
+ "epoch": 1.56,
2017
+ "learning_rate": 9.874339601166474e-06,
2018
+ "loss": 0.0242,
2019
+ "step": 335
2020
+ },
2021
+ {
2022
+ "epoch": 1.56,
2023
+ "learning_rate": 9.824079886589004e-06,
2024
+ "loss": 0.0219,
2025
+ "step": 336
2026
+ },
2027
+ {
2028
+ "epoch": 1.57,
2029
+ "learning_rate": 9.773824616832706e-06,
2030
+ "loss": 0.0169,
2031
+ "step": 337
2032
+ },
2033
+ {
2034
+ "epoch": 1.57,
2035
+ "learning_rate": 9.723575061653957e-06,
2036
+ "loss": 0.0302,
2037
+ "step": 338
2038
+ },
2039
+ {
2040
+ "epoch": 1.58,
2041
+ "learning_rate": 9.673332490664765e-06,
2042
+ "loss": 0.0225,
2043
+ "step": 339
2044
+ },
2045
+ {
2046
+ "epoch": 1.58,
2047
+ "learning_rate": 9.623098173300655e-06,
2048
+ "loss": 0.0224,
2049
+ "step": 340
2050
+ },
2051
+ {
2052
+ "epoch": 1.59,
2053
+ "learning_rate": 9.572873378788633e-06,
2054
+ "loss": 0.0199,
2055
+ "step": 341
2056
+ },
2057
+ {
2058
+ "epoch": 1.59,
2059
+ "learning_rate": 9.522659376115086e-06,
2060
+ "loss": 0.0271,
2061
+ "step": 342
2062
+ },
2063
+ {
2064
+ "epoch": 1.6,
2065
+ "learning_rate": 9.472457433993736e-06,
2066
+ "loss": 0.0248,
2067
+ "step": 343
2068
+ },
2069
+ {
2070
+ "epoch": 1.6,
2071
+ "learning_rate": 9.422268820833588e-06,
2072
+ "loss": 0.0208,
2073
+ "step": 344
2074
+ },
2075
+ {
2076
+ "epoch": 1.6,
2077
+ "learning_rate": 9.372094804706867e-06,
2078
+ "loss": 0.0227,
2079
+ "step": 345
2080
+ },
2081
+ {
2082
+ "epoch": 1.61,
2083
+ "learning_rate": 9.321936653316995e-06,
2084
+ "loss": 0.0203,
2085
+ "step": 346
2086
+ },
2087
+ {
2088
+ "epoch": 1.61,
2089
+ "learning_rate": 9.271795633966544e-06,
2090
+ "loss": 0.0236,
2091
+ "step": 347
2092
+ },
2093
+ {
2094
+ "epoch": 1.62,
2095
+ "learning_rate": 9.221673013525235e-06,
2096
+ "loss": 0.0219,
2097
+ "step": 348
2098
+ },
2099
+ {
2100
+ "epoch": 1.62,
2101
+ "learning_rate": 9.171570058397919e-06,
2102
+ "loss": 0.021,
2103
+ "step": 349
2104
+ },
2105
+ {
2106
+ "epoch": 1.63,
2107
+ "learning_rate": 9.121488034492569e-06,
2108
+ "loss": 0.0222,
2109
+ "step": 350
2110
+ },
2111
+ {
2112
+ "epoch": 1.63,
2113
+ "learning_rate": 9.071428207188324e-06,
2114
+ "loss": 0.0208,
2115
+ "step": 351
2116
+ },
2117
+ {
2118
+ "epoch": 1.64,
2119
+ "learning_rate": 9.021391841303484e-06,
2120
+ "loss": 0.0224,
2121
+ "step": 352
2122
+ },
2123
+ {
2124
+ "epoch": 1.64,
2125
+ "learning_rate": 8.971380201063586e-06,
2126
+ "loss": 0.0211,
2127
+ "step": 353
2128
+ },
2129
+ {
2130
+ "epoch": 1.65,
2131
+ "learning_rate": 8.921394550069434e-06,
2132
+ "loss": 0.022,
2133
+ "step": 354
2134
+ },
2135
+ {
2136
+ "epoch": 1.65,
2137
+ "learning_rate": 8.871436151265183e-06,
2138
+ "loss": 0.0192,
2139
+ "step": 355
2140
+ },
2141
+ {
2142
+ "epoch": 1.66,
2143
+ "learning_rate": 8.821506266906444e-06,
2144
+ "loss": 0.0202,
2145
+ "step": 356
2146
+ },
2147
+ {
2148
+ "epoch": 1.66,
2149
+ "learning_rate": 8.77160615852836e-06,
2150
+ "loss": 0.0217,
2151
+ "step": 357
2152
+ },
2153
+ {
2154
+ "epoch": 1.67,
2155
+ "learning_rate": 8.721737086913766e-06,
2156
+ "loss": 0.0193,
2157
+ "step": 358
2158
+ },
2159
+ {
2160
+ "epoch": 1.67,
2161
+ "learning_rate": 8.671900312061306e-06,
2162
+ "loss": 0.0167,
2163
+ "step": 359
2164
+ },
2165
+ {
2166
+ "epoch": 1.67,
2167
+ "learning_rate": 8.62209709315362e-06,
2168
+ "loss": 0.0234,
2169
+ "step": 360
2170
+ },
2171
+ {
2172
+ "epoch": 1.68,
2173
+ "learning_rate": 8.572328688525513e-06,
2174
+ "loss": 0.0233,
2175
+ "step": 361
2176
+ },
2177
+ {
2178
+ "epoch": 1.68,
2179
+ "learning_rate": 8.522596355632164e-06,
2180
+ "loss": 0.0247,
2181
+ "step": 362
2182
+ },
2183
+ {
2184
+ "epoch": 1.69,
2185
+ "learning_rate": 8.472901351017371e-06,
2186
+ "loss": 0.0209,
2187
+ "step": 363
2188
+ },
2189
+ {
2190
+ "epoch": 1.69,
2191
+ "learning_rate": 8.423244930281774e-06,
2192
+ "loss": 0.0231,
2193
+ "step": 364
2194
+ },
2195
+ {
2196
+ "epoch": 1.7,
2197
+ "learning_rate": 8.373628348051165e-06,
2198
+ "loss": 0.0196,
2199
+ "step": 365
2200
+ },
2201
+ {
2202
+ "epoch": 1.7,
2203
+ "learning_rate": 8.32405285794476e-06,
2204
+ "loss": 0.0188,
2205
+ "step": 366
2206
+ },
2207
+ {
2208
+ "epoch": 1.71,
2209
+ "learning_rate": 8.274519712543535e-06,
2210
+ "loss": 0.02,
2211
+ "step": 367
2212
+ },
2213
+ {
2214
+ "epoch": 1.71,
2215
+ "learning_rate": 8.22503016335859e-06,
2216
+ "loss": 0.0184,
2217
+ "step": 368
2218
+ },
2219
+ {
2220
+ "epoch": 1.72,
2221
+ "learning_rate": 8.175585460799499e-06,
2222
+ "loss": 0.0192,
2223
+ "step": 369
2224
+ },
2225
+ {
2226
+ "epoch": 1.72,
2227
+ "learning_rate": 8.126186854142752e-06,
2228
+ "loss": 0.0195,
2229
+ "step": 370
2230
+ },
2231
+ {
2232
+ "epoch": 1.73,
2233
+ "learning_rate": 8.076835591500172e-06,
2234
+ "loss": 0.0219,
2235
+ "step": 371
2236
+ },
2237
+ {
2238
+ "epoch": 1.73,
2239
+ "learning_rate": 8.027532919787372e-06,
2240
+ "loss": 0.0208,
2241
+ "step": 372
2242
+ },
2243
+ {
2244
+ "epoch": 1.73,
2245
+ "learning_rate": 7.978280084692268e-06,
2246
+ "loss": 0.0194,
2247
+ "step": 373
2248
+ },
2249
+ {
2250
+ "epoch": 1.74,
2251
+ "learning_rate": 7.92907833064359e-06,
2252
+ "loss": 0.0242,
2253
+ "step": 374
2254
+ },
2255
+ {
2256
+ "epoch": 1.74,
2257
+ "learning_rate": 7.879928900779457e-06,
2258
+ "loss": 0.0203,
2259
+ "step": 375
2260
+ },
2261
+ {
2262
+ "epoch": 1.75,
2263
+ "learning_rate": 7.830833036915947e-06,
2264
+ "loss": 0.0212,
2265
+ "step": 376
2266
+ },
2267
+ {
2268
+ "epoch": 1.75,
2269
+ "learning_rate": 7.781791979515749e-06,
2270
+ "loss": 0.0225,
2271
+ "step": 377
2272
+ },
2273
+ {
2274
+ "epoch": 1.76,
2275
+ "learning_rate": 7.732806967656785e-06,
2276
+ "loss": 0.0263,
2277
+ "step": 378
2278
+ },
2279
+ {
2280
+ "epoch": 1.76,
2281
+ "learning_rate": 7.683879239000936e-06,
2282
+ "loss": 0.0184,
2283
+ "step": 379
2284
+ },
2285
+ {
2286
+ "epoch": 1.77,
2287
+ "learning_rate": 7.635010029762755e-06,
2288
+ "loss": 0.0166,
2289
+ "step": 380
2290
+ },
2291
+ {
2292
+ "epoch": 1.77,
2293
+ "learning_rate": 7.586200574678231e-06,
2294
+ "loss": 0.0218,
2295
+ "step": 381
2296
+ },
2297
+ {
2298
+ "epoch": 1.78,
2299
+ "learning_rate": 7.5374521069736e-06,
2300
+ "loss": 0.0199,
2301
+ "step": 382
2302
+ },
2303
+ {
2304
+ "epoch": 1.78,
2305
+ "learning_rate": 7.488765858334188e-06,
2306
+ "loss": 0.0219,
2307
+ "step": 383
2308
+ },
2309
+ {
2310
+ "epoch": 1.79,
2311
+ "learning_rate": 7.440143058873275e-06,
2312
+ "loss": 0.0231,
2313
+ "step": 384
2314
+ },
2315
+ {
2316
+ "epoch": 1.79,
2317
+ "learning_rate": 7.391584937101034e-06,
2318
+ "loss": 0.0207,
2319
+ "step": 385
2320
+ },
2321
+ {
2322
+ "epoch": 1.8,
2323
+ "learning_rate": 7.3430927198934745e-06,
2324
+ "loss": 0.0189,
2325
+ "step": 386
2326
+ },
2327
+ {
2328
+ "epoch": 1.8,
2329
+ "learning_rate": 7.294667632461463e-06,
2330
+ "loss": 0.0255,
2331
+ "step": 387
2332
+ },
2333
+ {
2334
+ "epoch": 1.8,
2335
+ "learning_rate": 7.246310898319753e-06,
2336
+ "loss": 0.0229,
2337
+ "step": 388
2338
+ },
2339
+ {
2340
+ "epoch": 1.81,
2341
+ "learning_rate": 7.198023739256066e-06,
2342
+ "loss": 0.0194,
2343
+ "step": 389
2344
+ },
2345
+ {
2346
+ "epoch": 1.81,
2347
+ "learning_rate": 7.149807375300239e-06,
2348
+ "loss": 0.0204,
2349
+ "step": 390
2350
+ },
2351
+ {
2352
+ "epoch": 1.82,
2353
+ "learning_rate": 7.10166302469339e-06,
2354
+ "loss": 0.0213,
2355
+ "step": 391
2356
+ },
2357
+ {
2358
+ "epoch": 1.82,
2359
+ "learning_rate": 7.053591903857138e-06,
2360
+ "loss": 0.0206,
2361
+ "step": 392
2362
+ },
2363
+ {
2364
+ "epoch": 1.83,
2365
+ "learning_rate": 7.005595227362858e-06,
2366
+ "loss": 0.0239,
2367
+ "step": 393
2368
+ },
2369
+ {
2370
+ "epoch": 1.83,
2371
+ "learning_rate": 6.957674207901018e-06,
2372
+ "loss": 0.0214,
2373
+ "step": 394
2374
+ },
2375
+ {
2376
+ "epoch": 1.84,
2377
+ "learning_rate": 6.909830056250527e-06,
2378
+ "loss": 0.0198,
2379
+ "step": 395
2380
+ },
2381
+ {
2382
+ "epoch": 1.84,
2383
+ "learning_rate": 6.862063981248126e-06,
2384
+ "loss": 0.0221,
2385
+ "step": 396
2386
+ },
2387
+ {
2388
+ "epoch": 1.85,
2389
+ "learning_rate": 6.814377189757876e-06,
2390
+ "loss": 0.0245,
2391
+ "step": 397
2392
+ },
2393
+ {
2394
+ "epoch": 1.85,
2395
+ "learning_rate": 6.766770886640637e-06,
2396
+ "loss": 0.0219,
2397
+ "step": 398
2398
+ },
2399
+ {
2400
+ "epoch": 1.86,
2401
+ "learning_rate": 6.71924627472365e-06,
2402
+ "loss": 0.0206,
2403
+ "step": 399
2404
+ },
2405
+ {
2406
+ "epoch": 1.86,
2407
+ "learning_rate": 6.671804554770135e-06,
2408
+ "loss": 0.0215,
2409
+ "step": 400
2410
+ },
2411
+ {
2412
+ "epoch": 1.87,
2413
+ "learning_rate": 6.624446925448944e-06,
2414
+ "loss": 0.026,
2415
+ "step": 401
2416
+ },
2417
+ {
2418
+ "epoch": 1.87,
2419
+ "learning_rate": 6.577174583304289e-06,
2420
+ "loss": 0.0198,
2421
+ "step": 402
2422
+ },
2423
+ {
2424
+ "epoch": 1.87,
2425
+ "learning_rate": 6.529988722725506e-06,
2426
+ "loss": 0.0232,
2427
+ "step": 403
2428
+ },
2429
+ {
2430
+ "epoch": 1.88,
2431
+ "learning_rate": 6.482890535916876e-06,
2432
+ "loss": 0.0222,
2433
+ "step": 404
2434
+ },
2435
+ {
2436
+ "epoch": 1.88,
2437
+ "learning_rate": 6.435881212867494e-06,
2438
+ "loss": 0.0203,
2439
+ "step": 405
2440
+ },
2441
+ {
2442
+ "epoch": 1.89,
2443
+ "learning_rate": 6.388961941321225e-06,
2444
+ "loss": 0.0218,
2445
+ "step": 406
2446
+ },
2447
+ {
2448
+ "epoch": 1.89,
2449
+ "learning_rate": 6.342133906746676e-06,
2450
+ "loss": 0.0222,
2451
+ "step": 407
2452
+ },
2453
+ {
2454
+ "epoch": 1.9,
2455
+ "learning_rate": 6.295398292307242e-06,
2456
+ "loss": 0.0218,
2457
+ "step": 408
2458
+ },
2459
+ {
2460
+ "epoch": 1.9,
2461
+ "learning_rate": 6.2487562788312275e-06,
2462
+ "loss": 0.0242,
2463
+ "step": 409
2464
+ },
2465
+ {
2466
+ "epoch": 1.91,
2467
+ "learning_rate": 6.202209044781991e-06,
2468
+ "loss": 0.017,
2469
+ "step": 410
2470
+ },
2471
+ {
2472
+ "epoch": 1.91,
2473
+ "learning_rate": 6.155757766228192e-06,
2474
+ "loss": 0.0198,
2475
+ "step": 411
2476
+ },
2477
+ {
2478
+ "epoch": 1.92,
2479
+ "learning_rate": 6.109403616814069e-06,
2480
+ "loss": 0.0164,
2481
+ "step": 412
2482
+ },
2483
+ {
2484
+ "epoch": 1.92,
2485
+ "learning_rate": 6.063147767729764e-06,
2486
+ "loss": 0.0222,
2487
+ "step": 413
2488
+ },
2489
+ {
2490
+ "epoch": 1.93,
2491
+ "learning_rate": 6.016991387681775e-06,
2492
+ "loss": 0.0194,
2493
+ "step": 414
2494
+ },
2495
+ {
2496
+ "epoch": 1.93,
2497
+ "learning_rate": 5.970935642863375e-06,
2498
+ "loss": 0.0216,
2499
+ "step": 415
2500
+ },
2501
+ {
2502
+ "epoch": 1.93,
2503
+ "learning_rate": 5.924981696925192e-06,
2504
+ "loss": 0.0212,
2505
+ "step": 416
2506
+ },
2507
+ {
2508
+ "epoch": 1.94,
2509
+ "learning_rate": 5.879130710945791e-06,
2510
+ "loss": 0.018,
2511
+ "step": 417
2512
+ },
2513
+ {
2514
+ "epoch": 1.94,
2515
+ "learning_rate": 5.83338384340232e-06,
2516
+ "loss": 0.0201,
2517
+ "step": 418
2518
+ },
2519
+ {
2520
+ "epoch": 1.95,
2521
+ "learning_rate": 5.787742250141279e-06,
2522
+ "loss": 0.0185,
2523
+ "step": 419
2524
+ },
2525
+ {
2526
+ "epoch": 1.95,
2527
+ "learning_rate": 5.742207084349274e-06,
2528
+ "loss": 0.0213,
2529
+ "step": 420
2530
+ },
2531
+ {
2532
+ "epoch": 1.96,
2533
+ "learning_rate": 5.696779496523913e-06,
2534
+ "loss": 0.0219,
2535
+ "step": 421
2536
+ },
2537
+ {
2538
+ "epoch": 1.96,
2539
+ "learning_rate": 5.651460634444716e-06,
2540
+ "loss": 0.0176,
2541
+ "step": 422
2542
+ },
2543
+ {
2544
+ "epoch": 1.97,
2545
+ "learning_rate": 5.606251643144136e-06,
2546
+ "loss": 0.0191,
2547
+ "step": 423
2548
+ },
2549
+ {
2550
+ "epoch": 1.97,
2551
+ "learning_rate": 5.561153664878603e-06,
2552
+ "loss": 0.0213,
2553
+ "step": 424
2554
+ },
2555
+ {
2556
+ "epoch": 1.98,
2557
+ "learning_rate": 5.516167839099679e-06,
2558
+ "loss": 0.0181,
2559
+ "step": 425
2560
+ },
2561
+ {
2562
+ "epoch": 1.98,
2563
+ "learning_rate": 5.471295302425277e-06,
2564
+ "loss": 0.0229,
2565
+ "step": 426
2566
+ },
2567
+ {
2568
+ "epoch": 1.99,
2569
+ "learning_rate": 5.426537188610916e-06,
2570
+ "loss": 0.0214,
2571
+ "step": 427
2572
+ },
2573
+ {
2574
+ "epoch": 1.99,
2575
+ "learning_rate": 5.3818946285211045e-06,
2576
+ "loss": 0.0262,
2577
+ "step": 428
2578
+ },
2579
+ {
2580
+ "epoch": 2.0,
2581
+ "learning_rate": 5.337368750100756e-06,
2582
+ "loss": 0.0171,
2583
+ "step": 429
2584
+ },
2585
+ {
2586
+ "epoch": 2.0,
2587
+ "learning_rate": 5.292960678346674e-06,
2588
+ "loss": 0.0246,
2589
+ "step": 430
2590
+ },
2591
+ {
2592
+ "epoch": 2.0,
2593
+ "learning_rate": 5.248671535279164e-06,
2594
+ "loss": 0.0192,
2595
+ "step": 431
2596
+ },
2597
+ {
2598
+ "epoch": 2.01,
2599
+ "learning_rate": 5.204502439913641e-06,
2600
+ "loss": 0.018,
2601
+ "step": 432
2602
+ },
2603
+ {
2604
+ "epoch": 2.01,
2605
+ "learning_rate": 5.160454508232398e-06,
2606
+ "loss": 0.0168,
2607
+ "step": 433
2608
+ },
2609
+ {
2610
+ "epoch": 2.02,
2611
+ "learning_rate": 5.116528853156384e-06,
2612
+ "loss": 0.0208,
2613
+ "step": 434
2614
+ },
2615
+ {
2616
+ "epoch": 2.02,
2617
+ "learning_rate": 5.072726584517086e-06,
2618
+ "loss": 0.0207,
2619
+ "step": 435
2620
+ },
2621
+ {
2622
+ "epoch": 2.03,
2623
+ "learning_rate": 5.029048809028496e-06,
2624
+ "loss": 0.0172,
2625
+ "step": 436
2626
+ },
2627
+ {
2628
+ "epoch": 2.03,
2629
+ "learning_rate": 4.98549663025914e-06,
2630
+ "loss": 0.0198,
2631
+ "step": 437
2632
+ },
2633
+ {
2634
+ "epoch": 2.04,
2635
+ "learning_rate": 4.942071148604215e-06,
2636
+ "loss": 0.0199,
2637
+ "step": 438
2638
+ },
2639
+ {
2640
+ "epoch": 2.04,
2641
+ "learning_rate": 4.8987734612577544e-06,
2642
+ "loss": 0.0167,
2643
+ "step": 439
2644
+ },
2645
+ {
2646
+ "epoch": 2.05,
2647
+ "learning_rate": 4.855604662184935e-06,
2648
+ "loss": 0.02,
2649
+ "step": 440
2650
+ },
2651
+ {
2652
+ "epoch": 2.05,
2653
+ "learning_rate": 4.812565842094426e-06,
2654
+ "loss": 0.0175,
2655
+ "step": 441
2656
+ },
2657
+ {
2658
+ "epoch": 2.06,
2659
+ "learning_rate": 4.769658088410819e-06,
2660
+ "loss": 0.019,
2661
+ "step": 442
2662
+ },
2663
+ {
2664
+ "epoch": 2.06,
2665
+ "learning_rate": 4.726882485247177e-06,
2666
+ "loss": 0.0151,
2667
+ "step": 443
2668
+ },
2669
+ {
2670
+ "epoch": 2.07,
2671
+ "learning_rate": 4.684240113377619e-06,
2672
+ "loss": 0.0171,
2673
+ "step": 444
2674
+ },
2675
+ {
2676
+ "epoch": 2.07,
2677
+ "learning_rate": 4.641732050210032e-06,
2678
+ "loss": 0.018,
2679
+ "step": 445
2680
+ },
2681
+ {
2682
+ "epoch": 2.07,
2683
+ "learning_rate": 4.599359369758842e-06,
2684
+ "loss": 0.0175,
2685
+ "step": 446
2686
+ },
2687
+ {
2688
+ "epoch": 2.08,
2689
+ "learning_rate": 4.557123142617873e-06,
2690
+ "loss": 0.018,
2691
+ "step": 447
2692
+ },
2693
+ {
2694
+ "epoch": 2.08,
2695
+ "learning_rate": 4.5150244359333e-06,
2696
+ "loss": 0.0153,
2697
+ "step": 448
2698
+ },
2699
+ {
2700
+ "epoch": 2.09,
2701
+ "learning_rate": 4.4730643133766874e-06,
2702
+ "loss": 0.0216,
2703
+ "step": 449
2704
+ },
2705
+ {
2706
+ "epoch": 2.09,
2707
+ "learning_rate": 4.4312438351181246e-06,
2708
+ "loss": 0.0207,
2709
+ "step": 450
2710
+ },
2711
+ {
2712
+ "epoch": 2.1,
2713
+ "learning_rate": 4.389564057799413e-06,
2714
+ "loss": 0.0165,
2715
+ "step": 451
2716
+ },
2717
+ {
2718
+ "epoch": 2.1,
2719
+ "learning_rate": 4.348026034507402e-06,
2720
+ "loss": 0.0153,
2721
+ "step": 452
2722
+ },
2723
+ {
2724
+ "epoch": 2.11,
2725
+ "learning_rate": 4.306630814747358e-06,
2726
+ "loss": 0.0166,
2727
+ "step": 453
2728
+ },
2729
+ {
2730
+ "epoch": 2.11,
2731
+ "learning_rate": 4.265379444416446e-06,
2732
+ "loss": 0.017,
2733
+ "step": 454
2734
+ },
2735
+ {
2736
+ "epoch": 2.12,
2737
+ "learning_rate": 4.224272965777326e-06,
2738
+ "loss": 0.0275,
2739
+ "step": 455
2740
+ },
2741
+ {
2742
+ "epoch": 2.12,
2743
+ "learning_rate": 4.183312417431793e-06,
2744
+ "loss": 0.0152,
2745
+ "step": 456
2746
+ },
2747
+ {
2748
+ "epoch": 2.13,
2749
+ "learning_rate": 4.142498834294555e-06,
2750
+ "loss": 0.0235,
2751
+ "step": 457
2752
+ },
2753
+ {
2754
+ "epoch": 2.13,
2755
+ "learning_rate": 4.1018332475670795e-06,
2756
+ "loss": 0.021,
2757
+ "step": 458
2758
+ },
2759
+ {
2760
+ "epoch": 2.13,
2761
+ "learning_rate": 4.06131668471153e-06,
2762
+ "loss": 0.0194,
2763
+ "step": 459
2764
+ },
2765
+ {
2766
+ "epoch": 2.14,
2767
+ "learning_rate": 4.020950169424815e-06,
2768
+ "loss": 0.0193,
2769
+ "step": 460
2770
+ },
2771
+ {
2772
+ "epoch": 2.14,
2773
+ "learning_rate": 3.980734721612718e-06,
2774
+ "loss": 0.0181,
2775
+ "step": 461
2776
+ },
2777
+ {
2778
+ "epoch": 2.15,
2779
+ "learning_rate": 3.940671357364137e-06,
2780
+ "loss": 0.022,
2781
+ "step": 462
2782
+ },
2783
+ {
2784
+ "epoch": 2.15,
2785
+ "learning_rate": 3.900761088925411e-06,
2786
+ "loss": 0.0134,
2787
+ "step": 463
2788
+ },
2789
+ {
2790
+ "epoch": 2.16,
2791
+ "learning_rate": 3.861004924674724e-06,
2792
+ "loss": 0.0183,
2793
+ "step": 464
2794
+ },
2795
+ {
2796
+ "epoch": 2.16,
2797
+ "learning_rate": 3.821403869096658e-06,
2798
+ "loss": 0.0191,
2799
+ "step": 465
2800
+ },
2801
+ {
2802
+ "epoch": 2.17,
2803
+ "learning_rate": 3.7819589227567886e-06,
2804
+ "loss": 0.0175,
2805
+ "step": 466
2806
+ },
2807
+ {
2808
+ "epoch": 2.17,
2809
+ "learning_rate": 3.7426710822764234e-06,
2810
+ "loss": 0.0168,
2811
+ "step": 467
2812
+ },
2813
+ {
2814
+ "epoch": 2.18,
2815
+ "learning_rate": 3.7035413403073995e-06,
2816
+ "loss": 0.0169,
2817
+ "step": 468
2818
+ },
2819
+ {
2820
+ "epoch": 2.18,
2821
+ "learning_rate": 3.6645706855070296e-06,
2822
+ "loss": 0.0175,
2823
+ "step": 469
2824
+ },
2825
+ {
2826
+ "epoch": 2.19,
2827
+ "learning_rate": 3.625760102513103e-06,
2828
+ "loss": 0.0175,
2829
+ "step": 470
2830
+ },
2831
+ {
2832
+ "epoch": 2.19,
2833
+ "learning_rate": 3.5871105719190115e-06,
2834
+ "loss": 0.0173,
2835
+ "step": 471
2836
+ },
2837
+ {
2838
+ "epoch": 2.2,
2839
+ "learning_rate": 3.5486230702489767e-06,
2840
+ "loss": 0.0215,
2841
+ "step": 472
2842
+ },
2843
+ {
2844
+ "epoch": 2.2,
2845
+ "learning_rate": 3.5102985699333713e-06,
2846
+ "loss": 0.0201,
2847
+ "step": 473
2848
+ },
2849
+ {
2850
+ "epoch": 2.2,
2851
+ "learning_rate": 3.4721380392841632e-06,
2852
+ "loss": 0.0233,
2853
+ "step": 474
2854
+ },
2855
+ {
2856
+ "epoch": 2.21,
2857
+ "learning_rate": 3.4341424424704373e-06,
2858
+ "loss": 0.0213,
2859
+ "step": 475
2860
+ },
2861
+ {
2862
+ "epoch": 2.21,
2863
+ "learning_rate": 3.396312739494032e-06,
2864
+ "loss": 0.0279,
2865
+ "step": 476
2866
+ },
2867
+ {
2868
+ "epoch": 2.22,
2869
+ "learning_rate": 3.3586498861653004e-06,
2870
+ "loss": 0.0195,
2871
+ "step": 477
2872
+ },
2873
+ {
2874
+ "epoch": 2.22,
2875
+ "learning_rate": 3.3211548340789366e-06,
2876
+ "loss": 0.0187,
2877
+ "step": 478
2878
+ },
2879
+ {
2880
+ "epoch": 2.23,
2881
+ "learning_rate": 3.2838285305899586e-06,
2882
+ "loss": 0.0167,
2883
+ "step": 479
2884
+ },
2885
+ {
2886
+ "epoch": 2.23,
2887
+ "learning_rate": 3.2466719187897555e-06,
2888
+ "loss": 0.0178,
2889
+ "step": 480
2890
+ },
2891
+ {
2892
+ "epoch": 2.24,
2893
+ "learning_rate": 3.2096859374822587e-06,
2894
+ "loss": 0.018,
2895
+ "step": 481
2896
+ },
2897
+ {
2898
+ "epoch": 2.24,
2899
+ "learning_rate": 3.172871521160236e-06,
2900
+ "loss": 0.018,
2901
+ "step": 482
2902
+ },
2903
+ {
2904
+ "epoch": 2.25,
2905
+ "learning_rate": 3.1362295999816617e-06,
2906
+ "loss": 0.0172,
2907
+ "step": 483
2908
+ },
2909
+ {
2910
+ "epoch": 2.25,
2911
+ "learning_rate": 3.0997610997462292e-06,
2912
+ "loss": 0.0158,
2913
+ "step": 484
2914
+ },
2915
+ {
2916
+ "epoch": 2.26,
2917
+ "learning_rate": 3.063466941871952e-06,
2918
+ "loss": 0.0186,
2919
+ "step": 485
2920
+ },
2921
+ {
2922
+ "epoch": 2.26,
2923
+ "learning_rate": 3.027348043371893e-06,
2924
+ "loss": 0.0178,
2925
+ "step": 486
2926
+ },
2927
+ {
2928
+ "epoch": 2.27,
2929
+ "learning_rate": 2.991405316830985e-06,
2930
+ "loss": 0.0155,
2931
+ "step": 487
2932
+ },
2933
+ {
2934
+ "epoch": 2.27,
2935
+ "learning_rate": 2.9556396703829705e-06,
2936
+ "loss": 0.0144,
2937
+ "step": 488
2938
+ },
2939
+ {
2940
+ "epoch": 2.27,
2941
+ "learning_rate": 2.920052007687475e-06,
2942
+ "loss": 0.0181,
2943
+ "step": 489
2944
+ },
2945
+ {
2946
+ "epoch": 2.28,
2947
+ "learning_rate": 2.884643227907147e-06,
2948
+ "loss": 0.0132,
2949
+ "step": 490
2950
+ },
2951
+ {
2952
+ "epoch": 2.28,
2953
+ "learning_rate": 2.8494142256849676e-06,
2954
+ "loss": 0.0182,
2955
+ "step": 491
2956
+ },
2957
+ {
2958
+ "epoch": 2.29,
2959
+ "learning_rate": 2.814365891121634e-06,
2960
+ "loss": 0.0168,
2961
+ "step": 492
2962
+ },
2963
+ {
2964
+ "epoch": 2.29,
2965
+ "learning_rate": 2.77949910975306e-06,
2966
+ "loss": 0.0153,
2967
+ "step": 493
2968
+ },
2969
+ {
2970
+ "epoch": 2.3,
2971
+ "learning_rate": 2.744814762528022e-06,
2972
+ "loss": 0.0155,
2973
+ "step": 494
2974
+ },
2975
+ {
2976
+ "epoch": 2.3,
2977
+ "learning_rate": 2.7103137257858867e-06,
2978
+ "loss": 0.0192,
2979
+ "step": 495
2980
+ },
2981
+ {
2982
+ "epoch": 2.31,
2983
+ "learning_rate": 2.67599687123447e-06,
2984
+ "loss": 0.0208,
2985
+ "step": 496
2986
+ },
2987
+ {
2988
+ "epoch": 2.31,
2989
+ "learning_rate": 2.6418650659280253e-06,
2990
+ "loss": 0.021,
2991
+ "step": 497
2992
+ },
2993
+ {
2994
+ "epoch": 2.32,
2995
+ "learning_rate": 2.6079191722453157e-06,
2996
+ "loss": 0.0201,
2997
+ "step": 498
2998
+ },
2999
+ {
3000
+ "epoch": 2.32,
3001
+ "learning_rate": 2.5741600478678474e-06,
3002
+ "loss": 0.0199,
3003
+ "step": 499
3004
+ },
3005
+ {
3006
+ "epoch": 2.33,
3007
+ "learning_rate": 2.5405885457581793e-06,
3008
+ "loss": 0.0165,
3009
+ "step": 500
3010
+ },
3011
+ {
3012
+ "epoch": 2.33,
3013
+ "learning_rate": 2.5072055141383876e-06,
3014
+ "loss": 0.0199,
3015
+ "step": 501
3016
+ },
3017
+ {
3018
+ "epoch": 2.33,
3019
+ "learning_rate": 2.4740117964686218e-06,
3020
+ "loss": 0.0201,
3021
+ "step": 502
3022
+ },
3023
+ {
3024
+ "epoch": 2.34,
3025
+ "learning_rate": 2.4410082314258034e-06,
3026
+ "loss": 0.0174,
3027
+ "step": 503
3028
+ },
3029
+ {
3030
+ "epoch": 2.34,
3031
+ "learning_rate": 2.408195652882436e-06,
3032
+ "loss": 0.0189,
3033
+ "step": 504
3034
+ },
3035
+ {
3036
+ "epoch": 2.35,
3037
+ "learning_rate": 2.37557488988552e-06,
3038
+ "loss": 0.0148,
3039
+ "step": 505
3040
+ },
3041
+ {
3042
+ "epoch": 2.35,
3043
+ "learning_rate": 2.3431467666356356e-06,
3044
+ "loss": 0.017,
3045
+ "step": 506
3046
+ },
3047
+ {
3048
+ "epoch": 2.36,
3049
+ "learning_rate": 2.3109121024660873e-06,
3050
+ "loss": 0.0171,
3051
+ "step": 507
3052
+ },
3053
+ {
3054
+ "epoch": 2.36,
3055
+ "learning_rate": 2.2788717118222203e-06,
3056
+ "loss": 0.021,
3057
+ "step": 508
3058
+ },
3059
+ {
3060
+ "epoch": 2.37,
3061
+ "learning_rate": 2.2470264042408452e-06,
3062
+ "loss": 0.0183,
3063
+ "step": 509
3064
+ },
3065
+ {
3066
+ "epoch": 2.37,
3067
+ "learning_rate": 2.215376984329767e-06,
3068
+ "loss": 0.0144,
3069
+ "step": 510
3070
+ },
3071
+ {
3072
+ "epoch": 2.38,
3073
+ "learning_rate": 2.183924251747477e-06,
3074
+ "loss": 0.0179,
3075
+ "step": 511
3076
+ },
3077
+ {
3078
+ "epoch": 2.38,
3079
+ "learning_rate": 2.1526690011829297e-06,
3080
+ "loss": 0.0197,
3081
+ "step": 512
3082
+ },
3083
+ {
3084
+ "epoch": 2.39,
3085
+ "learning_rate": 2.1216120223354817e-06,
3086
+ "loss": 0.0158,
3087
+ "step": 513
3088
+ },
3089
+ {
3090
+ "epoch": 2.39,
3091
+ "learning_rate": 2.090754099894917e-06,
3092
+ "loss": 0.0185,
3093
+ "step": 514
3094
+ },
3095
+ {
3096
+ "epoch": 2.4,
3097
+ "learning_rate": 2.0600960135216463e-06,
3098
+ "loss": 0.0138,
3099
+ "step": 515
3100
+ },
3101
+ {
3102
+ "epoch": 2.4,
3103
+ "learning_rate": 2.02963853782699e-06,
3104
+ "loss": 0.0156,
3105
+ "step": 516
3106
+ },
3107
+ {
3108
+ "epoch": 2.4,
3109
+ "learning_rate": 1.99938244235361e-06,
3110
+ "loss": 0.022,
3111
+ "step": 517
3112
+ },
3113
+ {
3114
+ "epoch": 2.41,
3115
+ "learning_rate": 1.969328491556074e-06,
3116
+ "loss": 0.0201,
3117
+ "step": 518
3118
+ },
3119
+ {
3120
+ "epoch": 2.41,
3121
+ "learning_rate": 1.93947744478153e-06,
3122
+ "loss": 0.0171,
3123
+ "step": 519
3124
+ },
3125
+ {
3126
+ "epoch": 2.42,
3127
+ "learning_rate": 1.9098300562505266e-06,
3128
+ "loss": 0.0167,
3129
+ "step": 520
3130
+ },
3131
+ {
3132
+ "epoch": 2.42,
3133
+ "learning_rate": 1.8803870750379605e-06,
3134
+ "loss": 0.0173,
3135
+ "step": 521
3136
+ },
3137
+ {
3138
+ "epoch": 2.43,
3139
+ "learning_rate": 1.8511492450541379e-06,
3140
+ "loss": 0.0201,
3141
+ "step": 522
3142
+ },
3143
+ {
3144
+ "epoch": 2.43,
3145
+ "learning_rate": 1.8221173050259976e-06,
3146
+ "loss": 0.0171,
3147
+ "step": 523
3148
+ },
3149
+ {
3150
+ "epoch": 2.44,
3151
+ "learning_rate": 1.7932919884784216e-06,
3152
+ "loss": 0.0172,
3153
+ "step": 524
3154
+ },
3155
+ {
3156
+ "epoch": 2.44,
3157
+ "learning_rate": 1.7646740237157256e-06,
3158
+ "loss": 0.0215,
3159
+ "step": 525
3160
+ },
3161
+ {
3162
+ "epoch": 2.45,
3163
+ "learning_rate": 1.736264133803246e-06,
3164
+ "loss": 0.0174,
3165
+ "step": 526
3166
+ },
3167
+ {
3168
+ "epoch": 2.45,
3169
+ "learning_rate": 1.7080630365490658e-06,
3170
+ "loss": 0.0181,
3171
+ "step": 527
3172
+ },
3173
+ {
3174
+ "epoch": 2.46,
3175
+ "learning_rate": 1.6800714444858945e-06,
3176
+ "loss": 0.0224,
3177
+ "step": 528
3178
+ },
3179
+ {
3180
+ "epoch": 2.46,
3181
+ "learning_rate": 1.652290064853045e-06,
3182
+ "loss": 0.0203,
3183
+ "step": 529
3184
+ },
3185
+ {
3186
+ "epoch": 2.47,
3187
+ "learning_rate": 1.6247195995785836e-06,
3188
+ "loss": 0.0197,
3189
+ "step": 530
3190
+ },
3191
+ {
3192
+ "epoch": 2.47,
3193
+ "learning_rate": 1.5973607452615802e-06,
3194
+ "loss": 0.0167,
3195
+ "step": 531
3196
+ },
3197
+ {
3198
+ "epoch": 2.47,
3199
+ "learning_rate": 1.5702141931545144e-06,
3200
+ "loss": 0.0212,
3201
+ "step": 532
3202
+ },
3203
+ {
3204
+ "epoch": 2.48,
3205
+ "learning_rate": 1.5432806291458146e-06,
3206
+ "loss": 0.0196,
3207
+ "step": 533
3208
+ },
3209
+ {
3210
+ "epoch": 2.48,
3211
+ "learning_rate": 1.5165607337425158e-06,
3212
+ "loss": 0.0205,
3213
+ "step": 534
3214
+ },
3215
+ {
3216
+ "epoch": 2.49,
3217
+ "learning_rate": 1.490055182053083e-06,
3218
+ "loss": 0.0186,
3219
+ "step": 535
3220
+ },
3221
+ {
3222
+ "epoch": 2.49,
3223
+ "learning_rate": 1.4637646437703335e-06,
3224
+ "loss": 0.0177,
3225
+ "step": 536
3226
+ },
3227
+ {
3228
+ "epoch": 2.5,
3229
+ "learning_rate": 1.4376897831545345e-06,
3230
+ "loss": 0.0202,
3231
+ "step": 537
3232
+ },
3233
+ {
3234
+ "epoch": 2.5,
3235
+ "learning_rate": 1.4118312590166128e-06,
3236
+ "loss": 0.0176,
3237
+ "step": 538
3238
+ },
3239
+ {
3240
+ "epoch": 2.51,
3241
+ "learning_rate": 1.386189724701501e-06,
3242
+ "loss": 0.0166,
3243
+ "step": 539
3244
+ },
3245
+ {
3246
+ "epoch": 2.51,
3247
+ "learning_rate": 1.3607658280716474e-06,
3248
+ "loss": 0.0246,
3249
+ "step": 540
3250
+ },
3251
+ {
3252
+ "epoch": 2.52,
3253
+ "learning_rate": 1.3355602114906285e-06,
3254
+ "loss": 0.0159,
3255
+ "step": 541
3256
+ },
3257
+ {
3258
+ "epoch": 2.52,
3259
+ "learning_rate": 1.3105735118069373e-06,
3260
+ "loss": 0.0186,
3261
+ "step": 542
3262
+ },
3263
+ {
3264
+ "epoch": 2.53,
3265
+ "learning_rate": 1.2858063603378724e-06,
3266
+ "loss": 0.0165,
3267
+ "step": 543
3268
+ },
3269
+ {
3270
+ "epoch": 2.53,
3271
+ "learning_rate": 1.261259382853608e-06,
3272
+ "loss": 0.0188,
3273
+ "step": 544
3274
+ },
3275
+ {
3276
+ "epoch": 2.53,
3277
+ "learning_rate": 1.2369331995613664e-06,
3278
+ "loss": 0.0148,
3279
+ "step": 545
3280
+ },
3281
+ {
3282
+ "epoch": 2.54,
3283
+ "learning_rate": 1.2128284250897526e-06,
3284
+ "loss": 0.0222,
3285
+ "step": 546
3286
+ },
3287
+ {
3288
+ "epoch": 2.54,
3289
+ "learning_rate": 1.1889456684732358e-06,
3290
+ "loss": 0.0206,
3291
+ "step": 547
3292
+ },
3293
+ {
3294
+ "epoch": 2.55,
3295
+ "learning_rate": 1.165285533136743e-06,
3296
+ "loss": 0.0162,
3297
+ "step": 548
3298
+ },
3299
+ {
3300
+ "epoch": 2.55,
3301
+ "learning_rate": 1.1418486168804255e-06,
3302
+ "loss": 0.019,
3303
+ "step": 549
3304
+ },
3305
+ {
3306
+ "epoch": 2.56,
3307
+ "learning_rate": 1.1186355118645552e-06,
3308
+ "loss": 0.0214,
3309
+ "step": 550
3310
+ },
3311
+ {
3312
+ "epoch": 2.56,
3313
+ "learning_rate": 1.0956468045945512e-06,
3314
+ "loss": 0.0155,
3315
+ "step": 551
3316
+ },
3317
+ {
3318
+ "epoch": 2.57,
3319
+ "learning_rate": 1.0728830759061781e-06,
3320
+ "loss": 0.0186,
3321
+ "step": 552
3322
+ },
3323
+ {
3324
+ "epoch": 2.57,
3325
+ "learning_rate": 1.0503449009508516e-06,
3326
+ "loss": 0.0178,
3327
+ "step": 553
3328
+ },
3329
+ {
3330
+ "epoch": 2.58,
3331
+ "learning_rate": 1.0280328491811265e-06,
3332
+ "loss": 0.0153,
3333
+ "step": 554
3334
+ },
3335
+ {
3336
+ "epoch": 2.58,
3337
+ "learning_rate": 1.0059474843362893e-06,
3338
+ "loss": 0.0171,
3339
+ "step": 555
3340
+ },
3341
+ {
3342
+ "epoch": 2.59,
3343
+ "learning_rate": 9.840893644281347e-07,
3344
+ "loss": 0.0173,
3345
+ "step": 556
3346
+ },
3347
+ {
3348
+ "epoch": 2.59,
3349
+ "learning_rate": 9.624590417268475e-07,
3350
+ "loss": 0.0155,
3351
+ "step": 557
3352
+ },
3353
+ {
3354
+ "epoch": 2.6,
3355
+ "learning_rate": 9.410570627470594e-07,
3356
+ "loss": 0.0158,
3357
+ "step": 558
3358
+ },
3359
+ {
3360
+ "epoch": 2.6,
3361
+ "learning_rate": 9.198839682340432e-07,
3362
+ "loss": 0.0174,
3363
+ "step": 559
3364
+ },
3365
+ {
3366
+ "epoch": 2.6,
3367
+ "learning_rate": 8.989402931500434e-07,
3368
+ "loss": 0.0202,
3369
+ "step": 560
3370
+ },
3371
+ {
3372
+ "epoch": 2.61,
3373
+ "learning_rate": 8.782265666607614e-07,
3374
+ "loss": 0.0204,
3375
+ "step": 561
3376
+ },
3377
+ {
3378
+ "epoch": 2.61,
3379
+ "learning_rate": 8.577433121219903e-07,
3380
+ "loss": 0.024,
3381
+ "step": 562
3382
+ },
3383
+ {
3384
+ "epoch": 2.62,
3385
+ "learning_rate": 8.374910470663821e-07,
3386
+ "loss": 0.0172,
3387
+ "step": 563
3388
+ },
3389
+ {
3390
+ "epoch": 2.62,
3391
+ "learning_rate": 8.174702831903836e-07,
3392
+ "loss": 0.0147,
3393
+ "step": 564
3394
+ },
3395
+ {
3396
+ "epoch": 2.63,
3397
+ "learning_rate": 7.976815263412963e-07,
3398
+ "loss": 0.0205,
3399
+ "step": 565
3400
+ },
3401
+ {
3402
+ "epoch": 2.63,
3403
+ "learning_rate": 7.781252765045078e-07,
3404
+ "loss": 0.0201,
3405
+ "step": 566
3406
+ },
3407
+ {
3408
+ "epoch": 2.64,
3409
+ "learning_rate": 7.588020277908426e-07,
3410
+ "loss": 0.018,
3411
+ "step": 567
3412
+ },
3413
+ {
3414
+ "epoch": 2.64,
3415
+ "learning_rate": 7.397122684240998e-07,
3416
+ "loss": 0.0195,
3417
+ "step": 568
3418
+ },
3419
+ {
3420
+ "epoch": 2.65,
3421
+ "learning_rate": 7.208564807286945e-07,
3422
+ "loss": 0.0183,
3423
+ "step": 569
3424
+ },
3425
+ {
3426
+ "epoch": 2.65,
3427
+ "learning_rate": 7.022351411174866e-07,
3428
+ "loss": 0.0202,
3429
+ "step": 570
3430
+ },
3431
+ {
3432
+ "epoch": 2.66,
3433
+ "learning_rate": 6.838487200797406e-07,
3434
+ "loss": 0.0158,
3435
+ "step": 571
3436
+ },
3437
+ {
3438
+ "epoch": 2.66,
3439
+ "learning_rate": 6.656976821692374e-07,
3440
+ "loss": 0.0176,
3441
+ "step": 572
3442
+ },
3443
+ {
3444
+ "epoch": 2.67,
3445
+ "learning_rate": 6.477824859925297e-07,
3446
+ "loss": 0.018,
3447
+ "step": 573
3448
+ },
3449
+ {
3450
+ "epoch": 2.67,
3451
+ "learning_rate": 6.301035841973702e-07,
3452
+ "loss": 0.0239,
3453
+ "step": 574
3454
+ },
3455
+ {
3456
+ "epoch": 2.67,
3457
+ "learning_rate": 6.126614234612593e-07,
3458
+ "loss": 0.0159,
3459
+ "step": 575
3460
+ },
3461
+ {
3462
+ "epoch": 2.68,
3463
+ "learning_rate": 5.954564444801703e-07,
3464
+ "loss": 0.0203,
3465
+ "step": 576
3466
+ },
3467
+ {
3468
+ "epoch": 2.68,
3469
+ "learning_rate": 5.784890819574074e-07,
3470
+ "loss": 0.0207,
3471
+ "step": 577
3472
+ },
3473
+ {
3474
+ "epoch": 2.69,
3475
+ "learning_rate": 5.617597645926265e-07,
3476
+ "loss": 0.0215,
3477
+ "step": 578
3478
+ },
3479
+ {
3480
+ "epoch": 2.69,
3481
+ "learning_rate": 5.452689150710055e-07,
3482
+ "loss": 0.0161,
3483
+ "step": 579
3484
+ },
3485
+ {
3486
+ "epoch": 2.7,
3487
+ "learning_rate": 5.290169500525577e-07,
3488
+ "loss": 0.0204,
3489
+ "step": 580
3490
+ },
3491
+ {
3492
+ "epoch": 2.7,
3493
+ "learning_rate": 5.130042801616109e-07,
3494
+ "loss": 0.0196,
3495
+ "step": 581
3496
+ },
3497
+ {
3498
+ "epoch": 2.71,
3499
+ "learning_rate": 4.97231309976427e-07,
3500
+ "loss": 0.0194,
3501
+ "step": 582
3502
+ },
3503
+ {
3504
+ "epoch": 2.71,
3505
+ "learning_rate": 4.81698438018986e-07,
3506
+ "loss": 0.018,
3507
+ "step": 583
3508
+ },
3509
+ {
3510
+ "epoch": 2.72,
3511
+ "learning_rate": 4.664060567449169e-07,
3512
+ "loss": 0.0177,
3513
+ "step": 584
3514
+ },
3515
+ {
3516
+ "epoch": 2.72,
3517
+ "learning_rate": 4.5135455253357053e-07,
3518
+ "loss": 0.0197,
3519
+ "step": 585
3520
+ },
3521
+ {
3522
+ "epoch": 2.73,
3523
+ "learning_rate": 4.3654430567827367e-07,
3524
+ "loss": 0.0208,
3525
+ "step": 586
3526
+ },
3527
+ {
3528
+ "epoch": 2.73,
3529
+ "learning_rate": 4.2197569037670584e-07,
3530
+ "loss": 0.0151,
3531
+ "step": 587
3532
+ },
3533
+ {
3534
+ "epoch": 2.73,
3535
+ "learning_rate": 4.0764907472145677e-07,
3536
+ "loss": 0.0188,
3537
+ "step": 588
3538
+ },
3539
+ {
3540
+ "epoch": 2.74,
3541
+ "learning_rate": 3.935648206907194e-07,
3542
+ "loss": 0.0181,
3543
+ "step": 589
3544
+ },
3545
+ {
3546
+ "epoch": 2.74,
3547
+ "learning_rate": 3.7972328413914074e-07,
3548
+ "loss": 0.0148,
3549
+ "step": 590
3550
+ },
3551
+ {
3552
+ "epoch": 2.75,
3553
+ "learning_rate": 3.661248147888419e-07,
3554
+ "loss": 0.0173,
3555
+ "step": 591
3556
+ },
3557
+ {
3558
+ "epoch": 2.75,
3559
+ "learning_rate": 3.527697562205723e-07,
3560
+ "loss": 0.0174,
3561
+ "step": 592
3562
+ },
3563
+ {
3564
+ "epoch": 2.76,
3565
+ "learning_rate": 3.3965844586503095e-07,
3566
+ "loss": 0.0155,
3567
+ "step": 593
3568
+ },
3569
+ {
3570
+ "epoch": 2.76,
3571
+ "learning_rate": 3.267912149943431e-07,
3572
+ "loss": 0.0176,
3573
+ "step": 594
3574
+ },
3575
+ {
3576
+ "epoch": 2.77,
3577
+ "learning_rate": 3.1416838871368925e-07,
3578
+ "loss": 0.0157,
3579
+ "step": 595
3580
+ },
3581
+ {
3582
+ "epoch": 2.77,
3583
+ "learning_rate": 3.017902859530919e-07,
3584
+ "loss": 0.0166,
3585
+ "step": 596
3586
+ },
3587
+ {
3588
+ "epoch": 2.78,
3589
+ "learning_rate": 2.896572194593528e-07,
3590
+ "loss": 0.0173,
3591
+ "step": 597
3592
+ },
3593
+ {
3594
+ "epoch": 2.78,
3595
+ "learning_rate": 2.777694957881594e-07,
3596
+ "loss": 0.0173,
3597
+ "step": 598
3598
+ },
3599
+ {
3600
+ "epoch": 2.79,
3601
+ "learning_rate": 2.661274152963278e-07,
3602
+ "loss": 0.0179,
3603
+ "step": 599
3604
+ },
3605
+ {
3606
+ "epoch": 2.79,
3607
+ "learning_rate": 2.547312721342277e-07,
3608
+ "loss": 0.0221,
3609
+ "step": 600
3610
+ },
3611
+ {
3612
+ "epoch": 2.8,
3613
+ "learning_rate": 2.435813542383403e-07,
3614
+ "loss": 0.0207,
3615
+ "step": 601
3616
+ },
3617
+ {
3618
+ "epoch": 2.8,
3619
+ "learning_rate": 2.3267794332398454e-07,
3620
+ "loss": 0.0208,
3621
+ "step": 602
3622
+ },
3623
+ {
3624
+ "epoch": 2.8,
3625
+ "learning_rate": 2.2202131487820244e-07,
3626
+ "loss": 0.0157,
3627
+ "step": 603
3628
+ },
3629
+ {
3630
+ "epoch": 2.81,
3631
+ "learning_rate": 2.1161173815279824e-07,
3632
+ "loss": 0.0195,
3633
+ "step": 604
3634
+ },
3635
+ {
3636
+ "epoch": 2.81,
3637
+ "learning_rate": 2.014494761575314e-07,
3638
+ "loss": 0.0129,
3639
+ "step": 605
3640
+ },
3641
+ {
3642
+ "epoch": 2.82,
3643
+ "learning_rate": 1.9153478565347773e-07,
3644
+ "loss": 0.0157,
3645
+ "step": 606
3646
+ },
3647
+ {
3648
+ "epoch": 2.82,
3649
+ "learning_rate": 1.8186791714653428e-07,
3650
+ "loss": 0.0168,
3651
+ "step": 607
3652
+ },
3653
+ {
3654
+ "epoch": 2.83,
3655
+ "learning_rate": 1.7244911488109916e-07,
3656
+ "loss": 0.0188,
3657
+ "step": 608
3658
+ },
3659
+ {
3660
+ "epoch": 2.83,
3661
+ "learning_rate": 1.6327861683389179e-07,
3662
+ "loss": 0.0182,
3663
+ "step": 609
3664
+ },
3665
+ {
3666
+ "epoch": 2.84,
3667
+ "learning_rate": 1.543566547079467e-07,
3668
+ "loss": 0.0207,
3669
+ "step": 610
3670
+ },
3671
+ {
3672
+ "epoch": 2.84,
3673
+ "learning_rate": 1.4568345392675375e-07,
3674
+ "loss": 0.0173,
3675
+ "step": 611
3676
+ },
3677
+ {
3678
+ "epoch": 2.85,
3679
+ "learning_rate": 1.3725923362856607e-07,
3680
+ "loss": 0.0157,
3681
+ "step": 612
3682
+ },
3683
+ {
3684
+ "epoch": 2.85,
3685
+ "learning_rate": 1.290842066608655e-07,
3686
+ "loss": 0.0188,
3687
+ "step": 613
3688
+ },
3689
+ {
3690
+ "epoch": 2.86,
3691
+ "learning_rate": 1.2115857957497813e-07,
3692
+ "loss": 0.0199,
3693
+ "step": 614
3694
+ },
3695
+ {
3696
+ "epoch": 2.86,
3697
+ "learning_rate": 1.134825526208605e-07,
3698
+ "loss": 0.0182,
3699
+ "step": 615
3700
+ },
3701
+ {
3702
+ "epoch": 2.87,
3703
+ "learning_rate": 1.0605631974203834e-07,
3704
+ "loss": 0.0173,
3705
+ "step": 616
3706
+ },
3707
+ {
3708
+ "epoch": 2.87,
3709
+ "learning_rate": 9.888006857070698e-08,
3710
+ "loss": 0.016,
3711
+ "step": 617
3712
+ },
3713
+ {
3714
+ "epoch": 2.87,
3715
+ "learning_rate": 9.195398042299298e-08,
3716
+ "loss": 0.0161,
3717
+ "step": 618
3718
+ },
3719
+ {
3720
+ "epoch": 2.88,
3721
+ "learning_rate": 8.527823029436444e-08,
3722
+ "loss": 0.0178,
3723
+ "step": 619
3724
+ },
3725
+ {
3726
+ "epoch": 2.88,
3727
+ "learning_rate": 7.885298685522235e-08,
3728
+ "loss": 0.0172,
3729
+ "step": 620
3730
+ },
3731
+ {
3732
+ "epoch": 2.89,
3733
+ "learning_rate": 7.267841244662622e-08,
3734
+ "loss": 0.0173,
3735
+ "step": 621
3736
+ },
3737
+ {
3738
+ "epoch": 2.89,
3739
+ "learning_rate": 6.675466307620282e-08,
3740
+ "loss": 0.0171,
3741
+ "step": 622
3742
+ },
3743
+ {
3744
+ "epoch": 2.9,
3745
+ "learning_rate": 6.108188841419727e-08,
3746
+ "loss": 0.0175,
3747
+ "step": 623
3748
+ },
3749
+ {
3750
+ "epoch": 2.9,
3751
+ "learning_rate": 5.566023178969482e-08,
3752
+ "loss": 0.0269,
3753
+ "step": 624
3754
+ },
3755
+ {
3756
+ "epoch": 2.91,
3757
+ "learning_rate": 5.048983018699827e-08,
3758
+ "loss": 0.0169,
3759
+ "step": 625
3760
+ },
3761
+ {
3762
+ "epoch": 2.91,
3763
+ "learning_rate": 4.557081424216847e-08,
3764
+ "loss": 0.0157,
3765
+ "step": 626
3766
+ },
3767
+ {
3768
+ "epoch": 2.92,
3769
+ "learning_rate": 4.090330823972033e-08,
3770
+ "loss": 0.0157,
3771
+ "step": 627
3772
+ },
3773
+ {
3774
+ "epoch": 2.92,
3775
+ "learning_rate": 3.648743010948863e-08,
3776
+ "loss": 0.0159,
3777
+ "step": 628
3778
+ },
3779
+ {
3780
+ "epoch": 2.93,
3781
+ "learning_rate": 3.232329142363933e-08,
3782
+ "loss": 0.0178,
3783
+ "step": 629
3784
+ },
3785
+ {
3786
+ "epoch": 2.93,
3787
+ "learning_rate": 2.8410997393860663e-08,
3788
+ "loss": 0.0164,
3789
+ "step": 630
3790
+ },
3791
+ {
3792
+ "epoch": 2.93,
3793
+ "learning_rate": 2.4750646868697546e-08,
3794
+ "loss": 0.0187,
3795
+ "step": 631
3796
+ },
3797
+ {
3798
+ "epoch": 2.94,
3799
+ "learning_rate": 2.134233233105798e-08,
3800
+ "loss": 0.0184,
3801
+ "step": 632
3802
+ },
3803
+ {
3804
+ "epoch": 2.94,
3805
+ "learning_rate": 1.81861398958727e-08,
3806
+ "loss": 0.0148,
3807
+ "step": 633
3808
+ },
3809
+ {
3810
+ "epoch": 2.95,
3811
+ "learning_rate": 1.5282149307922488e-08,
3812
+ "loss": 0.018,
3813
+ "step": 634
3814
+ },
3815
+ {
3816
+ "epoch": 2.95,
3817
+ "learning_rate": 1.2630433939825326e-08,
3818
+ "loss": 0.0162,
3819
+ "step": 635
3820
+ },
3821
+ {
3822
+ "epoch": 2.96,
3823
+ "learning_rate": 1.0231060790174552e-08,
3824
+ "loss": 0.0152,
3825
+ "step": 636
3826
+ },
3827
+ {
3828
+ "epoch": 2.96,
3829
+ "learning_rate": 8.084090481855766e-09,
3830
+ "loss": 0.0169,
3831
+ "step": 637
3832
+ },
3833
+ {
3834
+ "epoch": 2.97,
3835
+ "learning_rate": 6.189577260508062e-09,
3836
+ "loss": 0.0168,
3837
+ "step": 638
3838
+ },
3839
+ {
3840
+ "epoch": 2.97,
3841
+ "learning_rate": 4.5475689931551155e-09,
3842
+ "loss": 0.0194,
3843
+ "step": 639
3844
+ },
3845
+ {
3846
+ "epoch": 2.98,
3847
+ "learning_rate": 3.1581071670006013e-09,
3848
+ "loss": 0.0199,
3849
+ "step": 640
3850
+ },
3851
+ {
3852
+ "epoch": 2.98,
3853
+ "learning_rate": 2.0212268883701424e-09,
3854
+ "loss": 0.0169,
3855
+ "step": 641
3856
+ },
3857
+ {
3858
+ "epoch": 2.99,
3859
+ "learning_rate": 1.1369568818364596e-09,
3860
+ "loss": 0.0161,
3861
+ "step": 642
3862
+ },
3863
+ {
3864
+ "epoch": 2.99,
3865
+ "learning_rate": 5.053194894821811e-10,
3866
+ "loss": 0.015,
3867
+ "step": 643
3868
+ },
3869
+ {
3870
+ "epoch": 3.0,
3871
+ "learning_rate": 1.2633067034251156e-10,
3872
+ "loss": 0.0139,
3873
+ "step": 644
3874
+ },
3875
+ {
3876
+ "epoch": 3.0,
3877
+ "learning_rate": 0.0,
3878
+ "loss": 0.0186,
3879
+ "step": 645
3880
+ },
3881
+ {
3882
+ "epoch": 3.0,
3883
+ "step": 645,
3884
+ "total_flos": 0.0,
3885
+ "train_loss": 0.031338086711286114,
3886
+ "train_runtime": 10862.5621,
3887
+ "train_samples_per_second": 6.647,
3888
+ "train_steps_per_second": 0.059
3889
+ }
3890
+ ],
3891
+ "logging_steps": 1.0,
3892
+ "max_steps": 645,
3893
+ "num_input_tokens_seen": 0,
3894
+ "num_train_epochs": 3,
3895
+ "save_steps": 500,
3896
+ "total_flos": 0.0,
3897
+ "train_batch_size": 16,
3898
+ "trial_name": null,
3899
+ "trial_params": null
3900
+ }