Hugging Face's logo

ycchen
/
final-lora-r64-ep5

PEFT
Model card Files Community
ycchen commited on
Commit
79cebda
•
1 Parent(s): 0ca3016

Upload folder using huggingface_hub

Browse files
Files changed (9) hide show
  1. README.md +27 -0
  2. adapter_config.json +219 -0
  3. adapter_model.bin +3 -0
  4. qwen.tiktoken +0 -0
  5. scheduler.pt +3 -0
  6. special_tokens_map.json +3 -0
  7. tokenizer_config.json +11 -0
  8. trainer_state.json +4279 -0
  9. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - quant_method: gptq
9
+ - bits: 8
10
+ - tokenizer: None
11
+ - dataset: None
12
+ - group_size: 128
13
+ - damp_percent: 0.01
14
+ - desc_act: False
15
+ - sym: True
16
+ - true_sequential: True
17
+ - use_cuda_fp16: False
18
+ - model_seqlen: None
19
+ - block_name_to_quantize: None
20
+ - module_name_preceding_first_block: None
21
+ - batch_size: 1
22
+ - pad_token_id: None
23
+ - disable_exllama: False
24
+ ### Framework versions
25
+
26
+
27
+ - PEFT 0.5.0
adapter_config.json ADDED
@@ -0,0 +1,219 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "ycchen/yc-test1",
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": 16,
11
+ "lora_dropout": 0.05,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 16,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "0.attn.c_attn",
18
+ "0.attn.c_proj",
19
+ "0.mlp.c_proj",
20
+ "0.mlp.w1",
21
+ "0.mlp.w2",
22
+ "1.attn.c_attn",
23
+ "1.attn.c_proj",
24
+ "1.mlp.c_proj",
25
+ "1.mlp.w1",
26
+ "1.mlp.w2",
27
+ "2.attn.c_attn",
28
+ "2.attn.c_proj",
29
+ "2.mlp.c_proj",
30
+ "2.mlp.w1",
31
+ "2.mlp.w2",
32
+ "3.attn.c_attn",
33
+ "3.attn.c_proj",
34
+ "3.mlp.c_proj",
35
+ "3.mlp.w1",
36
+ "3.mlp.w2",
37
+ "4.attn.c_attn",
38
+ "4.attn.c_proj",
39
+ "4.mlp.c_proj",
40
+ "4.mlp.w1",
41
+ "4.mlp.w2",
42
+ "5.attn.c_attn",
43
+ "5.attn.c_proj",
44
+ "5.mlp.c_proj",
45
+ "5.mlp.w1",
46
+ "5.mlp.w2",
47
+ "6.attn.c_attn",
48
+ "6.attn.c_proj",
49
+ "6.mlp.c_proj",
50
+ "6.mlp.w1",
51
+ "6.mlp.w2",
52
+ "7.attn.c_attn",
53
+ "7.attn.c_proj",
54
+ "7.mlp.c_proj",
55
+ "7.mlp.w1",
56
+ "7.mlp.w2",
57
+ "8.attn.c_attn",
58
+ "8.attn.c_proj",
59
+ "8.mlp.c_proj",
60
+ "8.mlp.w1",
61
+ "8.mlp.w2",
62
+ "9.attn.c_attn",
63
+ "9.attn.c_proj",
64
+ "9.mlp.c_proj",
65
+ "9.mlp.w1",
66
+ "9.mlp.w2",
67
+ "10.attn.c_attn",
68
+ "10.attn.c_proj",
69
+ "10.mlp.c_proj",
70
+ "10.mlp.w1",
71
+ "10.mlp.w2",
72
+ "11.attn.c_attn",
73
+ "11.attn.c_proj",
74
+ "11.mlp.c_proj",
75
+ "11.mlp.w1",
76
+ "11.mlp.w2",
77
+ "12.attn.c_attn",
78
+ "12.attn.c_proj",
79
+ "12.mlp.c_proj",
80
+ "12.mlp.w1",
81
+ "12.mlp.w2",
82
+ "13.attn.c_attn",
83
+ "13.attn.c_proj",
84
+ "13.mlp.c_proj",
85
+ "13.mlp.w1",
86
+ "13.mlp.w2",
87
+ "14.attn.c_attn",
88
+ "14.attn.c_proj",
89
+ "14.mlp.c_proj",
90
+ "14.mlp.w1",
91
+ "14.mlp.w2",
92
+ "15.attn.c_attn",
93
+ "15.attn.c_proj",
94
+ "15.mlp.c_proj",
95
+ "15.mlp.w1",
96
+ "15.mlp.w2",
97
+ "16.attn.c_attn",
98
+ "16.attn.c_proj",
99
+ "16.mlp.c_proj",
100
+ "16.mlp.w1",
101
+ "16.mlp.w2",
102
+ "17.attn.c_attn",
103
+ "17.attn.c_proj",
104
+ "17.mlp.c_proj",
105
+ "17.mlp.w1",
106
+ "17.mlp.w2",
107
+ "18.attn.c_attn",
108
+ "18.attn.c_proj",
109
+ "18.mlp.c_proj",
110
+ "18.mlp.w1",
111
+ "18.mlp.w2",
112
+ "19.attn.c_attn",
113
+ "19.attn.c_proj",
114
+ "19.mlp.c_proj",
115
+ "19.mlp.w1",
116
+ "19.mlp.w2",
117
+ "20.attn.c_attn",
118
+ "20.attn.c_proj",
119
+ "20.mlp.c_proj",
120
+ "20.mlp.w1",
121
+ "20.mlp.w2",
122
+ "21.attn.c_attn",
123
+ "21.attn.c_proj",
124
+ "21.mlp.c_proj",
125
+ "21.mlp.w1",
126
+ "21.mlp.w2",
127
+ "22.attn.c_attn",
128
+ "22.attn.c_proj",
129
+ "22.mlp.c_proj",
130
+ "22.mlp.w1",
131
+ "22.mlp.w2",
132
+ "23.attn.c_attn",
133
+ "23.attn.c_proj",
134
+ "23.mlp.c_proj",
135
+ "23.mlp.w1",
136
+ "23.mlp.w2",
137
+ "24.attn.c_attn",
138
+ "24.attn.c_proj",
139
+ "24.mlp.c_proj",
140
+ "24.mlp.w1",
141
+ "24.mlp.w2",
142
+ "25.attn.c_attn",
143
+ "25.attn.c_proj",
144
+ "25.mlp.c_proj",
145
+ "25.mlp.w1",
146
+ "25.mlp.w2",
147
+ "26.attn.c_attn",
148
+ "26.attn.c_proj",
149
+ "26.mlp.c_proj",
150
+ "26.mlp.w1",
151
+ "26.mlp.w2",
152
+ "27.attn.c_attn",
153
+ "27.attn.c_proj",
154
+ "27.mlp.c_proj",
155
+ "27.mlp.w1",
156
+ "27.mlp.w2",
157
+ "28.attn.c_attn",
158
+ "28.attn.c_proj",
159
+ "28.mlp.c_proj",
160
+ "28.mlp.w1",
161
+ "28.mlp.w2",
162
+ "29.attn.c_attn",
163
+ "29.attn.c_proj",
164
+ "29.mlp.c_proj",
165
+ "29.mlp.w1",
166
+ "29.mlp.w2",
167
+ "30.attn.c_attn",
168
+ "30.attn.c_proj",
169
+ "30.mlp.c_proj",
170
+ "30.mlp.w1",
171
+ "30.mlp.w2",
172
+ "31.attn.c_attn",
173
+ "31.attn.c_proj",
174
+ "31.mlp.c_proj",
175
+ "31.mlp.w1",
176
+ "31.mlp.w2",
177
+ "32.attn.c_attn",
178
+ "32.attn.c_proj",
179
+ "32.mlp.c_proj",
180
+ "32.mlp.w1",
181
+ "32.mlp.w2",
182
+ "33.attn.c_attn",
183
+ "33.attn.c_proj",
184
+ "33.mlp.c_proj",
185
+ "33.mlp.w1",
186
+ "33.mlp.w2",
187
+ "34.attn.c_attn",
188
+ "34.attn.c_proj",
189
+ "34.mlp.c_proj",
190
+ "34.mlp.w1",
191
+ "34.mlp.w2",
192
+ "35.attn.c_attn",
193
+ "35.attn.c_proj",
194
+ "35.mlp.c_proj",
195
+ "35.mlp.w1",
196
+ "35.mlp.w2",
197
+ "36.attn.c_attn",
198
+ "36.attn.c_proj",
199
+ "36.mlp.c_proj",
200
+ "36.mlp.w1",
201
+ "36.mlp.w2",
202
+ "37.attn.c_attn",
203
+ "37.attn.c_proj",
204
+ "37.mlp.c_proj",
205
+ "37.mlp.w1",
206
+ "37.mlp.w2",
207
+ "38.attn.c_attn",
208
+ "38.attn.c_proj",
209
+ "38.mlp.c_proj",
210
+ "38.mlp.w1",
211
+ "38.mlp.w2",
212
+ "39.attn.c_attn",
213
+ "39.attn.c_proj",
214
+ "39.mlp.c_proj",
215
+ "39.mlp.w1",
216
+ "39.mlp.w2"
217
+ ],
218
+ "task_type": "CAUSAL_LM"
219
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b1db46e9e5e1f8ca3d95147c751bf32ac97db994ddba0da5a3ee0522478375b
3
+ size 223292429
qwen.tiktoken ADDED
The diff for this file is too large to render. See raw diff
 
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34b6e9e4a8c949ba17d310417b25873e8ff48ceb195f75c447f0bc01856532d8
3
+ size 563
special_tokens_map.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "pad_token": "<|extra_21|>"
3
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_map": {
3
+ "AutoTokenizer": [
4
+ "Qwen/Qwen-14B--tokenization_qwen.QWenTokenizer",
5
+ null
6
+ ]
7
+ },
8
+ "clean_up_tokenization_spaces": true,
9
+ "model_max_length": 8192,
10
+ "tokenizer_class": "QWenTokenizer"
11
+ }
trainer_state.json ADDED
@@ -0,0 +1,4279 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 5.0,
5
+ "eval_steps": 500,
6
+ "global_step": 710,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.01,
13
+ "learning_rate": 2.3505215219626858e-06,
14
+ "loss": 1.5844,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 4.7010430439253716e-06,
20
+ "loss": 1.1717,
21
+ "step": 2
22
+ },
23
+ {
24
+ "epoch": 0.02,
25
+ "learning_rate": 7.051564565888058e-06,
26
+ "loss": 1.1708,
27
+ "step": 3
28
+ },
29
+ {
30
+ "epoch": 0.03,
31
+ "learning_rate": 9.402086087850743e-06,
32
+ "loss": 1.5208,
33
+ "step": 4
34
+ },
35
+ {
36
+ "epoch": 0.04,
37
+ "learning_rate": 1.175260760981343e-05,
38
+ "loss": 1.2781,
39
+ "step": 5
40
+ },
41
+ {
42
+ "epoch": 0.04,
43
+ "learning_rate": 1.4103129131776116e-05,
44
+ "loss": 1.3885,
45
+ "step": 6
46
+ },
47
+ {
48
+ "epoch": 0.05,
49
+ "learning_rate": 1.64536506537388e-05,
50
+ "loss": 1.4596,
51
+ "step": 7
52
+ },
53
+ {
54
+ "epoch": 0.06,
55
+ "learning_rate": 1.8804172175701486e-05,
56
+ "loss": 1.2032,
57
+ "step": 8
58
+ },
59
+ {
60
+ "epoch": 0.06,
61
+ "learning_rate": 2.1154693697664172e-05,
62
+ "loss": 1.3789,
63
+ "step": 9
64
+ },
65
+ {
66
+ "epoch": 0.07,
67
+ "learning_rate": 2.350521521962686e-05,
68
+ "loss": 1.1589,
69
+ "step": 10
70
+ },
71
+ {
72
+ "epoch": 0.08,
73
+ "learning_rate": 2.585573674158954e-05,
74
+ "loss": 1.3547,
75
+ "step": 11
76
+ },
77
+ {
78
+ "epoch": 0.08,
79
+ "learning_rate": 2.820625826355223e-05,
80
+ "loss": 1.6605,
81
+ "step": 12
82
+ },
83
+ {
84
+ "epoch": 0.09,
85
+ "learning_rate": 3.055677978551492e-05,
86
+ "loss": 1.2712,
87
+ "step": 13
88
+ },
89
+ {
90
+ "epoch": 0.1,
91
+ "learning_rate": 3.29073013074776e-05,
92
+ "loss": 1.2314,
93
+ "step": 14
94
+ },
95
+ {
96
+ "epoch": 0.11,
97
+ "learning_rate": 3.525782282944028e-05,
98
+ "loss": 1.3365,
99
+ "step": 15
100
+ },
101
+ {
102
+ "epoch": 0.11,
103
+ "learning_rate": 3.760834435140297e-05,
104
+ "loss": 1.3434,
105
+ "step": 16
106
+ },
107
+ {
108
+ "epoch": 0.12,
109
+ "learning_rate": 3.9958865873365655e-05,
110
+ "loss": 1.4921,
111
+ "step": 17
112
+ },
113
+ {
114
+ "epoch": 0.13,
115
+ "learning_rate": 4.2309387395328345e-05,
116
+ "loss": 1.3429,
117
+ "step": 18
118
+ },
119
+ {
120
+ "epoch": 0.13,
121
+ "learning_rate": 4.4659908917291035e-05,
122
+ "loss": 1.3414,
123
+ "step": 19
124
+ },
125
+ {
126
+ "epoch": 0.14,
127
+ "learning_rate": 4.701043043925372e-05,
128
+ "loss": 1.3829,
129
+ "step": 20
130
+ },
131
+ {
132
+ "epoch": 0.15,
133
+ "learning_rate": 4.93609519612164e-05,
134
+ "loss": 1.0409,
135
+ "step": 21
136
+ },
137
+ {
138
+ "epoch": 0.15,
139
+ "learning_rate": 5.171147348317908e-05,
140
+ "loss": 1.4251,
141
+ "step": 22
142
+ },
143
+ {
144
+ "epoch": 0.16,
145
+ "learning_rate": 5.406199500514177e-05,
146
+ "loss": 1.3312,
147
+ "step": 23
148
+ },
149
+ {
150
+ "epoch": 0.17,
151
+ "learning_rate": 5.641251652710446e-05,
152
+ "loss": 1.5272,
153
+ "step": 24
154
+ },
155
+ {
156
+ "epoch": 0.18,
157
+ "learning_rate": 5.8763038049067145e-05,
158
+ "loss": 1.3403,
159
+ "step": 25
160
+ },
161
+ {
162
+ "epoch": 0.18,
163
+ "learning_rate": 6.111355957102983e-05,
164
+ "loss": 1.2074,
165
+ "step": 26
166
+ },
167
+ {
168
+ "epoch": 0.19,
169
+ "learning_rate": 6.346408109299252e-05,
170
+ "loss": 1.2367,
171
+ "step": 27
172
+ },
173
+ {
174
+ "epoch": 0.2,
175
+ "learning_rate": 6.58146026149552e-05,
176
+ "loss": 1.0457,
177
+ "step": 28
178
+ },
179
+ {
180
+ "epoch": 0.2,
181
+ "learning_rate": 6.816512413691789e-05,
182
+ "loss": 1.2032,
183
+ "step": 29
184
+ },
185
+ {
186
+ "epoch": 0.21,
187
+ "learning_rate": 7.051564565888057e-05,
188
+ "loss": 1.2936,
189
+ "step": 30
190
+ },
191
+ {
192
+ "epoch": 0.22,
193
+ "learning_rate": 7.286616718084326e-05,
194
+ "loss": 1.4325,
195
+ "step": 31
196
+ },
197
+ {
198
+ "epoch": 0.23,
199
+ "learning_rate": 7.521668870280595e-05,
200
+ "loss": 1.1163,
201
+ "step": 32
202
+ },
203
+ {
204
+ "epoch": 0.23,
205
+ "learning_rate": 7.756721022476862e-05,
206
+ "loss": 1.2707,
207
+ "step": 33
208
+ },
209
+ {
210
+ "epoch": 0.24,
211
+ "learning_rate": 7.991773174673131e-05,
212
+ "loss": 1.2504,
213
+ "step": 34
214
+ },
215
+ {
216
+ "epoch": 0.25,
217
+ "learning_rate": 8.2268253268694e-05,
218
+ "loss": 1.3381,
219
+ "step": 35
220
+ },
221
+ {
222
+ "epoch": 0.25,
223
+ "learning_rate": 8.461877479065669e-05,
224
+ "loss": 1.4586,
225
+ "step": 36
226
+ },
227
+ {
228
+ "epoch": 0.26,
229
+ "learning_rate": 8.696929631261938e-05,
230
+ "loss": 1.3332,
231
+ "step": 37
232
+ },
233
+ {
234
+ "epoch": 0.27,
235
+ "learning_rate": 8.931981783458207e-05,
236
+ "loss": 1.2071,
237
+ "step": 38
238
+ },
239
+ {
240
+ "epoch": 0.27,
241
+ "learning_rate": 9.167033935654475e-05,
242
+ "loss": 1.2606,
243
+ "step": 39
244
+ },
245
+ {
246
+ "epoch": 0.28,
247
+ "learning_rate": 9.402086087850743e-05,
248
+ "loss": 1.2386,
249
+ "step": 40
250
+ },
251
+ {
252
+ "epoch": 0.29,
253
+ "learning_rate": 9.637138240047012e-05,
254
+ "loss": 1.163,
255
+ "step": 41
256
+ },
257
+ {
258
+ "epoch": 0.3,
259
+ "learning_rate": 9.87219039224328e-05,
260
+ "loss": 1.2538,
261
+ "step": 42
262
+ },
263
+ {
264
+ "epoch": 0.3,
265
+ "learning_rate": 9.999988438329202e-05,
266
+ "loss": 1.1698,
267
+ "step": 43
268
+ },
269
+ {
270
+ "epoch": 0.31,
271
+ "learning_rate": 9.999882216302589e-05,
272
+ "loss": 1.3658,
273
+ "step": 44
274
+ },
275
+ {
276
+ "epoch": 0.32,
277
+ "learning_rate": 9.999664914761532e-05,
278
+ "loss": 1.0131,
279
+ "step": 45
280
+ },
281
+ {
282
+ "epoch": 0.32,
283
+ "learning_rate": 9.999336538533694e-05,
284
+ "loss": 1.4838,
285
+ "step": 46
286
+ },
287
+ {
288
+ "epoch": 0.33,
289
+ "learning_rate": 9.998897094914424e-05,
290
+ "loss": 1.1761,
291
+ "step": 47
292
+ },
293
+ {
294
+ "epoch": 0.34,
295
+ "learning_rate": 9.998346593666587e-05,
296
+ "loss": 1.4351,
297
+ "step": 48
298
+ },
299
+ {
300
+ "epoch": 0.35,
301
+ "learning_rate": 9.997685047020353e-05,
302
+ "loss": 1.3387,
303
+ "step": 49
304
+ },
305
+ {
306
+ "epoch": 0.35,
307
+ "learning_rate": 9.996912469672925e-05,
308
+ "loss": 1.2005,
309
+ "step": 50
310
+ },
311
+ {
312
+ "epoch": 0.36,
313
+ "learning_rate": 9.99602887878821e-05,
314
+ "loss": 1.3915,
315
+ "step": 51
316
+ },
317
+ {
318
+ "epoch": 0.37,
319
+ "learning_rate": 9.995034293996442e-05,
320
+ "loss": 1.2307,
321
+ "step": 52
322
+ },
323
+ {
324
+ "epoch": 0.37,
325
+ "learning_rate": 9.993928737393738e-05,
326
+ "loss": 1.1234,
327
+ "step": 53
328
+ },
329
+ {
330
+ "epoch": 0.38,
331
+ "learning_rate": 9.992712233541616e-05,
332
+ "loss": 1.2151,
333
+ "step": 54
334
+ },
335
+ {
336
+ "epoch": 0.39,
337
+ "learning_rate": 9.991384809466445e-05,
338
+ "loss": 1.0929,
339
+ "step": 55
340
+ },
341
+ {
342
+ "epoch": 0.39,
343
+ "learning_rate": 9.989946494658839e-05,
344
+ "loss": 1.4025,
345
+ "step": 56
346
+ },
347
+ {
348
+ "epoch": 0.4,
349
+ "learning_rate": 9.98839732107302e-05,
350
+ "loss": 1.1941,
351
+ "step": 57
352
+ },
353
+ {
354
+ "epoch": 0.41,
355
+ "learning_rate": 9.986737323126084e-05,
356
+ "loss": 1.2325,
357
+ "step": 58
358
+ },
359
+ {
360
+ "epoch": 0.42,
361
+ "learning_rate": 9.984966537697254e-05,
362
+ "loss": 1.2219,
363
+ "step": 59
364
+ },
365
+ {
366
+ "epoch": 0.42,
367
+ "learning_rate": 9.983085004127055e-05,
368
+ "loss": 1.256,
369
+ "step": 60
370
+ },
371
+ {
372
+ "epoch": 0.43,
373
+ "learning_rate": 9.981092764216437e-05,
374
+ "loss": 1.3723,
375
+ "step": 61
376
+ },
377
+ {
378
+ "epoch": 0.44,
379
+ "learning_rate": 9.978989862225853e-05,
380
+ "loss": 1.1166,
381
+ "step": 62
382
+ },
383
+ {
384
+ "epoch": 0.44,
385
+ "learning_rate": 9.976776344874268e-05,
386
+ "loss": 1.2826,
387
+ "step": 63
388
+ },
389
+ {
390
+ "epoch": 0.45,
391
+ "learning_rate": 9.974452261338128e-05,
392
+ "loss": 1.1347,
393
+ "step": 64
394
+ },
395
+ {
396
+ "epoch": 0.46,
397
+ "learning_rate": 9.972017663250264e-05,
398
+ "loss": 1.1066,
399
+ "step": 65
400
+ },
401
+ {
402
+ "epoch": 0.46,
403
+ "learning_rate": 9.969472604698742e-05,
404
+ "loss": 1.3199,
405
+ "step": 66
406
+ },
407
+ {
408
+ "epoch": 0.47,
409
+ "learning_rate": 9.96681714222567e-05,
410
+ "loss": 1.3186,
411
+ "step": 67
412
+ },
413
+ {
414
+ "epoch": 0.48,
415
+ "learning_rate": 9.964051334825937e-05,
416
+ "loss": 1.3577,
417
+ "step": 68
418
+ },
419
+ {
420
+ "epoch": 0.49,
421
+ "learning_rate": 9.961175243945893e-05,
422
+ "loss": 1.1913,
423
+ "step": 69
424
+ },
425
+ {
426
+ "epoch": 0.49,
427
+ "learning_rate": 9.958188933482003e-05,
428
+ "loss": 1.3416,
429
+ "step": 70
430
+ },
431
+ {
432
+ "epoch": 0.5,
433
+ "learning_rate": 9.95509246977941e-05,
434
+ "loss": 1.4361,
435
+ "step": 71
436
+ },
437
+ {
438
+ "epoch": 0.51,
439
+ "learning_rate": 9.951885921630475e-05,
440
+ "loss": 1.2034,
441
+ "step": 72
442
+ },
443
+ {
444
+ "epoch": 0.51,
445
+ "learning_rate": 9.948569360273236e-05,
446
+ "loss": 1.147,
447
+ "step": 73
448
+ },
449
+ {
450
+ "epoch": 0.52,
451
+ "learning_rate": 9.945142859389835e-05,
452
+ "loss": 1.5043,
453
+ "step": 74
454
+ },
455
+ {
456
+ "epoch": 0.53,
457
+ "learning_rate": 9.941606495104877e-05,
458
+ "loss": 1.2877,
459
+ "step": 75
460
+ },
461
+ {
462
+ "epoch": 0.54,
463
+ "learning_rate": 9.937960345983736e-05,
464
+ "loss": 1.2024,
465
+ "step": 76
466
+ },
467
+ {
468
+ "epoch": 0.54,
469
+ "learning_rate": 9.934204493030816e-05,
470
+ "loss": 1.2718,
471
+ "step": 77
472
+ },
473
+ {
474
+ "epoch": 0.55,
475
+ "learning_rate": 9.930339019687746e-05,
476
+ "loss": 1.1279,
477
+ "step": 78
478
+ },
479
+ {
480
+ "epoch": 0.56,
481
+ "learning_rate": 9.92636401183153e-05,
482
+ "loss": 1.1138,
483
+ "step": 79
484
+ },
485
+ {
486
+ "epoch": 0.56,
487
+ "learning_rate": 9.922279557772639e-05,
488
+ "loss": 1.2727,
489
+ "step": 80
490
+ },
491
+ {
492
+ "epoch": 0.57,
493
+ "learning_rate": 9.918085748253044e-05,
494
+ "loss": 1.1073,
495
+ "step": 81
496
+ },
497
+ {
498
+ "epoch": 0.58,
499
+ "learning_rate": 9.9137826764442e-05,
500
+ "loss": 1.1801,
501
+ "step": 82
502
+ },
503
+ {
504
+ "epoch": 0.58,
505
+ "learning_rate": 9.909370437944993e-05,
506
+ "loss": 1.4271,
507
+ "step": 83
508
+ },
509
+ {
510
+ "epoch": 0.59,
511
+ "learning_rate": 9.904849130779588e-05,
512
+ "loss": 1.45,
513
+ "step": 84
514
+ },
515
+ {
516
+ "epoch": 0.6,
517
+ "learning_rate": 9.900218855395276e-05,
518
+ "loss": 1.3001,
519
+ "step": 85
520
+ },
521
+ {
522
+ "epoch": 0.61,
523
+ "learning_rate": 9.895479714660226e-05,
524
+ "loss": 1.3452,
525
+ "step": 86
526
+ },
527
+ {
528
+ "epoch": 0.61,
529
+ "learning_rate": 9.890631813861207e-05,
530
+ "loss": 1.2141,
531
+ "step": 87
532
+ },
533
+ {
534
+ "epoch": 0.62,
535
+ "learning_rate": 9.885675260701254e-05,
536
+ "loss": 1.213,
537
+ "step": 88
538
+ },
539
+ {
540
+ "epoch": 0.63,
541
+ "learning_rate": 9.880610165297262e-05,
542
+ "loss": 1.3351,
543
+ "step": 89
544
+ },
545
+ {
546
+ "epoch": 0.63,
547
+ "learning_rate": 9.87543664017755e-05,
548
+ "loss": 1.063,
549
+ "step": 90
550
+ },
551
+ {
552
+ "epoch": 0.64,
553
+ "learning_rate": 9.870154800279354e-05,
554
+ "loss": 1.0482,
555
+ "step": 91
556
+ },
557
+ {
558
+ "epoch": 0.65,
559
+ "learning_rate": 9.864764762946285e-05,
560
+ "loss": 1.3549,
561
+ "step": 92
562
+ },
563
+ {
564
+ "epoch": 0.65,
565
+ "learning_rate": 9.859266647925709e-05,
566
+ "loss": 1.2104,
567
+ "step": 93
568
+ },
569
+ {
570
+ "epoch": 0.66,
571
+ "learning_rate": 9.853660577366092e-05,
572
+ "loss": 1.3416,
573
+ "step": 94
574
+ },
575
+ {
576
+ "epoch": 0.67,
577
+ "learning_rate": 9.847946675814286e-05,
578
+ "loss": 1.0623,
579
+ "step": 95
580
+ },
581
+ {
582
+ "epoch": 0.68,
583
+ "learning_rate": 9.842125070212765e-05,
584
+ "loss": 1.0585,
585
+ "step": 96
586
+ },
587
+ {
588
+ "epoch": 0.68,
589
+ "learning_rate": 9.8361958898968e-05,
590
+ "loss": 1.123,
591
+ "step": 97
592
+ },
593
+ {
594
+ "epoch": 0.69,
595
+ "learning_rate": 9.830159266591589e-05,
596
+ "loss": 1.2343,
597
+ "step": 98
598
+ },
599
+ {
600
+ "epoch": 0.7,
601
+ "learning_rate": 9.824015334409325e-05,
602
+ "loss": 1.2386,
603
+ "step": 99
604
+ },
605
+ {
606
+ "epoch": 0.7,
607
+ "learning_rate": 9.817764229846228e-05,
608
+ "loss": 0.9396,
609
+ "step": 100
610
+ },
611
+ {
612
+ "epoch": 0.71,
613
+ "learning_rate": 9.811406091779501e-05,
614
+ "loss": 1.0644,
615
+ "step": 101
616
+ },
617
+ {
618
+ "epoch": 0.72,
619
+ "learning_rate": 9.804941061464246e-05,
620
+ "loss": 1.2164,
621
+ "step": 102
622
+ },
623
+ {
624
+ "epoch": 0.73,
625
+ "learning_rate": 9.798369282530334e-05,
626
+ "loss": 1.2597,
627
+ "step": 103
628
+ },
629
+ {
630
+ "epoch": 0.73,
631
+ "learning_rate": 9.79169090097921e-05,
632
+ "loss": 1.0657,
633
+ "step": 104
634
+ },
635
+ {
636
+ "epoch": 0.74,
637
+ "learning_rate": 9.784906065180639e-05,
638
+ "loss": 1.3115,
639
+ "step": 105
640
+ },
641
+ {
642
+ "epoch": 0.75,
643
+ "learning_rate": 9.77801492586943e-05,
644
+ "loss": 1.1609,
645
+ "step": 106
646
+ },
647
+ {
648
+ "epoch": 0.75,
649
+ "learning_rate": 9.771017636142069e-05,
650
+ "loss": 1.2352,
651
+ "step": 107
652
+ },
653
+ {
654
+ "epoch": 0.76,
655
+ "learning_rate": 9.763914351453328e-05,
656
+ "loss": 1.2551,
657
+ "step": 108
658
+ },
659
+ {
660
+ "epoch": 0.77,
661
+ "learning_rate": 9.756705229612806e-05,
662
+ "loss": 1.1841,
663
+ "step": 109
664
+ },
665
+ {
666
+ "epoch": 0.77,
667
+ "learning_rate": 9.749390430781428e-05,
668
+ "loss": 1.4215,
669
+ "step": 110
670
+ },
671
+ {
672
+ "epoch": 0.78,
673
+ "learning_rate": 9.741970117467885e-05,
674
+ "loss": 1.3392,
675
+ "step": 111
676
+ },
677
+ {
678
+ "epoch": 0.79,
679
+ "learning_rate": 9.734444454525015e-05,
680
+ "loss": 1.2668,
681
+ "step": 112
682
+ },
683
+ {
684
+ "epoch": 0.8,
685
+ "learning_rate": 9.726813609146156e-05,
686
+ "loss": 1.2107,
687
+ "step": 113
688
+ },
689
+ {
690
+ "epoch": 0.8,
691
+ "learning_rate": 9.719077750861422e-05,
692
+ "loss": 1.0586,
693
+ "step": 114
694
+ },
695
+ {
696
+ "epoch": 0.81,
697
+ "learning_rate": 9.711237051533939e-05,
698
+ "loss": 1.2463,
699
+ "step": 115
700
+ },
701
+ {
702
+ "epoch": 0.82,
703
+ "learning_rate": 9.703291685356022e-05,
704
+ "loss": 1.0728,
705
+ "step": 116
706
+ },
707
+ {
708
+ "epoch": 0.82,
709
+ "learning_rate": 9.695241828845314e-05,
710
+ "loss": 1.0965,
711
+ "step": 117
712
+ },
713
+ {
714
+ "epoch": 0.83,
715
+ "learning_rate": 9.68708766084086e-05,
716
+ "loss": 1.4591,
717
+ "step": 118
718
+ },
719
+ {
720
+ "epoch": 0.84,
721
+ "learning_rate": 9.678829362499131e-05,
722
+ "loss": 1.0267,
723
+ "step": 119
724
+ },
725
+ {
726
+ "epoch": 0.85,
727
+ "learning_rate": 9.670467117290003e-05,
728
+ "loss": 1.2758,
729
+ "step": 120
730
+ },
731
+ {
732
+ "epoch": 0.85,
733
+ "learning_rate": 9.662001110992679e-05,
734
+ "loss": 1.2481,
735
+ "step": 121
736
+ },
737
+ {
738
+ "epoch": 0.86,
739
+ "learning_rate": 9.653431531691564e-05,
740
+ "loss": 1.1977,
741
+ "step": 122
742
+ },
743
+ {
744
+ "epoch": 0.87,
745
+ "learning_rate": 9.644758569772084e-05,
746
+ "loss": 1.4683,
747
+ "step": 123
748
+ },
749
+ {
750
+ "epoch": 0.87,
751
+ "learning_rate": 9.635982417916459e-05,
752
+ "loss": 1.2774,
753
+ "step": 124
754
+ },
755
+ {
756
+ "epoch": 0.88,
757
+ "learning_rate": 9.627103271099418e-05,
758
+ "loss": 1.1035,
759
+ "step": 125
760
+ },
761
+ {
762
+ "epoch": 0.89,
763
+ "learning_rate": 9.618121326583874e-05,
764
+ "loss": 1.1808,
765
+ "step": 126
766
+ },
767
+ {
768
+ "epoch": 0.89,
769
+ "learning_rate": 9.609036783916529e-05,
770
+ "loss": 1.1988,
771
+ "step": 127
772
+ },
773
+ {
774
+ "epoch": 0.9,
775
+ "learning_rate": 9.59984984492346e-05,
776
+ "loss": 1.3405,
777
+ "step": 128
778
+ },
779
+ {
780
+ "epoch": 0.91,
781
+ "learning_rate": 9.590560713705619e-05,
782
+ "loss": 1.31,
783
+ "step": 129
784
+ },
785
+ {
786
+ "epoch": 0.92,
787
+ "learning_rate": 9.581169596634302e-05,
788
+ "loss": 1.2651,
789
+ "step": 130
790
+ },
791
+ {
792
+ "epoch": 0.92,
793
+ "learning_rate": 9.571676702346571e-05,
794
+ "loss": 1.2317,
795
+ "step": 131
796
+ },
797
+ {
798
+ "epoch": 0.93,
799
+ "learning_rate": 9.56208224174061e-05,
800
+ "loss": 1.2728,
801
+ "step": 132
802
+ },
803
+ {
804
+ "epoch": 0.94,
805
+ "learning_rate": 9.552386427971049e-05,
806
+ "loss": 1.2959,
807
+ "step": 133
808
+ },
809
+ {
810
+ "epoch": 0.94,
811
+ "learning_rate": 9.542589476444215e-05,
812
+ "loss": 1.2706,
813
+ "step": 134
814
+ },
815
+ {
816
+ "epoch": 0.95,
817
+ "learning_rate": 9.53269160481336e-05,
818
+ "loss": 1.1602,
819
+ "step": 135
820
+ },
821
+ {
822
+ "epoch": 0.96,
823
+ "learning_rate": 9.52269303297382e-05,
824
+ "loss": 1.2293,
825
+ "step": 136
826
+ },
827
+ {
828
+ "epoch": 0.96,
829
+ "learning_rate": 9.512593983058129e-05,
830
+ "loss": 1.2822,
831
+ "step": 137
832
+ },
833
+ {
834
+ "epoch": 0.97,
835
+ "learning_rate": 9.502394679431084e-05,
836
+ "loss": 1.0499,
837
+ "step": 138
838
+ },
839
+ {
840
+ "epoch": 0.98,
841
+ "learning_rate": 9.49209534868476e-05,
842
+ "loss": 1.3138,
843
+ "step": 139
844
+ },
845
+ {
846
+ "epoch": 0.99,
847
+ "learning_rate": 9.481696219633481e-05,
848
+ "loss": 1.4097,
849
+ "step": 140
850
+ },
851
+ {
852
+ "epoch": 0.99,
853
+ "learning_rate": 9.47119752330873e-05,
854
+ "loss": 1.1332,
855
+ "step": 141
856
+ },
857
+ {
858
+ "epoch": 1.0,
859
+ "learning_rate": 9.460599492954018e-05,
860
+ "loss": 1.1116,
861
+ "step": 142
862
+ },
863
+ {
864
+ "epoch": 1.01,
865
+ "learning_rate": 9.449902364019708e-05,
866
+ "loss": 1.3985,
867
+ "step": 143
868
+ },
869
+ {
870
+ "epoch": 1.01,
871
+ "learning_rate": 9.439106374157773e-05,
872
+ "loss": 1.0012,
873
+ "step": 144
874
+ },
875
+ {
876
+ "epoch": 1.02,
877
+ "learning_rate": 9.428211763216526e-05,
878
+ "loss": 0.9964,
879
+ "step": 145
880
+ },
881
+ {
882
+ "epoch": 1.03,
883
+ "learning_rate": 9.417218773235286e-05,
884
+ "loss": 1.339,
885
+ "step": 146
886
+ },
887
+ {
888
+ "epoch": 1.04,
889
+ "learning_rate": 9.406127648439008e-05,
890
+ "loss": 1.0616,
891
+ "step": 147
892
+ },
893
+ {
894
+ "epoch": 1.04,
895
+ "learning_rate": 9.394938635232844e-05,
896
+ "loss": 1.2071,
897
+ "step": 148
898
+ },
899
+ {
900
+ "epoch": 1.05,
901
+ "learning_rate": 9.383651982196682e-05,
902
+ "loss": 1.289,
903
+ "step": 149
904
+ },
905
+ {
906
+ "epoch": 1.06,
907
+ "learning_rate": 9.372267940079617e-05,
908
+ "loss": 1.0109,
909
+ "step": 150
910
+ },
911
+ {
912
+ "epoch": 1.06,
913
+ "learning_rate": 9.36078676179438e-05,
914
+ "loss": 1.1774,
915
+ "step": 151
916
+ },
917
+ {
918
+ "epoch": 1.07,
919
+ "learning_rate": 9.349208702411722e-05,
920
+ "loss": 0.9881,
921
+ "step": 152
922
+ },
923
+ {
924
+ "epoch": 1.08,
925
+ "learning_rate": 9.337534019154745e-05,
926
+ "loss": 1.1558,
927
+ "step": 153
928
+ },
929
+ {
930
+ "epoch": 1.08,
931
+ "learning_rate": 9.32576297139319e-05,
932
+ "loss": 1.468,
933
+ "step": 154
934
+ },
935
+ {
936
+ "epoch": 1.09,
937
+ "learning_rate": 9.313895820637669e-05,
938
+ "loss": 1.1095,
939
+ "step": 155
940
+ },
941
+ {
942
+ "epoch": 1.1,
943
+ "learning_rate": 9.301932830533862e-05,
944
+ "loss": 1.0915,
945
+ "step": 156
946
+ },
947
+ {
948
+ "epoch": 1.11,
949
+ "learning_rate": 9.289874266856655e-05,
950
+ "loss": 1.1427,
951
+ "step": 157
952
+ },
953
+ {
954
+ "epoch": 1.11,
955
+ "learning_rate": 9.277720397504243e-05,
956
+ "loss": 1.2019,
957
+ "step": 158
958
+ },
959
+ {
960
+ "epoch": 1.12,
961
+ "learning_rate": 9.265471492492163e-05,
962
+ "loss": 1.3101,
963
+ "step": 159
964
+ },
965
+ {
966
+ "epoch": 1.13,
967
+ "learning_rate": 9.253127823947311e-05,
968
+ "loss": 1.1997,
969
+ "step": 160
970
+ },
971
+ {
972
+ "epoch": 1.13,
973
+ "learning_rate": 9.240689666101892e-05,
974
+ "loss": 1.1938,
975
+ "step": 161
976
+ },
977
+ {
978
+ "epoch": 1.14,
979
+ "learning_rate": 9.228157295287317e-05,
980
+ "loss": 1.2199,
981
+ "step": 162
982
+ },
983
+ {
984
+ "epoch": 1.15,
985
+ "learning_rate": 9.215530989928081e-05,
986
+ "loss": 0.9386,
987
+ "step": 163
988
+ },
989
+ {
990
+ "epoch": 1.15,
991
+ "learning_rate": 9.202811030535565e-05,
992
+ "loss": 1.2752,
993
+ "step": 164
994
+ },
995
+ {
996
+ "epoch": 1.16,
997
+ "learning_rate": 9.189997699701807e-05,
998
+ "loss": 1.2073,
999
+ "step": 165
1000
+ },
1001
+ {
1002
+ "epoch": 1.17,
1003
+ "learning_rate": 9.177091282093231e-05,
1004
+ "loss": 1.3951,
1005
+ "step": 166
1006
+ },
1007
+ {
1008
+ "epoch": 1.18,
1009
+ "learning_rate": 9.164092064444306e-05,
1010
+ "loss": 1.1749,
1011
+ "step": 167
1012
+ },
1013
+ {
1014
+ "epoch": 1.18,
1015
+ "learning_rate": 9.151000335551194e-05,
1016
+ "loss": 1.0863,
1017
+ "step": 168
1018
+ },
1019
+ {
1020
+ "epoch": 1.19,
1021
+ "learning_rate": 9.137816386265327e-05,
1022
+ "loss": 1.1223,
1023
+ "step": 169
1024
+ },
1025
+ {
1026
+ "epoch": 1.2,
1027
+ "learning_rate": 9.124540509486939e-05,
1028
+ "loss": 0.948,
1029
+ "step": 170
1030
+ },
1031
+ {
1032
+ "epoch": 1.2,
1033
+ "learning_rate": 9.111173000158566e-05,
1034
+ "loss": 1.1039,
1035
+ "step": 171
1036
+ },
1037
+ {
1038
+ "epoch": 1.21,
1039
+ "learning_rate": 9.097714155258497e-05,
1040
+ "loss": 1.1733,
1041
+ "step": 172
1042
+ },
1043
+ {
1044
+ "epoch": 1.22,
1045
+ "learning_rate": 9.084164273794163e-05,
1046
+ "loss": 1.3096,
1047
+ "step": 173
1048
+ },
1049
+ {
1050
+ "epoch": 1.23,
1051
+ "learning_rate": 9.070523656795508e-05,
1052
+ "loss": 0.995,
1053
+ "step": 174
1054
+ },
1055
+ {
1056
+ "epoch": 1.23,
1057
+ "learning_rate": 9.056792607308293e-05,
1058
+ "loss": 1.1716,
1059
+ "step": 175
1060
+ },
1061
+ {
1062
+ "epoch": 1.24,
1063
+ "learning_rate": 9.042971430387366e-05,
1064
+ "loss": 1.1152,
1065
+ "step": 176
1066
+ },
1067
+ {
1068
+ "epoch": 1.25,
1069
+ "learning_rate": 9.029060433089885e-05,
1070
+ "loss": 1.2403,
1071
+ "step": 177
1072
+ },
1073
+ {
1074
+ "epoch": 1.25,
1075
+ "learning_rate": 9.015059924468494e-05,
1076
+ "loss": 1.3455,
1077
+ "step": 178
1078
+ },
1079
+ {
1080
+ "epoch": 1.26,
1081
+ "learning_rate": 9.000970215564462e-05,
1082
+ "loss": 1.1923,
1083
+ "step": 179
1084
+ },
1085
+ {
1086
+ "epoch": 1.27,
1087
+ "learning_rate": 8.986791619400768e-05,
1088
+ "loss": 1.1189,
1089
+ "step": 180
1090
+ },
1091
+ {
1092
+ "epoch": 1.27,
1093
+ "learning_rate": 8.972524450975143e-05,
1094
+ "loss": 1.166,
1095
+ "step": 181
1096
+ },
1097
+ {
1098
+ "epoch": 1.28,
1099
+ "learning_rate": 8.958169027253087e-05,
1100
+ "loss": 1.1154,
1101
+ "step": 182
1102
+ },
1103
+ {
1104
+ "epoch": 1.29,
1105
+ "learning_rate": 8.94372566716081e-05,
1106
+ "loss": 1.0369,
1107
+ "step": 183
1108
+ },
1109
+ {
1110
+ "epoch": 1.3,
1111
+ "learning_rate": 8.929194691578158e-05,
1112
+ "loss": 1.1153,
1113
+ "step": 184
1114
+ },
1115
+ {
1116
+ "epoch": 1.3,
1117
+ "learning_rate": 8.914576423331475e-05,
1118
+ "loss": 1.0748,
1119
+ "step": 185
1120
+ },
1121
+ {
1122
+ "epoch": 1.31,
1123
+ "learning_rate": 8.89987118718644e-05,
1124
+ "loss": 1.2589,
1125
+ "step": 186
1126
+ },
1127
+ {
1128
+ "epoch": 1.32,
1129
+ "learning_rate": 8.885079309840848e-05,
1130
+ "loss": 0.9202,
1131
+ "step": 187
1132
+ },
1133
+ {
1134
+ "epoch": 1.32,
1135
+ "learning_rate": 8.870201119917354e-05,
1136
+ "loss": 1.3762,
1137
+ "step": 188
1138
+ },
1139
+ {
1140
+ "epoch": 1.33,
1141
+ "learning_rate": 8.855236947956161e-05,
1142
+ "loss": 1.0431,
1143
+ "step": 189
1144
+ },
1145
+ {
1146
+ "epoch": 1.34,
1147
+ "learning_rate": 8.840187126407695e-05,
1148
+ "loss": 1.2946,
1149
+ "step": 190
1150
+ },
1151
+ {
1152
+ "epoch": 1.35,
1153
+ "learning_rate": 8.825051989625206e-05,
1154
+ "loss": 1.1731,
1155
+ "step": 191
1156
+ },
1157
+ {
1158
+ "epoch": 1.35,
1159
+ "learning_rate": 8.809831873857344e-05,
1160
+ "loss": 1.0521,
1161
+ "step": 192
1162
+ },
1163
+ {
1164
+ "epoch": 1.36,
1165
+ "learning_rate": 8.794527117240688e-05,
1166
+ "loss": 1.2539,
1167
+ "step": 193
1168
+ },
1169
+ {
1170
+ "epoch": 1.37,
1171
+ "learning_rate": 8.779138059792232e-05,
1172
+ "loss": 1.1325,
1173
+ "step": 194
1174
+ },
1175
+ {
1176
+ "epoch": 1.37,
1177
+ "learning_rate": 8.76366504340184e-05,
1178
+ "loss": 0.9917,
1179
+ "step": 195
1180
+ },
1181
+ {
1182
+ "epoch": 1.38,
1183
+ "learning_rate": 8.748108411824636e-05,
1184
+ "loss": 1.0688,
1185
+ "step": 196
1186
+ },
1187
+ {
1188
+ "epoch": 1.39,
1189
+ "learning_rate": 8.732468510673379e-05,
1190
+ "loss": 0.882,
1191
+ "step": 197
1192
+ },
1193
+ {
1194
+ "epoch": 1.39,
1195
+ "learning_rate": 8.716745687410779e-05,
1196
+ "loss": 1.2479,
1197
+ "step": 198
1198
+ },
1199
+ {
1200
+ "epoch": 1.4,
1201
+ "learning_rate": 8.700940291341782e-05,
1202
+ "loss": 1.0746,
1203
+ "step": 199
1204
+ },
1205
+ {
1206
+ "epoch": 1.41,
1207
+ "learning_rate": 8.685052673605798e-05,
1208
+ "loss": 1.0906,
1209
+ "step": 200
1210
+ },
1211
+ {
1212
+ "epoch": 1.42,
1213
+ "learning_rate": 8.66908318716892e-05,
1214
+ "loss": 1.0399,
1215
+ "step": 201
1216
+ },
1217
+ {
1218
+ "epoch": 1.42,
1219
+ "learning_rate": 8.653032186816062e-05,
1220
+ "loss": 1.1061,
1221
+ "step": 202
1222
+ },
1223
+ {
1224
+ "epoch": 1.43,
1225
+ "learning_rate": 8.636900029143091e-05,
1226
+ "loss": 1.2413,
1227
+ "step": 203
1228
+ },
1229
+ {
1230
+ "epoch": 1.44,
1231
+ "learning_rate": 8.620687072548899e-05,
1232
+ "loss": 0.9968,
1233
+ "step": 204
1234
+ },
1235
+ {
1236
+ "epoch": 1.44,
1237
+ "learning_rate": 8.604393677227436e-05,
1238
+ "loss": 1.0685,
1239
+ "step": 205
1240
+ },
1241
+ {
1242
+ "epoch": 1.45,
1243
+ "learning_rate": 8.588020205159724e-05,
1244
+ "loss": 1.0171,
1245
+ "step": 206
1246
+ },
1247
+ {
1248
+ "epoch": 1.46,
1249
+ "learning_rate": 8.57156702010579e-05,
1250
+ "loss": 0.9696,
1251
+ "step": 207
1252
+ },
1253
+ {
1254
+ "epoch": 1.46,
1255
+ "learning_rate": 8.555034487596612e-05,
1256
+ "loss": 1.1729,
1257
+ "step": 208
1258
+ },
1259
+ {
1260
+ "epoch": 1.47,
1261
+ "learning_rate": 8.538422974925975e-05,
1262
+ "loss": 1.1588,
1263
+ "step": 209
1264
+ },
1265
+ {
1266
+ "epoch": 1.48,
1267
+ "learning_rate": 8.521732851142327e-05,
1268
+ "loss": 1.2008,
1269
+ "step": 210
1270
+ },
1271
+ {
1272
+ "epoch": 1.49,
1273
+ "learning_rate": 8.50496448704057e-05,
1274
+ "loss": 1.0552,
1275
+ "step": 211
1276
+ },
1277
+ {
1278
+ "epoch": 1.49,
1279
+ "learning_rate": 8.488118255153831e-05,
1280
+ "loss": 1.1742,
1281
+ "step": 212
1282
+ },
1283
+ {
1284
+ "epoch": 1.5,
1285
+ "learning_rate": 8.471194529745176e-05,
1286
+ "loss": 1.2478,
1287
+ "step": 213
1288
+ },
1289
+ {
1290
+ "epoch": 1.51,
1291
+ "learning_rate": 8.454193686799306e-05,
1292
+ "loss": 1.0792,
1293
+ "step": 214
1294
+ },
1295
+ {
1296
+ "epoch": 1.51,
1297
+ "learning_rate": 8.437116104014187e-05,
1298
+ "loss": 0.9578,
1299
+ "step": 215
1300
+ },
1301
+ {
1302
+ "epoch": 1.52,
1303
+ "learning_rate": 8.419962160792683e-05,
1304
+ "loss": 1.3535,
1305
+ "step": 216
1306
+ },
1307
+ {
1308
+ "epoch": 1.53,
1309
+ "learning_rate": 8.402732238234105e-05,
1310
+ "loss": 1.0785,
1311
+ "step": 217
1312
+ },
1313
+ {
1314
+ "epoch": 1.54,
1315
+ "learning_rate": 8.385426719125759e-05,
1316
+ "loss": 1.0145,
1317
+ "step": 218
1318
+ },
1319
+ {
1320
+ "epoch": 1.54,
1321
+ "learning_rate": 8.368045987934432e-05,
1322
+ "loss": 1.1375,
1323
+ "step": 219
1324
+ },
1325
+ {
1326
+ "epoch": 1.55,
1327
+ "learning_rate": 8.350590430797858e-05,
1328
+ "loss": 0.976,
1329
+ "step": 220
1330
+ },
1331
+ {
1332
+ "epoch": 1.56,
1333
+ "learning_rate": 8.333060435516134e-05,
1334
+ "loss": 0.9385,
1335
+ "step": 221
1336
+ },
1337
+ {
1338
+ "epoch": 1.56,
1339
+ "learning_rate": 8.315456391543106e-05,
1340
+ "loss": 1.1096,
1341
+ "step": 222
1342
+ },
1343
+ {
1344
+ "epoch": 1.57,
1345
+ "learning_rate": 8.297778689977719e-05,
1346
+ "loss": 0.9635,
1347
+ "step": 223
1348
+ },
1349
+ {
1350
+ "epoch": 1.58,
1351
+ "learning_rate": 8.280027723555327e-05,
1352
+ "loss": 1.0106,
1353
+ "step": 224
1354
+ },
1355
+ {
1356
+ "epoch": 1.58,
1357
+ "learning_rate": 8.262203886638964e-05,
1358
+ "loss": 1.2063,
1359
+ "step": 225
1360
+ },
1361
+ {
1362
+ "epoch": 1.59,
1363
+ "learning_rate": 8.24430757521059e-05,
1364
+ "loss": 1.2897,
1365
+ "step": 226
1366
+ },
1367
+ {
1368
+ "epoch": 1.6,
1369
+ "learning_rate": 8.226339186862286e-05,
1370
+ "loss": 1.1298,
1371
+ "step": 227
1372
+ },
1373
+ {
1374
+ "epoch": 1.61,
1375
+ "learning_rate": 8.208299120787429e-05,
1376
+ "loss": 1.1559,
1377
+ "step": 228
1378
+ },
1379
+ {
1380
+ "epoch": 1.61,
1381
+ "learning_rate": 8.190187777771812e-05,
1382
+ "loss": 1.0587,
1383
+ "step": 229
1384
+ },
1385
+ {
1386
+ "epoch": 1.62,
1387
+ "learning_rate": 8.172005560184755e-05,
1388
+ "loss": 1.0188,
1389
+ "step": 230
1390
+ },
1391
+ {
1392
+ "epoch": 1.63,
1393
+ "learning_rate": 8.153752871970151e-05,
1394
+ "loss": 1.1215,
1395
+ "step": 231
1396
+ },
1397
+ {
1398
+ "epoch": 1.63,
1399
+ "learning_rate": 8.135430118637504e-05,
1400
+ "loss": 0.9491,
1401
+ "step": 232
1402
+ },
1403
+ {
1404
+ "epoch": 1.64,
1405
+ "learning_rate": 8.117037707252911e-05,
1406
+ "loss": 0.9292,
1407
+ "step": 233
1408
+ },
1409
+ {
1410
+ "epoch": 1.65,
1411
+ "learning_rate": 8.098576046430023e-05,
1412
+ "loss": 1.1974,
1413
+ "step": 234
1414
+ },
1415
+ {
1416
+ "epoch": 1.65,
1417
+ "learning_rate": 8.080045546320963e-05,
1418
+ "loss": 1.0007,
1419
+ "step": 235
1420
+ },
1421
+ {
1422
+ "epoch": 1.66,
1423
+ "learning_rate": 8.061446618607222e-05,
1424
+ "loss": 1.1207,
1425
+ "step": 236
1426
+ },
1427
+ {
1428
+ "epoch": 1.67,
1429
+ "learning_rate": 8.042779676490506e-05,
1430
+ "loss": 0.8835,
1431
+ "step": 237
1432
+ },
1433
+ {
1434
+ "epoch": 1.68,
1435
+ "learning_rate": 8.024045134683559e-05,
1436
+ "loss": 0.8961,
1437
+ "step": 238
1438
+ },
1439
+ {
1440
+ "epoch": 1.68,
1441
+ "learning_rate": 8.005243409400945e-05,
1442
+ "loss": 0.9905,
1443
+ "step": 239
1444
+ },
1445
+ {
1446
+ "epoch": 1.69,
1447
+ "learning_rate": 7.986374918349804e-05,
1448
+ "loss": 1.0613,
1449
+ "step": 240
1450
+ },
1451
+ {
1452
+ "epoch": 1.7,
1453
+ "learning_rate": 7.967440080720581e-05,
1454
+ "loss": 1.0781,
1455
+ "step": 241
1456
+ },
1457
+ {
1458
+ "epoch": 1.7,
1459
+ "learning_rate": 7.948439317177696e-05,
1460
+ "loss": 0.7686,
1461
+ "step": 242
1462
+ },
1463
+ {
1464
+ "epoch": 1.71,
1465
+ "learning_rate": 7.929373049850214e-05,
1466
+ "loss": 0.941,
1467
+ "step": 243
1468
+ },
1469
+ {
1470
+ "epoch": 1.72,
1471
+ "learning_rate": 7.910241702322452e-05,
1472
+ "loss": 1.0858,
1473
+ "step": 244
1474
+ },
1475
+ {
1476
+ "epoch": 1.73,
1477
+ "learning_rate": 7.891045699624586e-05,
1478
+ "loss": 1.0956,
1479
+ "step": 245
1480
+ },
1481
+ {
1482
+ "epoch": 1.73,
1483
+ "learning_rate": 7.87178546822319e-05,
1484
+ "loss": 0.9023,
1485
+ "step": 246
1486
+ },
1487
+ {
1488
+ "epoch": 1.74,
1489
+ "learning_rate": 7.852461436011777e-05,
1490
+ "loss": 1.1251,
1491
+ "step": 247
1492
+ },
1493
+ {
1494
+ "epoch": 1.75,
1495
+ "learning_rate": 7.833074032301279e-05,
1496
+ "loss": 1.0181,
1497
+ "step": 248
1498
+ },
1499
+ {
1500
+ "epoch": 1.75,
1501
+ "learning_rate": 7.813623687810528e-05,
1502
+ "loss": 1.0848,
1503
+ "step": 249
1504
+ },
1505
+ {
1506
+ "epoch": 1.76,
1507
+ "learning_rate": 7.79411083465666e-05,
1508
+ "loss": 1.0787,
1509
+ "step": 250
1510
+ },
1511
+ {
1512
+ "epoch": 1.77,
1513
+ "learning_rate": 7.774535906345541e-05,
1514
+ "loss": 1.0356,
1515
+ "step": 251
1516
+ },
1517
+ {
1518
+ "epoch": 1.77,
1519
+ "learning_rate": 7.754899337762122e-05,
1520
+ "loss": 1.1425,
1521
+ "step": 252
1522
+ },
1523
+ {
1524
+ "epoch": 1.78,
1525
+ "learning_rate": 7.73520156516078e-05,
1526
+ "loss": 1.1107,
1527
+ "step": 253
1528
+ },
1529
+ {
1530
+ "epoch": 1.79,
1531
+ "learning_rate": 7.715443026155628e-05,
1532
+ "loss": 1.1205,
1533
+ "step": 254
1534
+ },
1535
+ {
1536
+ "epoch": 1.8,
1537
+ "learning_rate": 7.695624159710788e-05,
1538
+ "loss": 1.0453,
1539
+ "step": 255
1540
+ },
1541
+ {
1542
+ "epoch": 1.8,
1543
+ "learning_rate": 7.675745406130647e-05,
1544
+ "loss": 0.8335,
1545
+ "step": 256
1546
+ },
1547
+ {
1548
+ "epoch": 1.81,
1549
+ "learning_rate": 7.655807207050066e-05,
1550
+ "loss": 1.0367,
1551
+ "step": 257
1552
+ },
1553
+ {
1554
+ "epoch": 1.82,
1555
+ "learning_rate": 7.635810005424575e-05,
1556
+ "loss": 0.9137,
1557
+ "step": 258
1558
+ },
1559
+ {
1560
+ "epoch": 1.82,
1561
+ "learning_rate": 7.615754245520526e-05,
1562
+ "loss": 0.9815,
1563
+ "step": 259
1564
+ },
1565
+ {
1566
+ "epoch": 1.83,
1567
+ "learning_rate": 7.595640372905235e-05,
1568
+ "loss": 1.1748,
1569
+ "step": 260
1570
+ },
1571
+ {
1572
+ "epoch": 1.84,
1573
+ "learning_rate": 7.575468834437065e-05,
1574
+ "loss": 0.8648,
1575
+ "step": 261
1576
+ },
1577
+ {
1578
+ "epoch": 1.85,
1579
+ "learning_rate": 7.555240078255512e-05,
1580
+ "loss": 1.0501,
1581
+ "step": 262
1582
+ },
1583
+ {
1584
+ "epoch": 1.85,
1585
+ "learning_rate": 7.53495455377125e-05,
1586
+ "loss": 1.0069,
1587
+ "step": 263
1588
+ },
1589
+ {
1590
+ "epoch": 1.86,
1591
+ "learning_rate": 7.514612711656132e-05,
1592
+ "loss": 1.0022,
1593
+ "step": 264
1594
+ },
1595
+ {
1596
+ "epoch": 1.87,
1597
+ "learning_rate": 7.494215003833197e-05,
1598
+ "loss": 1.2586,
1599
+ "step": 265
1600
+ },
1601
+ {
1602
+ "epoch": 1.87,
1603
+ "learning_rate": 7.473761883466614e-05,
1604
+ "loss": 1.1238,
1605
+ "step": 266
1606
+ },
1607
+ {
1608
+ "epoch": 1.88,
1609
+ "learning_rate": 7.453253804951623e-05,
1610
+ "loss": 0.9192,
1611
+ "step": 267
1612
+ },
1613
+ {
1614
+ "epoch": 1.89,
1615
+ "learning_rate": 7.432691223904439e-05,
1616
+ "loss": 1.0445,
1617
+ "step": 268
1618
+ },
1619
+ {
1620
+ "epoch": 1.89,
1621
+ "learning_rate": 7.412074597152125e-05,
1622
+ "loss": 1.0032,
1623
+ "step": 269
1624
+ },
1625
+ {
1626
+ "epoch": 1.9,
1627
+ "learning_rate": 7.39140438272245e-05,
1628
+ "loss": 1.147,
1629
+ "step": 270
1630
+ },
1631
+ {
1632
+ "epoch": 1.91,
1633
+ "learning_rate": 7.370681039833706e-05,
1634
+ "loss": 1.1687,
1635
+ "step": 271
1636
+ },
1637
+ {
1638
+ "epoch": 1.92,
1639
+ "learning_rate": 7.349905028884516e-05,
1640
+ "loss": 1.1281,
1641
+ "step": 272
1642
+ },
1643
+ {
1644
+ "epoch": 1.92,
1645
+ "learning_rate": 7.329076811443589e-05,
1646
+ "loss": 1.0926,
1647
+ "step": 273
1648
+ },
1649
+ {
1650
+ "epoch": 1.93,
1651
+ "learning_rate": 7.308196850239487e-05,
1652
+ "loss": 1.0286,
1653
+ "step": 274
1654
+ },
1655
+ {
1656
+ "epoch": 1.94,
1657
+ "learning_rate": 7.287265609150325e-05,
1658
+ "loss": 1.1251,
1659
+ "step": 275
1660
+ },
1661
+ {
1662
+ "epoch": 1.94,
1663
+ "learning_rate": 7.266283553193475e-05,
1664
+ "loss": 1.0484,
1665
+ "step": 276
1666
+ },
1667
+ {
1668
+ "epoch": 1.95,
1669
+ "learning_rate": 7.245251148515237e-05,
1670
+ "loss": 0.9548,
1671
+ "step": 277
1672
+ },
1673
+ {
1674
+ "epoch": 1.96,
1675
+ "learning_rate": 7.224168862380482e-05,
1676
+ "loss": 1.0533,
1677
+ "step": 278
1678
+ },
1679
+ {
1680
+ "epoch": 1.96,
1681
+ "learning_rate": 7.203037163162258e-05,
1682
+ "loss": 1.1724,
1683
+ "step": 279
1684
+ },
1685
+ {
1686
+ "epoch": 1.97,
1687
+ "learning_rate": 7.181856520331407e-05,
1688
+ "loss": 0.8868,
1689
+ "step": 280
1690
+ },
1691
+ {
1692
+ "epoch": 1.98,
1693
+ "learning_rate": 7.160627404446115e-05,
1694
+ "loss": 1.1498,
1695
+ "step": 281
1696
+ },
1697
+ {
1698
+ "epoch": 1.99,
1699
+ "learning_rate": 7.13935028714147e-05,
1700
+ "loss": 1.2512,
1701
+ "step": 282
1702
+ },
1703
+ {
1704
+ "epoch": 1.99,
1705
+ "learning_rate": 7.118025641118981e-05,
1706
+ "loss": 0.9648,
1707
+ "step": 283
1708
+ },
1709
+ {
1710
+ "epoch": 2.0,
1711
+ "learning_rate": 7.096653940136069e-05,
1712
+ "loss": 0.8932,
1713
+ "step": 284
1714
+ },
1715
+ {
1716
+ "epoch": 2.01,
1717
+ "learning_rate": 7.075235658995558e-05,
1718
+ "loss": 1.1335,
1719
+ "step": 285
1720
+ },
1721
+ {
1722
+ "epoch": 2.01,
1723
+ "learning_rate": 7.053771273535108e-05,
1724
+ "loss": 0.8604,
1725
+ "step": 286
1726
+ },
1727
+ {
1728
+ "epoch": 2.02,
1729
+ "learning_rate": 7.032261260616654e-05,
1730
+ "loss": 0.858,
1731
+ "step": 287
1732
+ },
1733
+ {
1734
+ "epoch": 2.03,
1735
+ "learning_rate": 7.010706098115812e-05,
1736
+ "loss": 1.152,
1737
+ "step": 288
1738
+ },
1739
+ {
1740
+ "epoch": 2.04,
1741
+ "learning_rate": 6.989106264911262e-05,
1742
+ "loss": 0.8461,
1743
+ "step": 289
1744
+ },
1745
+ {
1746
+ "epoch": 2.04,
1747
+ "learning_rate": 6.967462240874104e-05,
1748
+ "loss": 1.0286,
1749
+ "step": 290
1750
+ },
1751
+ {
1752
+ "epoch": 2.05,
1753
+ "learning_rate": 6.9457745068572e-05,
1754
+ "loss": 1.1287,
1755
+ "step": 291
1756
+ },
1757
+ {
1758
+ "epoch": 2.06,
1759
+ "learning_rate": 6.924043544684494e-05,
1760
+ "loss": 0.828,
1761
+ "step": 292
1762
+ },
1763
+ {
1764
+ "epoch": 2.06,
1765
+ "learning_rate": 6.902269837140304e-05,
1766
+ "loss": 0.9707,
1767
+ "step": 293
1768
+ },
1769
+ {
1770
+ "epoch": 2.07,
1771
+ "learning_rate": 6.880453867958599e-05,
1772
+ "loss": 0.8406,
1773
+ "step": 294
1774
+ },
1775
+ {
1776
+ "epoch": 2.08,
1777
+ "learning_rate": 6.858596121812247e-05,
1778
+ "loss": 0.9251,
1779
+ "step": 295
1780
+ },
1781
+ {
1782
+ "epoch": 2.08,
1783
+ "learning_rate": 6.836697084302253e-05,
1784
+ "loss": 1.1901,
1785
+ "step": 296
1786
+ },
1787
+ {
1788
+ "epoch": 2.09,
1789
+ "learning_rate": 6.814757241946969e-05,
1790
+ "loss": 0.9653,
1791
+ "step": 297
1792
+ },
1793
+ {
1794
+ "epoch": 2.1,
1795
+ "learning_rate": 6.792777082171286e-05,
1796
+ "loss": 0.9213,
1797
+ "step": 298
1798
+ },
1799
+ {
1800
+ "epoch": 2.11,
1801
+ "learning_rate": 6.770757093295798e-05,
1802
+ "loss": 0.8884,
1803
+ "step": 299
1804
+ },
1805
+ {
1806
+ "epoch": 2.11,
1807
+ "learning_rate": 6.74869776452597e-05,
1808
+ "loss": 1.0057,
1809
+ "step": 300
1810
+ },
1811
+ {
1812
+ "epoch": 2.12,
1813
+ "learning_rate": 6.726599585941247e-05,
1814
+ "loss": 1.0599,
1815
+ "step": 301
1816
+ },
1817
+ {
1818
+ "epoch": 2.13,
1819
+ "learning_rate": 6.704463048484184e-05,
1820
+ "loss": 1.0271,
1821
+ "step": 302
1822
+ },
1823
+ {
1824
+ "epoch": 2.13,
1825
+ "learning_rate": 6.682288643949534e-05,
1826
+ "loss": 0.9883,
1827
+ "step": 303
1828
+ },
1829
+ {
1830
+ "epoch": 2.14,
1831
+ "learning_rate": 6.66007686497332e-05,
1832
+ "loss": 1.0209,
1833
+ "step": 304
1834
+ },
1835
+ {
1836
+ "epoch": 2.15,
1837
+ "learning_rate": 6.637828205021891e-05,
1838
+ "loss": 0.8159,
1839
+ "step": 305
1840
+ },
1841
+ {
1842
+ "epoch": 2.15,
1843
+ "learning_rate": 6.615543158380956e-05,
1844
+ "loss": 1.0447,
1845
+ "step": 306
1846
+ },
1847
+ {
1848
+ "epoch": 2.16,
1849
+ "learning_rate": 6.593222220144617e-05,
1850
+ "loss": 1.0394,
1851
+ "step": 307
1852
+ },
1853
+ {
1854
+ "epoch": 2.17,
1855
+ "learning_rate": 6.570865886204347e-05,
1856
+ "loss": 1.146,
1857
+ "step": 308
1858
+ },
1859
+ {
1860
+ "epoch": 2.18,
1861
+ "learning_rate": 6.54847465323799e-05,
1862
+ "loss": 0.9289,
1863
+ "step": 309
1864
+ },
1865
+ {
1866
+ "epoch": 2.18,
1867
+ "learning_rate": 6.526049018698721e-05,
1868
+ "loss": 0.8888,
1869
+ "step": 310
1870
+ },
1871
+ {
1872
+ "epoch": 2.19,
1873
+ "learning_rate": 6.503589480803998e-05,
1874
+ "loss": 0.9346,
1875
+ "step": 311
1876
+ },
1877
+ {
1878
+ "epoch": 2.2,
1879
+ "learning_rate": 6.481096538524487e-05,
1880
+ "loss": 0.7988,
1881
+ "step": 312
1882
+ },
1883
+ {
1884
+ "epoch": 2.2,
1885
+ "learning_rate": 6.458570691572979e-05,
1886
+ "loss": 0.9219,
1887
+ "step": 313
1888
+ },
1889
+ {
1890
+ "epoch": 2.21,
1891
+ "learning_rate": 6.436012440393294e-05,
1892
+ "loss": 0.9391,
1893
+ "step": 314
1894
+ },
1895
+ {
1896
+ "epoch": 2.22,
1897
+ "learning_rate": 6.413422286149154e-05,
1898
+ "loss": 1.0681,
1899
+ "step": 315
1900
+ },
1901
+ {
1902
+ "epoch": 2.23,
1903
+ "learning_rate": 6.390800730713055e-05,
1904
+ "loss": 0.7624,
1905
+ "step": 316
1906
+ },
1907
+ {
1908
+ "epoch": 2.23,
1909
+ "learning_rate": 6.368148276655113e-05,
1910
+ "loss": 0.9216,
1911
+ "step": 317
1912
+ },
1913
+ {
1914
+ "epoch": 2.24,
1915
+ "learning_rate": 6.345465427231904e-05,
1916
+ "loss": 0.8041,
1917
+ "step": 318
1918
+ },
1919
+ {
1920
+ "epoch": 2.25,
1921
+ "learning_rate": 6.32275268637528e-05,
1922
+ "loss": 1.0059,
1923
+ "step": 319
1924
+ },
1925
+ {
1926
+ "epoch": 2.25,
1927
+ "learning_rate": 6.300010558681171e-05,
1928
+ "loss": 1.0607,
1929
+ "step": 320
1930
+ },
1931
+ {
1932
+ "epoch": 2.26,
1933
+ "learning_rate": 6.27723954939838e-05,
1934
+ "loss": 0.8981,
1935
+ "step": 321
1936
+ },
1937
+ {
1938
+ "epoch": 2.27,
1939
+ "learning_rate": 6.254440164417356e-05,
1940
+ "loss": 0.9132,
1941
+ "step": 322
1942
+ },
1943
+ {
1944
+ "epoch": 2.27,
1945
+ "learning_rate": 6.231612910258952e-05,
1946
+ "loss": 0.9141,
1947
+ "step": 323
1948
+ },
1949
+ {
1950
+ "epoch": 2.28,
1951
+ "learning_rate": 6.208758294063176e-05,
1952
+ "loss": 0.8164,
1953
+ "step": 324
1954
+ },
1955
+ {
1956
+ "epoch": 2.29,
1957
+ "learning_rate": 6.185876823577923e-05,
1958
+ "loss": 0.8132,
1959
+ "step": 325
1960
+ },
1961
+ {
1962
+ "epoch": 2.3,
1963
+ "learning_rate": 6.162969007147697e-05,
1964
+ "loss": 0.8809,
1965
+ "step": 326
1966
+ },
1967
+ {
1968
+ "epoch": 2.3,
1969
+ "learning_rate": 6.14003535370231e-05,
1970
+ "loss": 0.8769,
1971
+ "step": 327
1972
+ },
1973
+ {
1974
+ "epoch": 2.31,
1975
+ "learning_rate": 6.117076372745579e-05,
1976
+ "loss": 1.0372,
1977
+ "step": 328
1978
+ },
1979
+ {
1980
+ "epoch": 2.32,
1981
+ "learning_rate": 6.094092574344017e-05,
1982
+ "loss": 0.7325,
1983
+ "step": 329
1984
+ },
1985
+ {
1986
+ "epoch": 2.32,
1987
+ "learning_rate": 6.0710844691154824e-05,
1988
+ "loss": 1.1814,
1989
+ "step": 330
1990
+ },
1991
+ {
1992
+ "epoch": 2.33,
1993
+ "learning_rate": 6.0480525682178455e-05,
1994
+ "loss": 0.8201,
1995
+ "step": 331
1996
+ },
1997
+ {
1998
+ "epoch": 2.34,
1999
+ "learning_rate": 6.024997383337637e-05,
2000
+ "loss": 1.0517,
2001
+ "step": 332
2002
+ },
2003
+ {
2004
+ "epoch": 2.35,
2005
+ "learning_rate": 6.0019194266786715e-05,
2006
+ "loss": 0.8776,
2007
+ "step": 333
2008
+ },
2009
+ {
2010
+ "epoch": 2.35,
2011
+ "learning_rate": 5.97881921095067e-05,
2012
+ "loss": 0.8195,
2013
+ "step": 334
2014
+ },
2015
+ {
2016
+ "epoch": 2.36,
2017
+ "learning_rate": 5.955697249357869e-05,
2018
+ "loss": 0.9821,
2019
+ "step": 335
2020
+ },
2021
+ {
2022
+ "epoch": 2.37,
2023
+ "learning_rate": 5.932554055587624e-05,
2024
+ "loss": 0.9625,
2025
+ "step": 336
2026
+ },
2027
+ {
2028
+ "epoch": 2.37,
2029
+ "learning_rate": 5.909390143798995e-05,
2030
+ "loss": 0.7567,
2031
+ "step": 337
2032
+ },
2033
+ {
2034
+ "epoch": 2.38,
2035
+ "learning_rate": 5.8862060286113165e-05,
2036
+ "loss": 0.8004,
2037
+ "step": 338
2038
+ },
2039
+ {
2040
+ "epoch": 2.39,
2041
+ "learning_rate": 5.8630022250927716e-05,
2042
+ "loss": 0.5626,
2043
+ "step": 339
2044
+ },
2045
+ {
2046
+ "epoch": 2.39,
2047
+ "learning_rate": 5.839779248748954e-05,
2048
+ "loss": 1.015,
2049
+ "step": 340
2050
+ },
2051
+ {
2052
+ "epoch": 2.4,
2053
+ "learning_rate": 5.816537615511404e-05,
2054
+ "loss": 0.8443,
2055
+ "step": 341
2056
+ },
2057
+ {
2058
+ "epoch": 2.41,
2059
+ "learning_rate": 5.79327784172615e-05,
2060
+ "loss": 0.8472,
2061
+ "step": 342
2062
+ },
2063
+ {
2064
+ "epoch": 2.42,
2065
+ "learning_rate": 5.7700004441422395e-05,
2066
+ "loss": 0.7691,
2067
+ "step": 343
2068
+ },
2069
+ {
2070
+ "epoch": 2.42,
2071
+ "learning_rate": 5.746705939900262e-05,
2072
+ "loss": 0.8565,
2073
+ "step": 344
2074
+ },
2075
+ {
2076
+ "epoch": 2.43,
2077
+ "learning_rate": 5.723394846520852e-05,
2078
+ "loss": 1.0142,
2079
+ "step": 345
2080
+ },
2081
+ {
2082
+ "epoch": 2.44,
2083
+ "learning_rate": 5.7000676818931944e-05,
2084
+ "loss": 0.7965,
2085
+ "step": 346
2086
+ },
2087
+ {
2088
+ "epoch": 2.44,
2089
+ "learning_rate": 5.676724964263521e-05,
2090
+ "loss": 0.8013,
2091
+ "step": 347
2092
+ },
2093
+ {
2094
+ "epoch": 2.45,
2095
+ "learning_rate": 5.653367212223599e-05,
2096
+ "loss": 0.8015,
2097
+ "step": 348
2098
+ },
2099
+ {
2100
+ "epoch": 2.46,
2101
+ "learning_rate": 5.6299949446992026e-05,
2102
+ "loss": 0.742,
2103
+ "step": 349
2104
+ },
2105
+ {
2106
+ "epoch": 2.46,
2107
+ "learning_rate": 5.6066086809385873e-05,
2108
+ "loss": 0.9467,
2109
+ "step": 350
2110
+ },
2111
+ {
2112
+ "epoch": 2.47,
2113
+ "learning_rate": 5.583208940500959e-05,
2114
+ "loss": 0.8903,
2115
+ "step": 351
2116
+ },
2117
+ {
2118
+ "epoch": 2.48,
2119
+ "learning_rate": 5.559796243244926e-05,
2120
+ "loss": 0.9451,
2121
+ "step": 352
2122
+ },
2123
+ {
2124
+ "epoch": 2.49,
2125
+ "learning_rate": 5.536371109316947e-05,
2126
+ "loss": 0.8239,
2127
+ "step": 353
2128
+ },
2129
+ {
2130
+ "epoch": 2.49,
2131
+ "learning_rate": 5.5129340591397804e-05,
2132
+ "loss": 0.9108,
2133
+ "step": 354
2134
+ },
2135
+ {
2136
+ "epoch": 2.5,
2137
+ "learning_rate": 5.489485613400929e-05,
2138
+ "loss": 0.9606,
2139
+ "step": 355
2140
+ },
2141
+ {
2142
+ "epoch": 2.51,
2143
+ "learning_rate": 5.466026293041053e-05,
2144
+ "loss": 0.8744,
2145
+ "step": 356
2146
+ },
2147
+ {
2148
+ "epoch": 2.51,
2149
+ "learning_rate": 5.442556619242414e-05,
2150
+ "loss": 0.7085,
2151
+ "step": 357
2152
+ },
2153
+ {
2154
+ "epoch": 2.52,
2155
+ "learning_rate": 5.419077113417288e-05,
2156
+ "loss": 1.0677,
2157
+ "step": 358
2158
+ },
2159
+ {
2160
+ "epoch": 2.53,
2161
+ "learning_rate": 5.3955882971963856e-05,
2162
+ "loss": 0.8227,
2163
+ "step": 359
2164
+ },
2165
+ {
2166
+ "epoch": 2.54,
2167
+ "learning_rate": 5.3720906924172596e-05,
2168
+ "loss": 0.7129,
2169
+ "step": 360
2170
+ },
2171
+ {
2172
+ "epoch": 2.54,
2173
+ "learning_rate": 5.348584821112711e-05,
2174
+ "loss": 0.8936,
2175
+ "step": 361
2176
+ },
2177
+ {
2178
+ "epoch": 2.55,
2179
+ "learning_rate": 5.325071205499198e-05,
2180
+ "loss": 0.7626,
2181
+ "step": 362
2182
+ },
2183
+ {
2184
+ "epoch": 2.56,
2185
+ "learning_rate": 5.301550367965228e-05,
2186
+ "loss": 0.7006,
2187
+ "step": 363
2188
+ },
2189
+ {
2190
+ "epoch": 2.56,
2191
+ "learning_rate": 5.278022831059751e-05,
2192
+ "loss": 0.8868,
2193
+ "step": 364
2194
+ },
2195
+ {
2196
+ "epoch": 2.57,
2197
+ "learning_rate": 5.254489117480556e-05,
2198
+ "loss": 0.7274,
2199
+ "step": 365
2200
+ },
2201
+ {
2202
+ "epoch": 2.58,
2203
+ "learning_rate": 5.230949750062655e-05,
2204
+ "loss": 0.7568,
2205
+ "step": 366
2206
+ },
2207
+ {
2208
+ "epoch": 2.58,
2209
+ "learning_rate": 5.2074052517666685e-05,
2210
+ "loss": 0.8734,
2211
+ "step": 367
2212
+ },
2213
+ {
2214
+ "epoch": 2.59,
2215
+ "learning_rate": 5.183856145667204e-05,
2216
+ "loss": 1.0261,
2217
+ "step": 368
2218
+ },
2219
+ {
2220
+ "epoch": 2.6,
2221
+ "learning_rate": 5.160302954941241e-05,
2222
+ "loss": 0.9876,
2223
+ "step": 369
2224
+ },
2225
+ {
2226
+ "epoch": 2.61,
2227
+ "learning_rate": 5.136746202856506e-05,
2228
+ "loss": 0.9113,
2229
+ "step": 370
2230
+ },
2231
+ {
2232
+ "epoch": 2.61,
2233
+ "learning_rate": 5.113186412759842e-05,
2234
+ "loss": 0.8247,
2235
+ "step": 371
2236
+ },
2237
+ {
2238
+ "epoch": 2.62,
2239
+ "learning_rate": 5.089624108065587e-05,
2240
+ "loss": 0.7352,
2241
+ "step": 372
2242
+ },
2243
+ {
2244
+ "epoch": 2.63,
2245
+ "learning_rate": 5.066059812243946e-05,
2246
+ "loss": 0.8319,
2247
+ "step": 373
2248
+ },
2249
+ {
2250
+ "epoch": 2.63,
2251
+ "learning_rate": 5.042494048809361e-05,
2252
+ "loss": 0.7573,
2253
+ "step": 374
2254
+ },
2255
+ {
2256
+ "epoch": 2.64,
2257
+ "learning_rate": 5.018927341308874e-05,
2258
+ "loss": 0.729,
2259
+ "step": 375
2260
+ },
2261
+ {
2262
+ "epoch": 2.65,
2263
+ "learning_rate": 4.9953602133105056e-05,
2264
+ "loss": 0.9483,
2265
+ "step": 376
2266
+ },
2267
+ {
2268
+ "epoch": 2.65,
2269
+ "learning_rate": 4.971793188391615e-05,
2270
+ "loss": 0.6948,
2271
+ "step": 377
2272
+ },
2273
+ {
2274
+ "epoch": 2.66,
2275
+ "learning_rate": 4.948226790127273e-05,
2276
+ "loss": 0.7836,
2277
+ "step": 378
2278
+ },
2279
+ {
2280
+ "epoch": 2.67,
2281
+ "learning_rate": 4.9246615420786315e-05,
2282
+ "loss": 0.647,
2283
+ "step": 379
2284
+ },
2285
+ {
2286
+ "epoch": 2.68,
2287
+ "learning_rate": 4.901097967781281e-05,
2288
+ "loss": 0.6558,
2289
+ "step": 380
2290
+ },
2291
+ {
2292
+ "epoch": 2.68,
2293
+ "learning_rate": 4.8775365907336345e-05,
2294
+ "loss": 0.7926,
2295
+ "step": 381
2296
+ },
2297
+ {
2298
+ "epoch": 2.69,
2299
+ "learning_rate": 4.8539779343852884e-05,
2300
+ "loss": 0.7926,
2301
+ "step": 382
2302
+ },
2303
+ {
2304
+ "epoch": 2.7,
2305
+ "learning_rate": 4.8304225221253895e-05,
2306
+ "loss": 0.818,
2307
+ "step": 383
2308
+ },
2309
+ {
2310
+ "epoch": 2.7,
2311
+ "learning_rate": 4.80687087727102e-05,
2312
+ "loss": 0.5522,
2313
+ "step": 384
2314
+ },
2315
+ {
2316
+ "epoch": 2.71,
2317
+ "learning_rate": 4.783323523055559e-05,
2318
+ "loss": 0.7327,
2319
+ "step": 385
2320
+ },
2321
+ {
2322
+ "epoch": 2.72,
2323
+ "learning_rate": 4.7597809826170646e-05,
2324
+ "loss": 0.8745,
2325
+ "step": 386
2326
+ },
2327
+ {
2328
+ "epoch": 2.73,
2329
+ "learning_rate": 4.736243778986651e-05,
2330
+ "loss": 0.8599,
2331
+ "step": 387
2332
+ },
2333
+ {
2334
+ "epoch": 2.73,
2335
+ "learning_rate": 4.7127124350768646e-05,
2336
+ "loss": 0.6869,
2337
+ "step": 388
2338
+ },
2339
+ {
2340
+ "epoch": 2.74,
2341
+ "learning_rate": 4.689187473670073e-05,
2342
+ "loss": 0.8579,
2343
+ "step": 389
2344
+ },
2345
+ {
2346
+ "epoch": 2.75,
2347
+ "learning_rate": 4.665669417406847e-05,
2348
+ "loss": 0.7942,
2349
+ "step": 390
2350
+ },
2351
+ {
2352
+ "epoch": 2.75,
2353
+ "learning_rate": 4.642158788774345e-05,
2354
+ "loss": 0.8683,
2355
+ "step": 391
2356
+ },
2357
+ {
2358
+ "epoch": 2.76,
2359
+ "learning_rate": 4.618656110094717e-05,
2360
+ "loss": 0.8248,
2361
+ "step": 392
2362
+ },
2363
+ {
2364
+ "epoch": 2.77,
2365
+ "learning_rate": 4.5951619035134914e-05,
2366
+ "loss": 0.8052,
2367
+ "step": 393
2368
+ },
2369
+ {
2370
+ "epoch": 2.77,
2371
+ "learning_rate": 4.571676690987972e-05,
2372
+ "loss": 0.7724,
2373
+ "step": 394
2374
+ },
2375
+ {
2376
+ "epoch": 2.78,
2377
+ "learning_rate": 4.548200994275652e-05,
2378
+ "loss": 0.8052,
2379
+ "step": 395
2380
+ },
2381
+ {
2382
+ "epoch": 2.79,
2383
+ "learning_rate": 4.524735334922615e-05,
2384
+ "loss": 0.8825,
2385
+ "step": 396
2386
+ },
2387
+ {
2388
+ "epoch": 2.8,
2389
+ "learning_rate": 4.501280234251952e-05,
2390
+ "loss": 0.8237,
2391
+ "step": 397
2392
+ },
2393
+ {
2394
+ "epoch": 2.8,
2395
+ "learning_rate": 4.477836213352178e-05,
2396
+ "loss": 0.5574,
2397
+ "step": 398
2398
+ },
2399
+ {
2400
+ "epoch": 2.81,
2401
+ "learning_rate": 4.4544037930656526e-05,
2402
+ "loss": 0.7633,
2403
+ "step": 399
2404
+ },
2405
+ {
2406
+ "epoch": 2.82,
2407
+ "learning_rate": 4.430983493977012e-05,
2408
+ "loss": 0.6846,
2409
+ "step": 400
2410
+ },
2411
+ {
2412
+ "epoch": 2.82,
2413
+ "learning_rate": 4.407575836401605e-05,
2414
+ "loss": 0.7732,
2415
+ "step": 401
2416
+ },
2417
+ {
2418
+ "epoch": 2.83,
2419
+ "learning_rate": 4.384181340373927e-05,
2420
+ "loss": 0.8842,
2421
+ "step": 402
2422
+ },
2423
+ {
2424
+ "epoch": 2.84,
2425
+ "learning_rate": 4.360800525636076e-05,
2426
+ "loss": 0.618,
2427
+ "step": 403
2428
+ },
2429
+ {
2430
+ "epoch": 2.85,
2431
+ "learning_rate": 4.337433911626198e-05,
2432
+ "loss": 0.7508,
2433
+ "step": 404
2434
+ },
2435
+ {
2436
+ "epoch": 2.85,
2437
+ "learning_rate": 4.314082017466947e-05,
2438
+ "loss": 0.6947,
2439
+ "step": 405
2440
+ },
2441
+ {
2442
+ "epoch": 2.86,
2443
+ "learning_rate": 4.2907453619539596e-05,
2444
+ "loss": 0.7075,
2445
+ "step": 406
2446
+ },
2447
+ {
2448
+ "epoch": 2.87,
2449
+ "learning_rate": 4.267424463544318e-05,
2450
+ "loss": 0.9874,
2451
+ "step": 407
2452
+ },
2453
+ {
2454
+ "epoch": 2.87,
2455
+ "learning_rate": 4.244119840345045e-05,
2456
+ "loss": 0.8875,
2457
+ "step": 408
2458
+ },
2459
+ {
2460
+ "epoch": 2.88,
2461
+ "learning_rate": 4.2208320101015816e-05,
2462
+ "loss": 0.6921,
2463
+ "step": 409
2464
+ },
2465
+ {
2466
+ "epoch": 2.89,
2467
+ "learning_rate": 4.197561490186291e-05,
2468
+ "loss": 0.8342,
2469
+ "step": 410
2470
+ },
2471
+ {
2472
+ "epoch": 2.89,
2473
+ "learning_rate": 4.174308797586961e-05,
2474
+ "loss": 0.7022,
2475
+ "step": 411
2476
+ },
2477
+ {
2478
+ "epoch": 2.9,
2479
+ "learning_rate": 4.151074448895324e-05,
2480
+ "loss": 0.8656,
2481
+ "step": 412
2482
+ },
2483
+ {
2484
+ "epoch": 2.91,
2485
+ "learning_rate": 4.1278589602955706e-05,
2486
+ "loss": 0.9449,
2487
+ "step": 413
2488
+ },
2489
+ {
2490
+ "epoch": 2.92,
2491
+ "learning_rate": 4.1046628475528936e-05,
2492
+ "loss": 0.9049,
2493
+ "step": 414
2494
+ },
2495
+ {
2496
+ "epoch": 2.92,
2497
+ "learning_rate": 4.081486626002021e-05,
2498
+ "loss": 0.8814,
2499
+ "step": 415
2500
+ },
2501
+ {
2502
+ "epoch": 2.93,
2503
+ "learning_rate": 4.058330810535766e-05,
2504
+ "loss": 0.7045,
2505
+ "step": 416
2506
+ },
2507
+ {
2508
+ "epoch": 2.94,
2509
+ "learning_rate": 4.035195915593598e-05,
2510
+ "loss": 0.8558,
2511
+ "step": 417
2512
+ },
2513
+ {
2514
+ "epoch": 2.94,
2515
+ "learning_rate": 4.012082455150201e-05,
2516
+ "loss": 0.772,
2517
+ "step": 418
2518
+ },
2519
+ {
2520
+ "epoch": 2.95,
2521
+ "learning_rate": 3.988990942704066e-05,
2522
+ "loss": 0.7281,
2523
+ "step": 419
2524
+ },
2525
+ {
2526
+ "epoch": 2.96,
2527
+ "learning_rate": 3.9659218912660766e-05,
2528
+ "loss": 0.7919,
2529
+ "step": 420
2530
+ },
2531
+ {
2532
+ "epoch": 2.96,
2533
+ "learning_rate": 3.9428758133481106e-05,
2534
+ "loss": 0.9592,
2535
+ "step": 421
2536
+ },
2537
+ {
2538
+ "epoch": 2.97,
2539
+ "learning_rate": 3.919853220951658e-05,
2540
+ "loss": 0.6567,
2541
+ "step": 422
2542
+ },
2543
+ {
2544
+ "epoch": 2.98,
2545
+ "learning_rate": 3.896854625556448e-05,
2546
+ "loss": 0.9045,
2547
+ "step": 423
2548
+ },
2549
+ {
2550
+ "epoch": 2.99,
2551
+ "learning_rate": 3.873880538109073e-05,
2552
+ "loss": 0.9526,
2553
+ "step": 424
2554
+ },
2555
+ {
2556
+ "epoch": 2.99,
2557
+ "learning_rate": 3.850931469011656e-05,
2558
+ "loss": 0.6942,
2559
+ "step": 425
2560
+ },
2561
+ {
2562
+ "epoch": 3.0,
2563
+ "learning_rate": 3.8280079281104996e-05,
2564
+ "loss": 0.604,
2565
+ "step": 426
2566
+ },
2567
+ {
2568
+ "epoch": 3.01,
2569
+ "learning_rate": 3.805110424684754e-05,
2570
+ "loss": 0.7876,
2571
+ "step": 427
2572
+ },
2573
+ {
2574
+ "epoch": 3.01,
2575
+ "learning_rate": 3.782239467435119e-05,
2576
+ "loss": 0.6762,
2577
+ "step": 428
2578
+ },
2579
+ {
2580
+ "epoch": 3.02,
2581
+ "learning_rate": 3.7593955644725263e-05,
2582
+ "loss": 0.6553,
2583
+ "step": 429
2584
+ },
2585
+ {
2586
+ "epoch": 3.03,
2587
+ "learning_rate": 3.736579223306863e-05,
2588
+ "loss": 0.9052,
2589
+ "step": 430
2590
+ },
2591
+ {
2592
+ "epoch": 3.04,
2593
+ "learning_rate": 3.713790950835691e-05,
2594
+ "loss": 0.6197,
2595
+ "step": 431
2596
+ },
2597
+ {
2598
+ "epoch": 3.04,
2599
+ "learning_rate": 3.691031253332981e-05,
2600
+ "loss": 0.8027,
2601
+ "step": 432
2602
+ },
2603
+ {
2604
+ "epoch": 3.05,
2605
+ "learning_rate": 3.6683006364378745e-05,
2606
+ "loss": 0.8929,
2607
+ "step": 433
2608
+ },
2609
+ {
2610
+ "epoch": 3.06,
2611
+ "learning_rate": 3.64559960514345e-05,
2612
+ "loss": 0.5936,
2613
+ "step": 434
2614
+ },
2615
+ {
2616
+ "epoch": 3.06,
2617
+ "learning_rate": 3.6229286637854865e-05,
2618
+ "loss": 0.6819,
2619
+ "step": 435
2620
+ },
2621
+ {
2622
+ "epoch": 3.07,
2623
+ "learning_rate": 3.6002883160312865e-05,
2624
+ "loss": 0.6546,
2625
+ "step": 436
2626
+ },
2627
+ {
2628
+ "epoch": 3.08,
2629
+ "learning_rate": 3.577679064868466e-05,
2630
+ "loss": 0.6387,
2631
+ "step": 437
2632
+ },
2633
+ {
2634
+ "epoch": 3.08,
2635
+ "learning_rate": 3.555101412593781e-05,
2636
+ "loss": 0.8402,
2637
+ "step": 438
2638
+ },
2639
+ {
2640
+ "epoch": 3.09,
2641
+ "learning_rate": 3.532555860801982e-05,
2642
+ "loss": 0.7592,
2643
+ "step": 439
2644
+ },
2645
+ {
2646
+ "epoch": 3.1,
2647
+ "learning_rate": 3.5100429103746547e-05,
2648
+ "loss": 0.7132,
2649
+ "step": 440
2650
+ },
2651
+ {
2652
+ "epoch": 3.11,
2653
+ "learning_rate": 3.487563061469102e-05,
2654
+ "loss": 0.6006,
2655
+ "step": 441
2656
+ },
2657
+ {
2658
+ "epoch": 3.11,
2659
+ "learning_rate": 3.465116813507232e-05,
2660
+ "loss": 0.758,
2661
+ "step": 442
2662
+ },
2663
+ {
2664
+ "epoch": 3.12,
2665
+ "learning_rate": 3.4427046651644523e-05,
2666
+ "loss": 0.7745,
2667
+ "step": 443
2668
+ },
2669
+ {
2670
+ "epoch": 3.13,
2671
+ "learning_rate": 3.420327114358606e-05,
2672
+ "loss": 0.8207,
2673
+ "step": 444
2674
+ },
2675
+ {
2676
+ "epoch": 3.13,
2677
+ "learning_rate": 3.397984658238902e-05,
2678
+ "loss": 0.7432,
2679
+ "step": 445
2680
+ },
2681
+ {
2682
+ "epoch": 3.14,
2683
+ "learning_rate": 3.375677793174862e-05,
2684
+ "loss": 0.7818,
2685
+ "step": 446
2686
+ },
2687
+ {
2688
+ "epoch": 3.15,
2689
+ "learning_rate": 3.3534070147453135e-05,
2690
+ "loss": 0.6546,
2691
+ "step": 447
2692
+ },
2693
+ {
2694
+ "epoch": 3.15,
2695
+ "learning_rate": 3.331172817727363e-05,
2696
+ "loss": 0.8025,
2697
+ "step": 448
2698
+ },
2699
+ {
2700
+ "epoch": 3.16,
2701
+ "learning_rate": 3.308975696085405e-05,
2702
+ "loss": 0.854,
2703
+ "step": 449
2704
+ },
2705
+ {
2706
+ "epoch": 3.17,
2707
+ "learning_rate": 3.2868161429601584e-05,
2708
+ "loss": 0.8708,
2709
+ "step": 450
2710
+ },
2711
+ {
2712
+ "epoch": 3.18,
2713
+ "learning_rate": 3.264694650657698e-05,
2714
+ "loss": 0.6904,
2715
+ "step": 451
2716
+ },
2717
+ {
2718
+ "epoch": 3.18,
2719
+ "learning_rate": 3.242611710638526e-05,
2720
+ "loss": 0.6629,
2721
+ "step": 452
2722
+ },
2723
+ {
2724
+ "epoch": 3.19,
2725
+ "learning_rate": 3.220567813506653e-05,
2726
+ "loss": 0.7047,
2727
+ "step": 453
2728
+ },
2729
+ {
2730
+ "epoch": 3.2,
2731
+ "learning_rate": 3.198563448998691e-05,
2732
+ "loss": 0.6103,
2733
+ "step": 454
2734
+ },
2735
+ {
2736
+ "epoch": 3.2,
2737
+ "learning_rate": 3.176599105972982e-05,
2738
+ "loss": 0.7088,
2739
+ "step": 455
2740
+ },
2741
+ {
2742
+ "epoch": 3.21,
2743
+ "learning_rate": 3.154675272398734e-05,
2744
+ "loss": 0.692,
2745
+ "step": 456
2746
+ },
2747
+ {
2748
+ "epoch": 3.22,
2749
+ "learning_rate": 3.132792435345173e-05,
2750
+ "loss": 0.7956,
2751
+ "step": 457
2752
+ },
2753
+ {
2754
+ "epoch": 3.23,
2755
+ "learning_rate": 3.1109510809707395e-05,
2756
+ "loss": 0.5122,
2757
+ "step": 458
2758
+ },
2759
+ {
2760
+ "epoch": 3.23,
2761
+ "learning_rate": 3.0891516945122735e-05,
2762
+ "loss": 0.6636,
2763
+ "step": 459
2764
+ },
2765
+ {
2766
+ "epoch": 3.24,
2767
+ "learning_rate": 3.06739476027423e-05,
2768
+ "loss": 0.5402,
2769
+ "step": 460
2770
+ },
2771
+ {
2772
+ "epoch": 3.25,
2773
+ "learning_rate": 3.0456807616179385e-05,
2774
+ "loss": 0.7909,
2775
+ "step": 461
2776
+ },
2777
+ {
2778
+ "epoch": 3.25,
2779
+ "learning_rate": 3.0240101809508447e-05,
2780
+ "loss": 0.7802,
2781
+ "step": 462
2782
+ },
2783
+ {
2784
+ "epoch": 3.26,
2785
+ "learning_rate": 3.0023834997158084e-05,
2786
+ "loss": 0.6254,
2787
+ "step": 463
2788
+ },
2789
+ {
2790
+ "epoch": 3.27,
2791
+ "learning_rate": 2.9808011983803974e-05,
2792
+ "loss": 0.7085,
2793
+ "step": 464
2794
+ },
2795
+ {
2796
+ "epoch": 3.27,
2797
+ "learning_rate": 2.9592637564262155e-05,
2798
+ "loss": 0.6821,
2799
+ "step": 465
2800
+ },
2801
+ {
2802
+ "epoch": 3.28,
2803
+ "learning_rate": 2.9377716523382548e-05,
2804
+ "loss": 0.5261,
2805
+ "step": 466
2806
+ },
2807
+ {
2808
+ "epoch": 3.29,
2809
+ "learning_rate": 2.916325363594261e-05,
2810
+ "loss": 0.6004,
2811
+ "step": 467
2812
+ },
2813
+ {
2814
+ "epoch": 3.3,
2815
+ "learning_rate": 2.894925366654125e-05,
2816
+ "loss": 0.7009,
2817
+ "step": 468
2818
+ },
2819
+ {
2820
+ "epoch": 3.3,
2821
+ "learning_rate": 2.8735721369492998e-05,
2822
+ "loss": 0.6891,
2823
+ "step": 469
2824
+ },
2825
+ {
2826
+ "epoch": 3.31,
2827
+ "learning_rate": 2.8522661488722402e-05,
2828
+ "loss": 0.7982,
2829
+ "step": 470
2830
+ },
2831
+ {
2832
+ "epoch": 3.32,
2833
+ "learning_rate": 2.8310078757658552e-05,
2834
+ "loss": 0.5411,
2835
+ "step": 471
2836
+ },
2837
+ {
2838
+ "epoch": 3.32,
2839
+ "learning_rate": 2.809797789913004e-05,
2840
+ "loss": 0.9861,
2841
+ "step": 472
2842
+ },
2843
+ {
2844
+ "epoch": 3.33,
2845
+ "learning_rate": 2.788636362525995e-05,
2846
+ "loss": 0.6231,
2847
+ "step": 473
2848
+ },
2849
+ {
2850
+ "epoch": 3.34,
2851
+ "learning_rate": 2.7675240637361127e-05,
2852
+ "loss": 0.8455,
2853
+ "step": 474
2854
+ },
2855
+ {
2856
+ "epoch": 3.35,
2857
+ "learning_rate": 2.7464613625831947e-05,
2858
+ "loss": 0.6272,
2859
+ "step": 475
2860
+ },
2861
+ {
2862
+ "epoch": 3.35,
2863
+ "learning_rate": 2.725448727005181e-05,
2864
+ "loss": 0.6168,
2865
+ "step": 476
2866
+ },
2867
+ {
2868
+ "epoch": 3.36,
2869
+ "learning_rate": 2.704486623827742e-05,
2870
+ "loss": 0.7557,
2871
+ "step": 477
2872
+ },
2873
+ {
2874
+ "epoch": 3.37,
2875
+ "learning_rate": 2.683575518753899e-05,
2876
+ "loss": 0.8038,
2877
+ "step": 478
2878
+ },
2879
+ {
2880
+ "epoch": 3.37,
2881
+ "learning_rate": 2.662715876353672e-05,
2882
+ "loss": 0.5731,
2883
+ "step": 479
2884
+ },
2885
+ {
2886
+ "epoch": 3.38,
2887
+ "learning_rate": 2.6419081600537676e-05,
2888
+ "loss": 0.5925,
2889
+ "step": 480
2890
+ },
2891
+ {
2892
+ "epoch": 3.39,
2893
+ "learning_rate": 2.6211528321272815e-05,
2894
+ "loss": 0.3563,
2895
+ "step": 481
2896
+ },
2897
+ {
2898
+ "epoch": 3.39,
2899
+ "learning_rate": 2.600450353683426e-05,
2900
+ "loss": 0.8047,
2901
+ "step": 482
2902
+ },
2903
+ {
2904
+ "epoch": 3.4,
2905
+ "learning_rate": 2.5798011846572845e-05,
2906
+ "loss": 0.6462,
2907
+ "step": 483
2908
+ },
2909
+ {
2910
+ "epoch": 3.41,
2911
+ "learning_rate": 2.5592057837996008e-05,
2912
+ "loss": 0.6358,
2913
+ "step": 484
2914
+ },
2915
+ {
2916
+ "epoch": 3.42,
2917
+ "learning_rate": 2.5386646086665756e-05,
2918
+ "loss": 0.5355,
2919
+ "step": 485
2920
+ },
2921
+ {
2922
+ "epoch": 3.42,
2923
+ "learning_rate": 2.518178115609714e-05,
2924
+ "loss": 0.6574,
2925
+ "step": 486
2926
+ },
2927
+ {
2928
+ "epoch": 3.43,
2929
+ "learning_rate": 2.4977467597656733e-05,
2930
+ "loss": 0.8254,
2931
+ "step": 487
2932
+ },
2933
+ {
2934
+ "epoch": 3.44,
2935
+ "learning_rate": 2.4773709950461665e-05,
2936
+ "loss": 0.6084,
2937
+ "step": 488
2938
+ },
2939
+ {
2940
+ "epoch": 3.44,
2941
+ "learning_rate": 2.4570512741278672e-05,
2942
+ "loss": 0.5879,
2943
+ "step": 489
2944
+ },
2945
+ {
2946
+ "epoch": 3.45,
2947
+ "learning_rate": 2.436788048442356e-05,
2948
+ "loss": 0.5913,
2949
+ "step": 490
2950
+ },
2951
+ {
2952
+ "epoch": 3.46,
2953
+ "learning_rate": 2.4165817681660923e-05,
2954
+ "loss": 0.5537,
2955
+ "step": 491
2956
+ },
2957
+ {
2958
+ "epoch": 3.46,
2959
+ "learning_rate": 2.3964328822104127e-05,
2960
+ "loss": 0.7386,
2961
+ "step": 492
2962
+ },
2963
+ {
2964
+ "epoch": 3.47,
2965
+ "learning_rate": 2.3763418382115522e-05,
2966
+ "loss": 0.6689,
2967
+ "step": 493
2968
+ },
2969
+ {
2970
+ "epoch": 3.48,
2971
+ "learning_rate": 2.356309082520709e-05,
2972
+ "loss": 0.72,
2973
+ "step": 494
2974
+ },
2975
+ {
2976
+ "epoch": 3.49,
2977
+ "learning_rate": 2.3363350601941254e-05,
2978
+ "loss": 0.6096,
2979
+ "step": 495
2980
+ },
2981
+ {
2982
+ "epoch": 3.49,
2983
+ "learning_rate": 2.3164202149831875e-05,
2984
+ "loss": 0.6886,
2985
+ "step": 496
2986
+ },
2987
+ {
2988
+ "epoch": 3.5,
2989
+ "learning_rate": 2.2965649893245943e-05,
2990
+ "loss": 0.75,
2991
+ "step": 497
2992
+ },
2993
+ {
2994
+ "epoch": 3.51,
2995
+ "learning_rate": 2.276769824330496e-05,
2996
+ "loss": 0.7009,
2997
+ "step": 498
2998
+ },
2999
+ {
3000
+ "epoch": 3.51,
3001
+ "learning_rate": 2.2570351597787166e-05,
3002
+ "loss": 0.527,
3003
+ "step": 499
3004
+ },
3005
+ {
3006
+ "epoch": 3.52,
3007
+ "learning_rate": 2.2373614341029776e-05,
3008
+ "loss": 0.8298,
3009
+ "step": 500
3010
+ },
3011
+ {
3012
+ "epoch": 3.53,
3013
+ "learning_rate": 2.217749084383151e-05,
3014
+ "loss": 0.6151,
3015
+ "step": 501
3016
+ },
3017
+ {
3018
+ "epoch": 3.54,
3019
+ "learning_rate": 2.1981985463355587e-05,
3020
+ "loss": 0.4804,
3021
+ "step": 502
3022
+ },
3023
+ {
3024
+ "epoch": 3.54,
3025
+ "learning_rate": 2.178710254303288e-05,
3026
+ "loss": 0.6926,
3027
+ "step": 503
3028
+ },
3029
+ {
3030
+ "epoch": 3.55,
3031
+ "learning_rate": 2.1592846412465408e-05,
3032
+ "loss": 0.5709,
3033
+ "step": 504
3034
+ },
3035
+ {
3036
+ "epoch": 3.56,
3037
+ "learning_rate": 2.139922138733016e-05,
3038
+ "loss": 0.4891,
3039
+ "step": 505
3040
+ },
3041
+ {
3042
+ "epoch": 3.56,
3043
+ "learning_rate": 2.1206231769283253e-05,
3044
+ "loss": 0.7093,
3045
+ "step": 506
3046
+ },
3047
+ {
3048
+ "epoch": 3.57,
3049
+ "learning_rate": 2.101388184586426e-05,
3050
+ "loss": 0.5319,
3051
+ "step": 507
3052
+ },
3053
+ {
3054
+ "epoch": 3.58,
3055
+ "learning_rate": 2.0822175890401102e-05,
3056
+ "loss": 0.5793,
3057
+ "step": 508
3058
+ },
3059
+ {
3060
+ "epoch": 3.58,
3061
+ "learning_rate": 2.0631118161914965e-05,
3062
+ "loss": 0.6307,
3063
+ "step": 509
3064
+ },
3065
+ {
3066
+ "epoch": 3.59,
3067
+ "learning_rate": 2.0440712905025788e-05,
3068
+ "loss": 0.7734,
3069
+ "step": 510
3070
+ },
3071
+ {
3072
+ "epoch": 3.6,
3073
+ "learning_rate": 2.025096434985793e-05,
3074
+ "loss": 0.7496,
3075
+ "step": 511
3076
+ },
3077
+ {
3078
+ "epoch": 3.61,
3079
+ "learning_rate": 2.006187671194616e-05,
3080
+ "loss": 0.7035,
3081
+ "step": 512
3082
+ },
3083
+ {
3084
+ "epoch": 3.61,
3085
+ "learning_rate": 1.9873454192142038e-05,
3086
+ "loss": 0.6277,
3087
+ "step": 513
3088
+ },
3089
+ {
3090
+ "epoch": 3.62,
3091
+ "learning_rate": 1.9685700976520605e-05,
3092
+ "loss": 0.504,
3093
+ "step": 514
3094
+ },
3095
+ {
3096
+ "epoch": 3.63,
3097
+ "learning_rate": 1.9498621236287275e-05,
3098
+ "loss": 0.6165,
3099
+ "step": 515
3100
+ },
3101
+ {
3102
+ "epoch": 3.63,
3103
+ "learning_rate": 1.931221912768532e-05,
3104
+ "loss": 0.5932,
3105
+ "step": 516
3106
+ },
3107
+ {
3108
+ "epoch": 3.64,
3109
+ "learning_rate": 1.9126498791903464e-05,
3110
+ "loss": 0.5616,
3111
+ "step": 517
3112
+ },
3113
+ {
3114
+ "epoch": 3.65,
3115
+ "learning_rate": 1.894146435498378e-05,
3116
+ "loss": 0.7417,
3117
+ "step": 518
3118
+ },
3119
+ {
3120
+ "epoch": 3.65,
3121
+ "learning_rate": 1.8757119927730266e-05,
3122
+ "loss": 0.4589,
3123
+ "step": 519
3124
+ },
3125
+ {
3126
+ "epoch": 3.66,
3127
+ "learning_rate": 1.8573469605617234e-05,
3128
+ "loss": 0.5317,
3129
+ "step": 520
3130
+ },
3131
+ {
3132
+ "epoch": 3.67,
3133
+ "learning_rate": 1.839051746869854e-05,
3134
+ "loss": 0.4569,
3135
+ "step": 521
3136
+ },
3137
+ {
3138
+ "epoch": 3.68,
3139
+ "learning_rate": 1.8208267581516853e-05,
3140
+ "loss": 0.4805,
3141
+ "step": 522
3142
+ },
3143
+ {
3144
+ "epoch": 3.68,
3145
+ "learning_rate": 1.8026723993013317e-05,
3146
+ "loss": 0.6394,
3147
+ "step": 523
3148
+ },
3149
+ {
3150
+ "epoch": 3.69,
3151
+ "learning_rate": 1.7845890736437686e-05,
3152
+ "loss": 0.5725,
3153
+ "step": 524
3154
+ },
3155
+ {
3156
+ "epoch": 3.7,
3157
+ "learning_rate": 1.766577182925867e-05,
3158
+ "loss": 0.6262,
3159
+ "step": 525
3160
+ },
3161
+ {
3162
+ "epoch": 3.7,
3163
+ "learning_rate": 1.748637127307468e-05,
3164
+ "loss": 0.4061,
3165
+ "step": 526
3166
+ },
3167
+ {
3168
+ "epoch": 3.71,
3169
+ "learning_rate": 1.7307693053524948e-05,
3170
+ "loss": 0.5609,
3171
+ "step": 527
3172
+ },
3173
+ {
3174
+ "epoch": 3.72,
3175
+ "learning_rate": 1.7129741140200966e-05,
3176
+ "loss": 0.6813,
3177
+ "step": 528
3178
+ },
3179
+ {
3180
+ "epoch": 3.73,
3181
+ "learning_rate": 1.6952519486558255e-05,
3182
+ "loss": 0.6705,
3183
+ "step": 529
3184
+ },
3185
+ {
3186
+ "epoch": 3.73,
3187
+ "learning_rate": 1.6776032029828652e-05,
3188
+ "loss": 0.5283,
3189
+ "step": 530
3190
+ },
3191
+ {
3192
+ "epoch": 3.74,
3193
+ "learning_rate": 1.6600282690932685e-05,
3194
+ "loss": 0.647,
3195
+ "step": 531
3196
+ },
3197
+ {
3198
+ "epoch": 3.75,
3199
+ "learning_rate": 1.642527537439259e-05,
3200
+ "loss": 0.6204,
3201
+ "step": 532
3202
+ },
3203
+ {
3204
+ "epoch": 3.75,
3205
+ "learning_rate": 1.625101396824551e-05,
3206
+ "loss": 0.7142,
3207
+ "step": 533
3208
+ },
3209
+ {
3210
+ "epoch": 3.76,
3211
+ "learning_rate": 1.6077502343957117e-05,
3212
+ "loss": 0.6346,
3213
+ "step": 534
3214
+ },
3215
+ {
3216
+ "epoch": 3.77,
3217
+ "learning_rate": 1.5904744356335615e-05,
3218
+ "loss": 0.6174,
3219
+ "step": 535
3220
+ },
3221
+ {
3222
+ "epoch": 3.77,
3223
+ "learning_rate": 1.5732743843446112e-05,
3224
+ "loss": 0.5119,
3225
+ "step": 536
3226
+ },
3227
+ {
3228
+ "epoch": 3.78,
3229
+ "learning_rate": 1.5561504626525274e-05,
3230
+ "loss": 0.5757,
3231
+ "step": 537
3232
+ },
3233
+ {
3234
+ "epoch": 3.79,
3235
+ "learning_rate": 1.539103050989656e-05,
3236
+ "loss": 0.6991,
3237
+ "step": 538
3238
+ },
3239
+ {
3240
+ "epoch": 3.8,
3241
+ "learning_rate": 1.5221325280885606e-05,
3242
+ "loss": 0.6397,
3243
+ "step": 539
3244
+ },
3245
+ {
3246
+ "epoch": 3.8,
3247
+ "learning_rate": 1.5052392709736101e-05,
3248
+ "loss": 0.3825,
3249
+ "step": 540
3250
+ },
3251
+ {
3252
+ "epoch": 3.81,
3253
+ "learning_rate": 1.488423654952606e-05,
3254
+ "loss": 0.5771,
3255
+ "step": 541
3256
+ },
3257
+ {
3258
+ "epoch": 3.82,
3259
+ "learning_rate": 1.4716860536084403e-05,
3260
+ "loss": 0.5234,
3261
+ "step": 542
3262
+ },
3263
+ {
3264
+ "epoch": 3.82,
3265
+ "learning_rate": 1.4550268387907996e-05,
3266
+ "loss": 0.6009,
3267
+ "step": 543
3268
+ },
3269
+ {
3270
+ "epoch": 3.83,
3271
+ "learning_rate": 1.4384463806079024e-05,
3272
+ "loss": 0.7015,
3273
+ "step": 544
3274
+ },
3275
+ {
3276
+ "epoch": 3.84,
3277
+ "learning_rate": 1.4219450474182722e-05,
3278
+ "loss": 0.4498,
3279
+ "step": 545
3280
+ },
3281
+ {
3282
+ "epoch": 3.85,
3283
+ "learning_rate": 1.4055232058225609e-05,
3284
+ "loss": 0.553,
3285
+ "step": 546
3286
+ },
3287
+ {
3288
+ "epoch": 3.85,
3289
+ "learning_rate": 1.3891812206554023e-05,
3290
+ "loss": 0.503,
3291
+ "step": 547
3292
+ },
3293
+ {
3294
+ "epoch": 3.86,
3295
+ "learning_rate": 1.3729194549773039e-05,
3296
+ "loss": 0.5165,
3297
+ "step": 548
3298
+ },
3299
+ {
3300
+ "epoch": 3.87,
3301
+ "learning_rate": 1.3567382700665826e-05,
3302
+ "loss": 0.8116,
3303
+ "step": 549
3304
+ },
3305
+ {
3306
+ "epoch": 3.87,
3307
+ "learning_rate": 1.3406380254113421e-05,
3308
+ "loss": 0.7193,
3309
+ "step": 550
3310
+ },
3311
+ {
3312
+ "epoch": 3.88,
3313
+ "learning_rate": 1.3246190787014767e-05,
3314
+ "loss": 0.5503,
3315
+ "step": 551
3316
+ },
3317
+ {
3318
+ "epoch": 3.89,
3319
+ "learning_rate": 1.3086817858207368e-05,
3320
+ "loss": 0.6821,
3321
+ "step": 552
3322
+ },
3323
+ {
3324
+ "epoch": 3.89,
3325
+ "learning_rate": 1.2928265008388119e-05,
3326
+ "loss": 0.5011,
3327
+ "step": 553
3328
+ },
3329
+ {
3330
+ "epoch": 3.9,
3331
+ "learning_rate": 1.2770535760034718e-05,
3332
+ "loss": 0.6798,
3333
+ "step": 554
3334
+ },
3335
+ {
3336
+ "epoch": 3.91,
3337
+ "learning_rate": 1.2613633617327386e-05,
3338
+ "loss": 0.7772,
3339
+ "step": 555
3340
+ },
3341
+ {
3342
+ "epoch": 3.92,
3343
+ "learning_rate": 1.245756206607101e-05,
3344
+ "loss": 0.7508,
3345
+ "step": 556
3346
+ },
3347
+ {
3348
+ "epoch": 3.92,
3349
+ "learning_rate": 1.2302324573617674e-05,
3350
+ "loss": 0.7464,
3351
+ "step": 557
3352
+ },
3353
+ {
3354
+ "epoch": 3.93,
3355
+ "learning_rate": 1.21479245887897e-05,
3356
+ "loss": 0.494,
3357
+ "step": 558
3358
+ },
3359
+ {
3360
+ "epoch": 3.94,
3361
+ "learning_rate": 1.1994365541802938e-05,
3362
+ "loss": 0.6598,
3363
+ "step": 559
3364
+ },
3365
+ {
3366
+ "epoch": 3.94,
3367
+ "learning_rate": 1.1841650844190639e-05,
3368
+ "loss": 0.6079,
3369
+ "step": 560
3370
+ },
3371
+ {
3372
+ "epoch": 3.95,
3373
+ "learning_rate": 1.1689783888727612e-05,
3374
+ "loss": 0.5854,
3375
+ "step": 561
3376
+ },
3377
+ {
3378
+ "epoch": 3.96,
3379
+ "learning_rate": 1.1538768049354886e-05,
3380
+ "loss": 0.6023,
3381
+ "step": 562
3382
+ },
3383
+ {
3384
+ "epoch": 3.96,
3385
+ "learning_rate": 1.1388606681104701e-05,
3386
+ "loss": 0.8083,
3387
+ "step": 563
3388
+ },
3389
+ {
3390
+ "epoch": 3.97,
3391
+ "learning_rate": 1.1239303120026045e-05,
3392
+ "loss": 0.5125,
3393
+ "step": 564
3394
+ },
3395
+ {
3396
+ "epoch": 3.98,
3397
+ "learning_rate": 1.1090860683110432e-05,
3398
+ "loss": 0.7413,
3399
+ "step": 565
3400
+ },
3401
+ {
3402
+ "epoch": 3.99,
3403
+ "learning_rate": 1.094328266821833e-05,
3404
+ "loss": 0.7644,
3405
+ "step": 566
3406
+ },
3407
+ {
3408
+ "epoch": 3.99,
3409
+ "learning_rate": 1.0796572354005836e-05,
3410
+ "loss": 0.5222,
3411
+ "step": 567
3412
+ },
3413
+ {
3414
+ "epoch": 4.0,
3415
+ "learning_rate": 1.0650732999851793e-05,
3416
+ "loss": 0.4381,
3417
+ "step": 568
3418
+ },
3419
+ {
3420
+ "epoch": 4.01,
3421
+ "learning_rate": 1.050576784578547e-05,
3422
+ "loss": 0.5834,
3423
+ "step": 569
3424
+ },
3425
+ {
3426
+ "epoch": 4.01,
3427
+ "learning_rate": 1.0361680112414518e-05,
3428
+ "loss": 0.5592,
3429
+ "step": 570
3430
+ },
3431
+ {
3432
+ "epoch": 4.02,
3433
+ "learning_rate": 1.0218473000853435e-05,
3434
+ "loss": 0.54,
3435
+ "step": 571
3436
+ },
3437
+ {
3438
+ "epoch": 4.03,
3439
+ "learning_rate": 1.007614969265247e-05,
3440
+ "loss": 0.7499,
3441
+ "step": 572
3442
+ },
3443
+ {
3444
+ "epoch": 4.04,
3445
+ "learning_rate": 9.93471334972687e-06,
3446
+ "loss": 0.5057,
3447
+ "step": 573
3448
+ },
3449
+ {
3450
+ "epoch": 4.04,
3451
+ "learning_rate": 9.794167114286728e-06,
3452
+ "loss": 0.6646,
3453
+ "step": 574
3454
+ },
3455
+ {
3456
+ "epoch": 4.05,
3457
+ "learning_rate": 9.654514108767122e-06,
3458
+ "loss": 0.7452,
3459
+ "step": 575
3460
+ },
3461
+ {
3462
+ "epoch": 4.06,
3463
+ "learning_rate": 9.515757435758743e-06,
3464
+ "loss": 0.4576,
3465
+ "step": 576
3466
+ },
3467
+ {
3468
+ "epoch": 4.06,
3469
+ "learning_rate": 9.377900177938981e-06,
3470
+ "loss": 0.5076,
3471
+ "step": 577
3472
+ },
3473
+ {
3474
+ "epoch": 4.07,
3475
+ "learning_rate": 9.240945398003448e-06,
3476
+ "loss": 0.5335,
3477
+ "step": 578
3478
+ },
3479
+ {
3480
+ "epoch": 4.08,
3481
+ "learning_rate": 9.104896138597884e-06,
3482
+ "loss": 0.4748,
3483
+ "step": 579
3484
+ },
3485
+ {
3486
+ "epoch": 4.08,
3487
+ "learning_rate": 8.969755422250642e-06,
3488
+ "loss": 0.6551,
3489
+ "step": 580
3490
+ },
3491
+ {
3492
+ "epoch": 4.09,
3493
+ "learning_rate": 8.835526251305487e-06,
3494
+ "loss": 0.6218,
3495
+ "step": 581
3496
+ },
3497
+ {
3498
+ "epoch": 4.1,
3499
+ "learning_rate": 8.702211607854871e-06,
3500
+ "loss": 0.5859,
3501
+ "step": 582
3502
+ },
3503
+ {
3504
+ "epoch": 4.11,
3505
+ "learning_rate": 8.569814453673747e-06,
3506
+ "loss": 0.4408,
3507
+ "step": 583
3508
+ },
3509
+ {
3510
+ "epoch": 4.11,
3511
+ "learning_rate": 8.438337730153728e-06,
3512
+ "loss": 0.6113,
3513
+ "step": 584
3514
+ },
3515
+ {
3516
+ "epoch": 4.12,
3517
+ "learning_rate": 8.307784358237736e-06,
3518
+ "loss": 0.6236,
3519
+ "step": 585
3520
+ },
3521
+ {
3522
+ "epoch": 4.13,
3523
+ "learning_rate": 8.178157238355149e-06,
3524
+ "loss": 0.6943,
3525
+ "step": 586
3526
+ },
3527
+ {
3528
+ "epoch": 4.13,
3529
+ "learning_rate": 8.049459250357284e-06,
3530
+ "loss": 0.6044,
3531
+ "step": 587
3532
+ },
3533
+ {
3534
+ "epoch": 4.14,
3535
+ "learning_rate": 7.92169325345351e-06,
3536
+ "loss": 0.6424,
3537
+ "step": 588
3538
+ },
3539
+ {
3540
+ "epoch": 4.15,
3541
+ "learning_rate": 7.794862086147686e-06,
3542
+ "loss": 0.5592,
3543
+ "step": 589
3544
+ },
3545
+ {
3546
+ "epoch": 4.15,
3547
+ "learning_rate": 7.668968566175067e-06,
3548
+ "loss": 0.6441,
3549
+ "step": 590
3550
+ },
3551
+ {
3552
+ "epoch": 4.16,
3553
+ "learning_rate": 7.544015490439743e-06,
3554
+ "loss": 0.7385,
3555
+ "step": 591
3556
+ },
3557
+ {
3558
+ "epoch": 4.17,
3559
+ "learning_rate": 7.420005634952548e-06,
3560
+ "loss": 0.7177,
3561
+ "step": 592
3562
+ },
3563
+ {
3564
+ "epoch": 4.18,
3565
+ "learning_rate": 7.296941754769243e-06,
3566
+ "loss": 0.5613,
3567
+ "step": 593
3568
+ },
3569
+ {
3570
+ "epoch": 4.18,
3571
+ "learning_rate": 7.174826583929478e-06,
3572
+ "loss": 0.5203,
3573
+ "step": 594
3574
+ },
3575
+ {
3576
+ "epoch": 4.19,
3577
+ "learning_rate": 7.053662835395908e-06,
3578
+ "loss": 0.5635,
3579
+ "step": 595
3580
+ },
3581
+ {
3582
+ "epoch": 4.2,
3583
+ "learning_rate": 6.933453200994033e-06,
3584
+ "loss": 0.4943,
3585
+ "step": 596
3586
+ },
3587
+ {
3588
+ "epoch": 4.2,
3589
+ "learning_rate": 6.814200351352323e-06,
3590
+ "loss": 0.5714,
3591
+ "step": 597
3592
+ },
3593
+ {
3594
+ "epoch": 4.21,
3595
+ "learning_rate": 6.695906935842916e-06,
3596
+ "loss": 0.5431,
3597
+ "step": 598
3598
+ },
3599
+ {
3600
+ "epoch": 4.22,
3601
+ "learning_rate": 6.5785755825227554e-06,
3602
+ "loss": 0.6152,
3603
+ "step": 599
3604
+ },
3605
+ {
3606
+ "epoch": 4.23,
3607
+ "learning_rate": 6.462208898075201e-06,
3608
+ "loss": 0.3967,
3609
+ "step": 600
3610
+ },
3611
+ {
3612
+ "epoch": 4.23,
3613
+ "learning_rate": 6.346809467752101e-06,
3614
+ "loss": 0.51,
3615
+ "step": 601
3616
+ },
3617
+ {
3618
+ "epoch": 4.24,
3619
+ "learning_rate": 6.232379855316395e-06,
3620
+ "loss": 0.3982,
3621
+ "step": 602
3622
+ },
3623
+ {
3624
+ "epoch": 4.25,
3625
+ "learning_rate": 6.118922602985133e-06,
3626
+ "loss": 0.6577,
3627
+ "step": 603
3628
+ },
3629
+ {
3630
+ "epoch": 4.25,
3631
+ "learning_rate": 6.006440231372973e-06,
3632
+ "loss": 0.6096,
3633
+ "step": 604
3634
+ },
3635
+ {
3636
+ "epoch": 4.26,
3637
+ "learning_rate": 5.894935239436239e-06,
3638
+ "loss": 0.4768,
3639
+ "step": 605
3640
+ },
3641
+ {
3642
+ "epoch": 4.27,
3643
+ "learning_rate": 5.78441010441736e-06,
3644
+ "loss": 0.582,
3645
+ "step": 606
3646
+ },
3647
+ {
3648
+ "epoch": 4.27,
3649
+ "learning_rate": 5.674867281789842e-06,
3650
+ "loss": 0.5339,
3651
+ "step": 607
3652
+ },
3653
+ {
3654
+ "epoch": 4.28,
3655
+ "learning_rate": 5.566309205203735e-06,
3656
+ "loss": 0.3802,
3657
+ "step": 608
3658
+ },
3659
+ {
3660
+ "epoch": 4.29,
3661
+ "learning_rate": 5.458738286431525e-06,
3662
+ "loss": 0.4993,
3663
+ "step": 609
3664
+ },
3665
+ {
3666
+ "epoch": 4.3,
3667
+ "learning_rate": 5.352156915314594e-06,
3668
+ "loss": 0.5956,
3669
+ "step": 610
3670
+ },
3671
+ {
3672
+ "epoch": 4.3,
3673
+ "learning_rate": 5.2465674597101275e-06,
3674
+ "loss": 0.5637,
3675
+ "step": 611
3676
+ },
3677
+ {
3678
+ "epoch": 4.31,
3679
+ "learning_rate": 5.141972265438455e-06,
3680
+ "loss": 0.6436,
3681
+ "step": 612
3682
+ },
3683
+ {
3684
+ "epoch": 4.32,
3685
+ "learning_rate": 5.0383736562310105e-06,
3686
+ "loss": 0.4381,
3687
+ "step": 613
3688
+ },
3689
+ {
3690
+ "epoch": 4.32,
3691
+ "learning_rate": 4.935773933678645e-06,
3692
+ "loss": 0.8673,
3693
+ "step": 614
3694
+ },
3695
+ {
3696
+ "epoch": 4.33,
3697
+ "learning_rate": 4.8341753771805275e-06,
3698
+ "loss": 0.5128,
3699
+ "step": 615
3700
+ },
3701
+ {
3702
+ "epoch": 4.34,
3703
+ "learning_rate": 4.733580243893509e-06,
3704
+ "loss": 0.7217,
3705
+ "step": 616
3706
+ },
3707
+ {
3708
+ "epoch": 4.35,
3709
+ "learning_rate": 4.633990768681928e-06,
3710
+ "loss": 0.5057,
3711
+ "step": 617
3712
+ },
3713
+ {
3714
+ "epoch": 4.35,
3715
+ "learning_rate": 4.535409164068022e-06,
3716
+ "loss": 0.4874,
3717
+ "step": 618
3718
+ },
3719
+ {
3720
+ "epoch": 4.36,
3721
+ "learning_rate": 4.437837620182745e-06,
3722
+ "loss": 0.5992,
3723
+ "step": 619
3724
+ },
3725
+ {
3726
+ "epoch": 4.37,
3727
+ "learning_rate": 4.341278304717078e-06,
3728
+ "loss": 0.7009,
3729
+ "step": 620
3730
+ },
3731
+ {
3732
+ "epoch": 4.37,
3733
+ "learning_rate": 4.2457333628739605e-06,
3734
+ "loss": 0.4728,
3735
+ "step": 621
3736
+ },
3737
+ {
3738
+ "epoch": 4.38,
3739
+ "learning_rate": 4.1512049173205445e-06,
3740
+ "loss": 0.4825,
3741
+ "step": 622
3742
+ },
3743
+ {
3744
+ "epoch": 4.39,
3745
+ "learning_rate": 4.057695068141043e-06,
3746
+ "loss": 0.264,
3747
+ "step": 623
3748
+ },
3749
+ {
3750
+ "epoch": 4.39,
3751
+ "learning_rate": 3.9652058927901415e-06,
3752
+ "loss": 0.6728,
3753
+ "step": 624
3754
+ },
3755
+ {
3756
+ "epoch": 4.4,
3757
+ "learning_rate": 3.873739446046787e-06,
3758
+ "loss": 0.5413,
3759
+ "step": 625
3760
+ },
3761
+ {
3762
+ "epoch": 4.41,
3763
+ "learning_rate": 3.7832977599685236e-06,
3764
+ "loss": 0.5163,
3765
+ "step": 626
3766
+ },
3767
+ {
3768
+ "epoch": 4.42,
3769
+ "learning_rate": 3.693882843846419e-06,
3770
+ "loss": 0.4204,
3771
+ "step": 627
3772
+ },
3773
+ {
3774
+ "epoch": 4.42,
3775
+ "learning_rate": 3.605496684160348e-06,
3776
+ "loss": 0.5551,
3777
+ "step": 628
3778
+ },
3779
+ {
3780
+ "epoch": 4.43,
3781
+ "learning_rate": 3.518141244534917e-06,
3782
+ "loss": 0.7123,
3783
+ "step": 629
3784
+ },
3785
+ {
3786
+ "epoch": 4.44,
3787
+ "learning_rate": 3.4318184656958074e-06,
3788
+ "loss": 0.5059,
3789
+ "step": 630
3790
+ },
3791
+ {
3792
+ "epoch": 4.44,
3793
+ "learning_rate": 3.3465302654266616e-06,
3794
+ "loss": 0.4831,
3795
+ "step": 631
3796
+ },
3797
+ {
3798
+ "epoch": 4.45,
3799
+ "learning_rate": 3.262278538526492e-06,
3800
+ "loss": 0.4672,
3801
+ "step": 632
3802
+ },
3803
+ {
3804
+ "epoch": 4.46,
3805
+ "learning_rate": 3.17906515676758e-06,
3806
+ "loss": 0.4423,
3807
+ "step": 633
3808
+ },
3809
+ {
3810
+ "epoch": 4.46,
3811
+ "learning_rate": 3.0968919688538656e-06,
3812
+ "loss": 0.6304,
3813
+ "step": 634
3814
+ },
3815
+ {
3816
+ "epoch": 4.47,
3817
+ "learning_rate": 3.0157608003799355e-06,
3818
+ "loss": 0.547,
3819
+ "step": 635
3820
+ },
3821
+ {
3822
+ "epoch": 4.48,
3823
+ "learning_rate": 2.9356734537903997e-06,
3824
+ "loss": 0.6087,
3825
+ "step": 636
3826
+ },
3827
+ {
3828
+ "epoch": 4.49,
3829
+ "learning_rate": 2.8566317083398907e-06,
3830
+ "loss": 0.5036,
3831
+ "step": 637
3832
+ },
3833
+ {
3834
+ "epoch": 4.49,
3835
+ "learning_rate": 2.7786373200535175e-06,
3836
+ "loss": 0.5607,
3837
+ "step": 638
3838
+ },
3839
+ {
3840
+ "epoch": 4.5,
3841
+ "learning_rate": 2.701692021687846e-06,
3842
+ "loss": 0.6446,
3843
+ "step": 639
3844
+ },
3845
+ {
3846
+ "epoch": 4.51,
3847
+ "learning_rate": 2.625797522692425e-06,
3848
+ "loss": 0.6112,
3849
+ "step": 640
3850
+ },
3851
+ {
3852
+ "epoch": 4.51,
3853
+ "learning_rate": 2.550955509171804e-06,
3854
+ "loss": 0.4442,
3855
+ "step": 641
3856
+ },
3857
+ {
3858
+ "epoch": 4.52,
3859
+ "learning_rate": 2.477167643848033e-06,
3860
+ "loss": 0.7012,
3861
+ "step": 642
3862
+ },
3863
+ {
3864
+ "epoch": 4.53,
3865
+ "learning_rate": 2.404435566023794e-06,
3866
+ "loss": 0.4942,
3867
+ "step": 643
3868
+ },
3869
+ {
3870
+ "epoch": 4.54,
3871
+ "learning_rate": 2.3327608915459378e-06,
3872
+ "loss": 0.3784,
3873
+ "step": 644
3874
+ },
3875
+ {
3876
+ "epoch": 4.54,
3877
+ "learning_rate": 2.262145212769562e-06,
3878
+ "loss": 0.5823,
3879
+ "step": 645
3880
+ },
3881
+ {
3882
+ "epoch": 4.55,
3883
+ "learning_rate": 2.192590098522701e-06,
3884
+ "loss": 0.4819,
3885
+ "step": 646
3886
+ },
3887
+ {
3888
+ "epoch": 4.56,
3889
+ "learning_rate": 2.1240970940714146e-06,
3890
+ "loss": 0.3844,
3891
+ "step": 647
3892
+ },
3893
+ {
3894
+ "epoch": 4.56,
3895
+ "learning_rate": 2.056667721085481e-06,
3896
+ "loss": 0.622,
3897
+ "step": 648
3898
+ },
3899
+ {
3900
+ "epoch": 4.57,
3901
+ "learning_rate": 1.9903034776045935e-06,
3902
+ "loss": 0.438,
3903
+ "step": 649
3904
+ },
3905
+ {
3906
+ "epoch": 4.58,
3907
+ "learning_rate": 1.9250058380050674e-06,
3908
+ "loss": 0.5024,
3909
+ "step": 650
3910
+ },
3911
+ {
3912
+ "epoch": 4.58,
3913
+ "learning_rate": 1.8607762529671059e-06,
3914
+ "loss": 0.5313,
3915
+ "step": 651
3916
+ },
3917
+ {
3918
+ "epoch": 4.59,
3919
+ "learning_rate": 1.7976161494425493e-06,
3920
+ "loss": 0.6647,
3921
+ "step": 652
3922
+ },
3923
+ {
3924
+ "epoch": 4.6,
3925
+ "learning_rate": 1.7355269306231602e-06,
3926
+ "loss": 0.6355,
3927
+ "step": 653
3928
+ },
3929
+ {
3930
+ "epoch": 4.61,
3931
+ "learning_rate": 1.6745099759094996e-06,
3932
+ "loss": 0.6087,
3933
+ "step": 654
3934
+ },
3935
+ {
3936
+ "epoch": 4.61,
3937
+ "learning_rate": 1.6145666408802496e-06,
3938
+ "loss": 0.5331,
3939
+ "step": 655
3940
+ },
3941
+ {
3942
+ "epoch": 4.62,
3943
+ "learning_rate": 1.5556982572620783e-06,
3944
+ "loss": 0.419,
3945
+ "step": 656
3946
+ },
3947
+ {
3948
+ "epoch": 4.63,
3949
+ "learning_rate": 1.49790613290009e-06,
3950
+ "loss": 0.5193,
3951
+ "step": 657
3952
+ },
3953
+ {
3954
+ "epoch": 4.63,
3955
+ "learning_rate": 1.4411915517287543e-06,
3956
+ "loss": 0.5199,
3957
+ "step": 658
3958
+ },
3959
+ {
3960
+ "epoch": 4.64,
3961
+ "learning_rate": 1.38555577374338e-06,
3962
+ "loss": 0.4913,
3963
+ "step": 659
3964
+ },
3965
+ {
3966
+ "epoch": 4.65,
3967
+ "learning_rate": 1.3310000349721297e-06,
3968
+ "loss": 0.6488,
3969
+ "step": 660
3970
+ },
3971
+ {
3972
+ "epoch": 4.65,
3973
+ "learning_rate": 1.2775255474485436e-06,
3974
+ "loss": 0.3613,
3975
+ "step": 661
3976
+ },
3977
+ {
3978
+ "epoch": 4.66,
3979
+ "learning_rate": 1.225133499184633e-06,
3980
+ "loss": 0.4291,
3981
+ "step": 662
3982
+ },
3983
+ {
3984
+ "epoch": 4.67,
3985
+ "learning_rate": 1.1738250541444728e-06,
3986
+ "loss": 0.3886,
3987
+ "step": 663
3988
+ },
3989
+ {
3990
+ "epoch": 4.68,
3991
+ "learning_rate": 1.1236013522183341e-06,
3992
+ "loss": 0.4071,
3993
+ "step": 664
3994
+ },
3995
+ {
3996
+ "epoch": 4.68,
3997
+ "learning_rate": 1.0744635091974098e-06,
3998
+ "loss": 0.5659,
3999
+ "step": 665
4000
+ },
4001
+ {
4002
+ "epoch": 4.69,
4003
+ "learning_rate": 1.0264126167489508e-06,
4004
+ "loss": 0.4816,
4005
+ "step": 666
4006
+ },
4007
+ {
4008
+ "epoch": 4.7,
4009
+ "learning_rate": 9.794497423920635e-07,
4010
+ "loss": 0.5466,
4011
+ "step": 667
4012
+ },
4013
+ {
4014
+ "epoch": 4.7,
4015
+ "learning_rate": 9.335759294739899e-07,
4016
+ "loss": 0.3509,
4017
+ "step": 668
4018
+ },
4019
+ {
4020
+ "epoch": 4.71,
4021
+ "learning_rate": 8.887921971469037e-07,
4022
+ "loss": 0.485,
4023
+ "step": 669
4024
+ },
4025
+ {
4026
+ "epoch": 4.72,
4027
+ "learning_rate": 8.450995403452955e-07,
4028
+ "loss": 0.6001,
4029
+ "step": 670
4030
+ },
4031
+ {
4032
+ "epoch": 4.73,
4033
+ "learning_rate": 8.024989297638452e-07,
4034
+ "loss": 0.5909,
4035
+ "step": 671
4036
+ },
4037
+ {
4038
+ "epoch": 4.73,
4039
+ "learning_rate": 7.609913118358791e-07,
4040
+ "loss": 0.4665,
4041
+ "step": 672
4042
+ },
4043
+ {
4044
+ "epoch": 4.74,
4045
+ "learning_rate": 7.205776087123362e-07,
4046
+ "loss": 0.5521,
4047
+ "step": 673
4048
+ },
4049
+ {
4050
+ "epoch": 4.75,
4051
+ "learning_rate": 6.812587182412622e-07,
4052
+ "loss": 0.5523,
4053
+ "step": 674
4054
+ },
4055
+ {
4056
+ "epoch": 4.75,
4057
+ "learning_rate": 6.430355139478871e-07,
4058
+ "loss": 0.6478,
4059
+ "step": 675
4060
+ },
4061
+ {
4062
+ "epoch": 4.76,
4063
+ "learning_rate": 6.059088450152128e-07,
4064
+ "loss": 0.5499,
4065
+ "step": 676
4066
+ },
4067
+ {
4068
+ "epoch": 4.77,
4069
+ "learning_rate": 5.698795362651443e-07,
4070
+ "loss": 0.5434,
4071
+ "step": 677
4072
+ },
4073
+ {
4074
+ "epoch": 4.77,
4075
+ "learning_rate": 5.349483881401496e-07,
4076
+ "loss": 0.4236,
4077
+ "step": 678
4078
+ },
4079
+ {
4080
+ "epoch": 4.78,
4081
+ "learning_rate": 5.011161766855177e-07,
4082
+ "loss": 0.4867,
4083
+ "step": 679
4084
+ },
4085
+ {
4086
+ "epoch": 4.79,
4087
+ "learning_rate": 4.683836535320674e-07,
4088
+ "loss": 0.6196,
4089
+ "step": 680
4090
+ },
4091
+ {
4092
+ "epoch": 4.8,
4093
+ "learning_rate": 4.3675154587949907e-07,
4094
+ "loss": 0.5592,
4095
+ "step": 681
4096
+ },
4097
+ {
4098
+ "epoch": 4.8,
4099
+ "learning_rate": 4.0622055648018573e-07,
4100
+ "loss": 0.3231,
4101
+ "step": 682
4102
+ },
4103
+ {
4104
+ "epoch": 4.81,
4105
+ "learning_rate": 3.7679136362360757e-07,
4106
+ "loss": 0.5,
4107
+ "step": 683
4108
+ },
4109
+ {
4110
+ "epoch": 4.82,
4111
+ "learning_rate": 3.484646211212528e-07,
4112
+ "loss": 0.4595,
4113
+ "step": 684
4114
+ },
4115
+ {
4116
+ "epoch": 4.82,
4117
+ "learning_rate": 3.212409582921128e-07,
4118
+ "loss": 0.5326,
4119
+ "step": 685
4120
+ },
4121
+ {
4122
+ "epoch": 4.83,
4123
+ "learning_rate": 2.951209799486765e-07,
4124
+ "loss": 0.6091,
4125
+ "step": 686
4126
+ },
4127
+ {
4128
+ "epoch": 4.84,
4129
+ "learning_rate": 2.7010526638351906e-07,
4130
+ "loss": 0.3778,
4131
+ "step": 687
4132
+ },
4133
+ {
4134
+ "epoch": 4.85,
4135
+ "learning_rate": 2.4619437335640647e-07,
4136
+ "loss": 0.4767,
4137
+ "step": 688
4138
+ },
4139
+ {
4140
+ "epoch": 4.85,
4141
+ "learning_rate": 2.233888320819111e-07,
4142
+ "loss": 0.4401,
4143
+ "step": 689
4144
+ },
4145
+ {
4146
+ "epoch": 4.86,
4147
+ "learning_rate": 2.0168914921767644e-07,
4148
+ "loss": 0.4482,
4149
+ "step": 690
4150
+ },
4151
+ {
4152
+ "epoch": 4.87,
4153
+ "learning_rate": 1.8109580685309324e-07,
4154
+ "loss": 0.7395,
4155
+ "step": 691
4156
+ },
4157
+ {
4158
+ "epoch": 4.87,
4159
+ "learning_rate": 1.616092624986354e-07,
4160
+ "loss": 0.6509,
4161
+ "step": 692
4162
+ },
4163
+ {
4164
+ "epoch": 4.88,
4165
+ "learning_rate": 1.4322994907568498e-07,
4166
+ "loss": 0.495,
4167
+ "step": 693
4168
+ },
4169
+ {
4170
+ "epoch": 4.89,
4171
+ "learning_rate": 1.259582749068955e-07,
4172
+ "loss": 0.6137,
4173
+ "step": 694
4174
+ },
4175
+ {
4176
+ "epoch": 4.89,
4177
+ "learning_rate": 1.0979462370714344e-07,
4178
+ "loss": 0.4285,
4179
+ "step": 695
4180
+ },
4181
+ {
4182
+ "epoch": 4.9,
4183
+ "learning_rate": 9.473935457500194e-08,
4184
+ "loss": 0.5902,
4185
+ "step": 696
4186
+ },
4187
+ {
4188
+ "epoch": 4.91,
4189
+ "learning_rate": 8.079280198473038e-08,
4190
+ "loss": 0.702,
4191
+ "step": 697
4192
+ },
4193
+ {
4194
+ "epoch": 4.92,
4195
+ "learning_rate": 6.79552757789026e-08,
4196
+ "loss": 0.6918,
4197
+ "step": 698
4198
+ },
4199
+ {
4200
+ "epoch": 4.92,
4201
+ "learning_rate": 5.622706116146237e-08,
4202
+ "loss": 0.6861,
4203
+ "step": 699
4204
+ },
4205
+ {
4206
+ "epoch": 4.93,
4207
+ "learning_rate": 4.560841869142851e-08,
4208
+ "loss": 0.4192,
4209
+ "step": 700
4210
+ },
4211
+ {
4212
+ "epoch": 4.94,
4213
+ "learning_rate": 3.609958427708282e-08,
4214
+ "loss": 0.5755,
4215
+ "step": 701
4216
+ },
4217
+ {
4218
+ "epoch": 4.94,
4219
+ "learning_rate": 2.770076917075204e-08,
4220
+ "loss": 0.5318,
4221
+ "step": 702
4222
+ },
4223
+ {
4224
+ "epoch": 4.95,
4225
+ "learning_rate": 2.0412159964083853e-08,
4226
+ "loss": 0.5324,
4227
+ "step": 703
4228
+ },
4229
+ {
4230
+ "epoch": 4.96,
4231
+ "learning_rate": 1.4233918583939077e-08,
4232
+ "loss": 0.5296,
4233
+ "step": 704
4234
+ },
4235
+ {
4236
+ "epoch": 4.96,
4237
+ "learning_rate": 9.166182288750103e-09,
4238
+ "loss": 0.7441,
4239
+ "step": 705
4240
+ },
4241
+ {
4242
+ "epoch": 4.97,
4243
+ "learning_rate": 5.209063665506664e-09,
4244
+ "loss": 0.4587,
4245
+ "step": 706
4246
+ },
4247
+ {
4248
+ "epoch": 4.98,
4249
+ "learning_rate": 2.362650627252272e-09,
4250
+ "loss": 0.671,
4251
+ "step": 707
4252
+ },
4253
+ {
4254
+ "epoch": 4.99,
4255
+ "learning_rate": 6.270064111024709e-10,
4256
+ "loss": 0.6712,
4257
+ "step": 708
4258
+ },
4259
+ {
4260
+ "epoch": 4.99,
4261
+ "learning_rate": 2.169576873711776e-12,
4262
+ "loss": 0.4572,
4263
+ "step": 709
4264
+ },
4265
+ {
4266
+ "epoch": 5.0,
4267
+ "learning_rate": 4.881540062118273e-10,
4268
+ "loss": 0.3704,
4269
+ "step": 710
4270
+ }
4271
+ ],
4272
+ "logging_steps": 1,
4273
+ "max_steps": 710,
4274
+ "num_train_epochs": 5,
4275
+ "save_steps": 500,
4276
+ "total_flos": 5.732084547059712e+16,
4277
+ "trial_name": null,
4278
+ "trial_params": null
4279
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f72f18edadc0963679b06a76f5557e051ff7b8426dc196148038e8e3448eb90
3
+ size 4091