Hugging Face's logo

Devy1
/
Qwen2.5-Coder-CONTROL-checkpoints_python_only_2k-1.5B-Base-5

Safetensors
qwen2
Model card Files
xet
 Community
Devy1 commited on
Commit
009e1f2
·
verified ·
1 Parent(s): 8bf2229

Upload checkpoint Qwen2.5-Coder-CONTROL-checkpoints_python_only_2k-1.5B-Base-5 at epoch 5

Browse files
Files changed (15) hide show
  1. .gitattributes +1 -0
  2. added_tokens.json +25 -0
  3. config.json +59 -0
  4. generation_config.json +9 -0
  5. merges.txt +0 -0
  6. model.safetensors +3 -0
  7. optimizer.pt +3 -0
  8. rng_state.pth +3 -0
  9. scheduler.pt +3 -0
  10. special_tokens_map.json +31 -0
  11. tokenizer.json +3 -0
  12. tokenizer_config.json +216 -0
  13. trainer_state.json +4059 -0
  14. training_args.bin +3 -0
  15. vocab.json +0 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
added_tokens.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|PAD_TOKEN|>": 151665,
5
+ "<|box_end|>": 151649,
6
+ "<|box_start|>": 151648,
7
+ "<|endoftext|>": 151643,
8
+ "<|file_sep|>": 151664,
9
+ "<|fim_middle|>": 151660,
10
+ "<|fim_pad|>": 151662,
11
+ "<|fim_prefix|>": 151659,
12
+ "<|fim_suffix|>": 151661,
13
+ "<|im_end|>": 151645,
14
+ "<|im_start|>": 151644,
15
+ "<|image_pad|>": 151655,
16
+ "<|object_ref_end|>": 151647,
17
+ "<|object_ref_start|>": 151646,
18
+ "<|quad_end|>": 151651,
19
+ "<|quad_start|>": 151650,
20
+ "<|repo_name|>": 151663,
21
+ "<|video_pad|>": 151656,
22
+ "<|vision_end|>": 151653,
23
+ "<|vision_pad|>": 151654,
24
+ "<|vision_start|>": 151652
25
+ }
config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "dtype": "bfloat16",
7
+ "eos_token_id": 151643,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 1536,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 8960,
12
+ "layer_types": [
13
+ "full_attention",
14
+ "full_attention",
15
+ "full_attention",
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention"
41
+ ],
42
+ "max_position_embeddings": 32768,
43
+ "max_window_layers": 21,
44
+ "model_type": "qwen2",
45
+ "num_attention_heads": 12,
46
+ "num_hidden_layers": 28,
47
+ "num_key_value_heads": 2,
48
+ "pad_token_id": 151665,
49
+ "rms_norm_eps": 1e-06,
50
+ "rope_scaling": null,
51
+ "rope_theta": 1000000.0,
52
+ "sliding_window": null,
53
+ "tie_word_embeddings": true,
54
+ "transformers_version": "4.57.6",
55
+ "unsloth_fixed": true,
56
+ "use_cache": true,
57
+ "use_sliding_window": false,
58
+ "vocab_size": 151936
59
+ }
generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "eos_token_id": [
3
+ 151643
4
+ ],
5
+ "max_length": 32768,
6
+ "max_new_tokens": 2048,
7
+ "pad_token_id": 151665,
8
+ "transformers_version": "4.57.6"
9
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:170e3a68e99854c4a60b0501df17112c52107f7745603990010f58f1364a0eff
3
+ size 3087467144
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28e0aab9e1d13a1b06ac9a590a3ebf4404e9ab99826591c2234462ec5db2dbd7
3
+ size 6175153765
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:718a0f3db00824213036a2c0441849791319b7d9cf189065873bb26a7020738e
3
+ size 14645
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8b2f36b1ee7eb13062fd9fd60674496580c34a8f6f1660d2e4b523009896f2b
3
+ size 1465
special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|endoftext|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|PAD_TOKEN|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fab42efe8d17406525a9154b728cf9e957629a8ed7ce997770efdd71128c6a1a
3
+ size 11422086
tokenizer_config.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ },
181
+ "151665": {
182
+ "content": "<|PAD_TOKEN|>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": true
188
+ }
189
+ },
190
+ "additional_special_tokens": [
191
+ "<|im_start|>",
192
+ "<|im_end|>",
193
+ "<|object_ref_start|>",
194
+ "<|object_ref_end|>",
195
+ "<|box_start|>",
196
+ "<|box_end|>",
197
+ "<|quad_start|>",
198
+ "<|quad_end|>",
199
+ "<|vision_start|>",
200
+ "<|vision_end|>",
201
+ "<|vision_pad|>",
202
+ "<|image_pad|>",
203
+ "<|video_pad|>"
204
+ ],
205
+ "bos_token": null,
206
+ "clean_up_tokenization_spaces": false,
207
+ "eos_token": "<|endoftext|>",
208
+ "errors": "replace",
209
+ "extra_special_tokens": {},
210
+ "model_max_length": 2048,
211
+ "pad_token": "<|PAD_TOKEN|>",
212
+ "padding_side": "left",
213
+ "split_special_tokens": false,
214
+ "tokenizer_class": "Qwen2Tokenizer",
215
+ "unk_token": null
216
+ }
trainer_state.json ADDED
@@ -0,0 +1,4059 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 5.0,
6
+ "eval_steps": 500,
7
+ "global_step": 575,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.008705114254624592,
14
+ "grad_norm": 1.703125,
15
+ "learning_rate": 5e-05,
16
+ "loss": 0.8632,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.017410228509249184,
21
+ "grad_norm": 2.015625,
22
+ "learning_rate": 4.9956521739130436e-05,
23
+ "loss": 0.7891,
24
+ "step": 2
25
+ },
26
+ {
27
+ "epoch": 0.026115342763873776,
28
+ "grad_norm": 8.375,
29
+ "learning_rate": 4.9913043478260876e-05,
30
+ "loss": 0.7896,
31
+ "step": 3
32
+ },
33
+ {
34
+ "epoch": 0.03482045701849837,
35
+ "grad_norm": 3.328125,
36
+ "learning_rate": 4.986956521739131e-05,
37
+ "loss": 0.9271,
38
+ "step": 4
39
+ },
40
+ {
41
+ "epoch": 0.04352557127312296,
42
+ "grad_norm": 1.640625,
43
+ "learning_rate": 4.9826086956521736e-05,
44
+ "loss": 0.9154,
45
+ "step": 5
46
+ },
47
+ {
48
+ "epoch": 0.05223068552774755,
49
+ "grad_norm": 1.5390625,
50
+ "learning_rate": 4.9782608695652176e-05,
51
+ "loss": 0.9267,
52
+ "step": 6
53
+ },
54
+ {
55
+ "epoch": 0.060935799782372145,
56
+ "grad_norm": 2.71875,
57
+ "learning_rate": 4.973913043478261e-05,
58
+ "loss": 0.8189,
59
+ "step": 7
60
+ },
61
+ {
62
+ "epoch": 0.06964091403699674,
63
+ "grad_norm": 1.5859375,
64
+ "learning_rate": 4.969565217391304e-05,
65
+ "loss": 0.8018,
66
+ "step": 8
67
+ },
68
+ {
69
+ "epoch": 0.07834602829162132,
70
+ "grad_norm": 1.359375,
71
+ "learning_rate": 4.9652173913043483e-05,
72
+ "loss": 0.9439,
73
+ "step": 9
74
+ },
75
+ {
76
+ "epoch": 0.08705114254624592,
77
+ "grad_norm": 2.375,
78
+ "learning_rate": 4.960869565217392e-05,
79
+ "loss": 0.8444,
80
+ "step": 10
81
+ },
82
+ {
83
+ "epoch": 0.0957562568008705,
84
+ "grad_norm": 2.125,
85
+ "learning_rate": 4.956521739130435e-05,
86
+ "loss": 1.013,
87
+ "step": 11
88
+ },
89
+ {
90
+ "epoch": 0.1044613710554951,
91
+ "grad_norm": 1.7109375,
92
+ "learning_rate": 4.9521739130434784e-05,
93
+ "loss": 1.1719,
94
+ "step": 12
95
+ },
96
+ {
97
+ "epoch": 0.11316648531011969,
98
+ "grad_norm": 1.4296875,
99
+ "learning_rate": 4.947826086956522e-05,
100
+ "loss": 1.0034,
101
+ "step": 13
102
+ },
103
+ {
104
+ "epoch": 0.12187159956474429,
105
+ "grad_norm": 1.25,
106
+ "learning_rate": 4.943478260869566e-05,
107
+ "loss": 1.0015,
108
+ "step": 14
109
+ },
110
+ {
111
+ "epoch": 0.1305767138193689,
112
+ "grad_norm": 1.25,
113
+ "learning_rate": 4.939130434782609e-05,
114
+ "loss": 0.9413,
115
+ "step": 15
116
+ },
117
+ {
118
+ "epoch": 0.13928182807399347,
119
+ "grad_norm": 1.5546875,
120
+ "learning_rate": 4.9347826086956524e-05,
121
+ "loss": 1.0435,
122
+ "step": 16
123
+ },
124
+ {
125
+ "epoch": 0.14798694232861806,
126
+ "grad_norm": 2.265625,
127
+ "learning_rate": 4.930434782608696e-05,
128
+ "loss": 0.9285,
129
+ "step": 17
130
+ },
131
+ {
132
+ "epoch": 0.15669205658324264,
133
+ "grad_norm": 1.3046875,
134
+ "learning_rate": 4.926086956521739e-05,
135
+ "loss": 0.91,
136
+ "step": 18
137
+ },
138
+ {
139
+ "epoch": 0.16539717083786726,
140
+ "grad_norm": 1.296875,
141
+ "learning_rate": 4.9217391304347824e-05,
142
+ "loss": 0.937,
143
+ "step": 19
144
+ },
145
+ {
146
+ "epoch": 0.17410228509249184,
147
+ "grad_norm": 1.28125,
148
+ "learning_rate": 4.9173913043478265e-05,
149
+ "loss": 0.9194,
150
+ "step": 20
151
+ },
152
+ {
153
+ "epoch": 0.18280739934711643,
154
+ "grad_norm": 1.203125,
155
+ "learning_rate": 4.91304347826087e-05,
156
+ "loss": 0.9392,
157
+ "step": 21
158
+ },
159
+ {
160
+ "epoch": 0.191512513601741,
161
+ "grad_norm": 1.3203125,
162
+ "learning_rate": 4.908695652173913e-05,
163
+ "loss": 0.935,
164
+ "step": 22
165
+ },
166
+ {
167
+ "epoch": 0.20021762785636563,
168
+ "grad_norm": 1.2421875,
169
+ "learning_rate": 4.904347826086957e-05,
170
+ "loss": 0.7352,
171
+ "step": 23
172
+ },
173
+ {
174
+ "epoch": 0.2089227421109902,
175
+ "grad_norm": 1.2578125,
176
+ "learning_rate": 4.9e-05,
177
+ "loss": 0.8006,
178
+ "step": 24
179
+ },
180
+ {
181
+ "epoch": 0.2176278563656148,
182
+ "grad_norm": 1.3046875,
183
+ "learning_rate": 4.895652173913044e-05,
184
+ "loss": 1.1003,
185
+ "step": 25
186
+ },
187
+ {
188
+ "epoch": 0.22633297062023938,
189
+ "grad_norm": 1.578125,
190
+ "learning_rate": 4.891304347826087e-05,
191
+ "loss": 0.8786,
192
+ "step": 26
193
+ },
194
+ {
195
+ "epoch": 0.235038084874864,
196
+ "grad_norm": 1.390625,
197
+ "learning_rate": 4.8869565217391305e-05,
198
+ "loss": 0.9133,
199
+ "step": 27
200
+ },
201
+ {
202
+ "epoch": 0.24374319912948858,
203
+ "grad_norm": 1.109375,
204
+ "learning_rate": 4.8826086956521746e-05,
205
+ "loss": 0.9822,
206
+ "step": 28
207
+ },
208
+ {
209
+ "epoch": 0.25244831338411317,
210
+ "grad_norm": 1.6640625,
211
+ "learning_rate": 4.878260869565218e-05,
212
+ "loss": 1.0016,
213
+ "step": 29
214
+ },
215
+ {
216
+ "epoch": 0.2611534276387378,
217
+ "grad_norm": 1.34375,
218
+ "learning_rate": 4.873913043478261e-05,
219
+ "loss": 1.1215,
220
+ "step": 30
221
+ },
222
+ {
223
+ "epoch": 0.26985854189336234,
224
+ "grad_norm": 1.171875,
225
+ "learning_rate": 4.8695652173913046e-05,
226
+ "loss": 1.1457,
227
+ "step": 31
228
+ },
229
+ {
230
+ "epoch": 0.27856365614798695,
231
+ "grad_norm": 1.3671875,
232
+ "learning_rate": 4.865217391304348e-05,
233
+ "loss": 1.0368,
234
+ "step": 32
235
+ },
236
+ {
237
+ "epoch": 0.28726877040261156,
238
+ "grad_norm": 1.140625,
239
+ "learning_rate": 4.860869565217391e-05,
240
+ "loss": 0.9539,
241
+ "step": 33
242
+ },
243
+ {
244
+ "epoch": 0.2959738846572361,
245
+ "grad_norm": 1.4296875,
246
+ "learning_rate": 4.856521739130435e-05,
247
+ "loss": 0.9165,
248
+ "step": 34
249
+ },
250
+ {
251
+ "epoch": 0.30467899891186073,
252
+ "grad_norm": 1.2265625,
253
+ "learning_rate": 4.8521739130434786e-05,
254
+ "loss": 1.03,
255
+ "step": 35
256
+ },
257
+ {
258
+ "epoch": 0.3133841131664853,
259
+ "grad_norm": 1.5703125,
260
+ "learning_rate": 4.847826086956522e-05,
261
+ "loss": 1.2848,
262
+ "step": 36
263
+ },
264
+ {
265
+ "epoch": 0.3220892274211099,
266
+ "grad_norm": 1.21875,
267
+ "learning_rate": 4.843478260869565e-05,
268
+ "loss": 0.8663,
269
+ "step": 37
270
+ },
271
+ {
272
+ "epoch": 0.3307943416757345,
273
+ "grad_norm": 1.125,
274
+ "learning_rate": 4.839130434782609e-05,
275
+ "loss": 0.9547,
276
+ "step": 38
277
+ },
278
+ {
279
+ "epoch": 0.3394994559303591,
280
+ "grad_norm": 1.328125,
281
+ "learning_rate": 4.834782608695652e-05,
282
+ "loss": 0.872,
283
+ "step": 39
284
+ },
285
+ {
286
+ "epoch": 0.3482045701849837,
287
+ "grad_norm": 1.140625,
288
+ "learning_rate": 4.830434782608696e-05,
289
+ "loss": 0.8424,
290
+ "step": 40
291
+ },
292
+ {
293
+ "epoch": 0.35690968443960824,
294
+ "grad_norm": 1.2578125,
295
+ "learning_rate": 4.8260869565217394e-05,
296
+ "loss": 0.8246,
297
+ "step": 41
298
+ },
299
+ {
300
+ "epoch": 0.36561479869423286,
301
+ "grad_norm": 1.3046875,
302
+ "learning_rate": 4.8217391304347834e-05,
303
+ "loss": 0.8326,
304
+ "step": 42
305
+ },
306
+ {
307
+ "epoch": 0.37431991294885747,
308
+ "grad_norm": 1.2109375,
309
+ "learning_rate": 4.817391304347826e-05,
310
+ "loss": 0.9043,
311
+ "step": 43
312
+ },
313
+ {
314
+ "epoch": 0.383025027203482,
315
+ "grad_norm": 1.296875,
316
+ "learning_rate": 4.8130434782608694e-05,
317
+ "loss": 0.8888,
318
+ "step": 44
319
+ },
320
+ {
321
+ "epoch": 0.39173014145810664,
322
+ "grad_norm": 1.2265625,
323
+ "learning_rate": 4.8086956521739134e-05,
324
+ "loss": 0.778,
325
+ "step": 45
326
+ },
327
+ {
328
+ "epoch": 0.40043525571273125,
329
+ "grad_norm": 1.0546875,
330
+ "learning_rate": 4.804347826086957e-05,
331
+ "loss": 0.8684,
332
+ "step": 46
333
+ },
334
+ {
335
+ "epoch": 0.4091403699673558,
336
+ "grad_norm": 1.5078125,
337
+ "learning_rate": 4.8e-05,
338
+ "loss": 0.9351,
339
+ "step": 47
340
+ },
341
+ {
342
+ "epoch": 0.4178454842219804,
343
+ "grad_norm": 1.2421875,
344
+ "learning_rate": 4.795652173913044e-05,
345
+ "loss": 0.8708,
346
+ "step": 48
347
+ },
348
+ {
349
+ "epoch": 0.426550598476605,
350
+ "grad_norm": 1.28125,
351
+ "learning_rate": 4.7913043478260875e-05,
352
+ "loss": 0.9919,
353
+ "step": 49
354
+ },
355
+ {
356
+ "epoch": 0.4352557127312296,
357
+ "grad_norm": 1.1953125,
358
+ "learning_rate": 4.78695652173913e-05,
359
+ "loss": 0.8166,
360
+ "step": 50
361
+ },
362
+ {
363
+ "epoch": 0.4439608269858542,
364
+ "grad_norm": 1.3203125,
365
+ "learning_rate": 4.782608695652174e-05,
366
+ "loss": 0.9542,
367
+ "step": 51
368
+ },
369
+ {
370
+ "epoch": 0.45266594124047876,
371
+ "grad_norm": 1.2109375,
372
+ "learning_rate": 4.7782608695652175e-05,
373
+ "loss": 1.0959,
374
+ "step": 52
375
+ },
376
+ {
377
+ "epoch": 0.4613710554951034,
378
+ "grad_norm": 1.140625,
379
+ "learning_rate": 4.773913043478261e-05,
380
+ "loss": 0.8695,
381
+ "step": 53
382
+ },
383
+ {
384
+ "epoch": 0.470076169749728,
385
+ "grad_norm": 1.5234375,
386
+ "learning_rate": 4.769565217391305e-05,
387
+ "loss": 1.1411,
388
+ "step": 54
389
+ },
390
+ {
391
+ "epoch": 0.47878128400435255,
392
+ "grad_norm": 1.125,
393
+ "learning_rate": 4.765217391304348e-05,
394
+ "loss": 0.8354,
395
+ "step": 55
396
+ },
397
+ {
398
+ "epoch": 0.48748639825897716,
399
+ "grad_norm": 1.390625,
400
+ "learning_rate": 4.7608695652173916e-05,
401
+ "loss": 0.9133,
402
+ "step": 56
403
+ },
404
+ {
405
+ "epoch": 0.4961915125136017,
406
+ "grad_norm": 1.2578125,
407
+ "learning_rate": 4.756521739130435e-05,
408
+ "loss": 0.9042,
409
+ "step": 57
410
+ },
411
+ {
412
+ "epoch": 0.5048966267682263,
413
+ "grad_norm": 1.28125,
414
+ "learning_rate": 4.752173913043478e-05,
415
+ "loss": 0.9707,
416
+ "step": 58
417
+ },
418
+ {
419
+ "epoch": 0.5136017410228509,
420
+ "grad_norm": 1.2734375,
421
+ "learning_rate": 4.747826086956522e-05,
422
+ "loss": 0.7845,
423
+ "step": 59
424
+ },
425
+ {
426
+ "epoch": 0.5223068552774756,
427
+ "grad_norm": 1.3046875,
428
+ "learning_rate": 4.7434782608695656e-05,
429
+ "loss": 0.955,
430
+ "step": 60
431
+ },
432
+ {
433
+ "epoch": 0.5310119695321001,
434
+ "grad_norm": 1.546875,
435
+ "learning_rate": 4.739130434782609e-05,
436
+ "loss": 0.9521,
437
+ "step": 61
438
+ },
439
+ {
440
+ "epoch": 0.5397170837867247,
441
+ "grad_norm": 1.2578125,
442
+ "learning_rate": 4.734782608695652e-05,
443
+ "loss": 1.0155,
444
+ "step": 62
445
+ },
446
+ {
447
+ "epoch": 0.5484221980413493,
448
+ "grad_norm": 1.3125,
449
+ "learning_rate": 4.7304347826086956e-05,
450
+ "loss": 1.0093,
451
+ "step": 63
452
+ },
453
+ {
454
+ "epoch": 0.5571273122959739,
455
+ "grad_norm": 1.234375,
456
+ "learning_rate": 4.726086956521739e-05,
457
+ "loss": 0.8372,
458
+ "step": 64
459
+ },
460
+ {
461
+ "epoch": 0.5658324265505985,
462
+ "grad_norm": 1.0703125,
463
+ "learning_rate": 4.721739130434783e-05,
464
+ "loss": 0.8776,
465
+ "step": 65
466
+ },
467
+ {
468
+ "epoch": 0.5745375408052231,
469
+ "grad_norm": 1.2265625,
470
+ "learning_rate": 4.7173913043478264e-05,
471
+ "loss": 0.8854,
472
+ "step": 66
473
+ },
474
+ {
475
+ "epoch": 0.5832426550598476,
476
+ "grad_norm": 1.1171875,
477
+ "learning_rate": 4.71304347826087e-05,
478
+ "loss": 0.8966,
479
+ "step": 67
480
+ },
481
+ {
482
+ "epoch": 0.5919477693144722,
483
+ "grad_norm": 1.296875,
484
+ "learning_rate": 4.708695652173914e-05,
485
+ "loss": 0.9694,
486
+ "step": 68
487
+ },
488
+ {
489
+ "epoch": 0.6006528835690969,
490
+ "grad_norm": 1.1875,
491
+ "learning_rate": 4.7043478260869564e-05,
492
+ "loss": 1.062,
493
+ "step": 69
494
+ },
495
+ {
496
+ "epoch": 0.6093579978237215,
497
+ "grad_norm": 1.078125,
498
+ "learning_rate": 4.7e-05,
499
+ "loss": 0.8369,
500
+ "step": 70
501
+ },
502
+ {
503
+ "epoch": 0.6180631120783461,
504
+ "grad_norm": 1.1953125,
505
+ "learning_rate": 4.695652173913044e-05,
506
+ "loss": 0.7927,
507
+ "step": 71
508
+ },
509
+ {
510
+ "epoch": 0.6267682263329706,
511
+ "grad_norm": 1.4140625,
512
+ "learning_rate": 4.691304347826087e-05,
513
+ "loss": 1.3626,
514
+ "step": 72
515
+ },
516
+ {
517
+ "epoch": 0.6354733405875952,
518
+ "grad_norm": 1.1328125,
519
+ "learning_rate": 4.686956521739131e-05,
520
+ "loss": 0.7737,
521
+ "step": 73
522
+ },
523
+ {
524
+ "epoch": 0.6441784548422198,
525
+ "grad_norm": 1.0859375,
526
+ "learning_rate": 4.6826086956521745e-05,
527
+ "loss": 0.9044,
528
+ "step": 74
529
+ },
530
+ {
531
+ "epoch": 0.6528835690968444,
532
+ "grad_norm": 1.078125,
533
+ "learning_rate": 4.678260869565218e-05,
534
+ "loss": 0.9451,
535
+ "step": 75
536
+ },
537
+ {
538
+ "epoch": 0.661588683351469,
539
+ "grad_norm": 1.234375,
540
+ "learning_rate": 4.673913043478261e-05,
541
+ "loss": 1.0073,
542
+ "step": 76
543
+ },
544
+ {
545
+ "epoch": 0.6702937976060935,
546
+ "grad_norm": 1.1484375,
547
+ "learning_rate": 4.6695652173913045e-05,
548
+ "loss": 1.0206,
549
+ "step": 77
550
+ },
551
+ {
552
+ "epoch": 0.6789989118607181,
553
+ "grad_norm": 1.25,
554
+ "learning_rate": 4.665217391304348e-05,
555
+ "loss": 0.7854,
556
+ "step": 78
557
+ },
558
+ {
559
+ "epoch": 0.6877040261153428,
560
+ "grad_norm": 1.0859375,
561
+ "learning_rate": 4.660869565217392e-05,
562
+ "loss": 0.9661,
563
+ "step": 79
564
+ },
565
+ {
566
+ "epoch": 0.6964091403699674,
567
+ "grad_norm": 1.296875,
568
+ "learning_rate": 4.656521739130435e-05,
569
+ "loss": 1.0657,
570
+ "step": 80
571
+ },
572
+ {
573
+ "epoch": 0.705114254624592,
574
+ "grad_norm": 1.0390625,
575
+ "learning_rate": 4.6521739130434785e-05,
576
+ "loss": 0.6992,
577
+ "step": 81
578
+ },
579
+ {
580
+ "epoch": 0.7138193688792165,
581
+ "grad_norm": 1.0703125,
582
+ "learning_rate": 4.647826086956522e-05,
583
+ "loss": 0.8135,
584
+ "step": 82
585
+ },
586
+ {
587
+ "epoch": 0.7225244831338411,
588
+ "grad_norm": 1.359375,
589
+ "learning_rate": 4.643478260869565e-05,
590
+ "loss": 1.1033,
591
+ "step": 83
592
+ },
593
+ {
594
+ "epoch": 0.7312295973884657,
595
+ "grad_norm": 1.3828125,
596
+ "learning_rate": 4.6391304347826086e-05,
597
+ "loss": 0.8077,
598
+ "step": 84
599
+ },
600
+ {
601
+ "epoch": 0.7399347116430903,
602
+ "grad_norm": 1.0859375,
603
+ "learning_rate": 4.6347826086956526e-05,
604
+ "loss": 0.8582,
605
+ "step": 85
606
+ },
607
+ {
608
+ "epoch": 0.7486398258977149,
609
+ "grad_norm": 1.140625,
610
+ "learning_rate": 4.630434782608696e-05,
611
+ "loss": 1.0223,
612
+ "step": 86
613
+ },
614
+ {
615
+ "epoch": 0.7573449401523396,
616
+ "grad_norm": 1.2734375,
617
+ "learning_rate": 4.62608695652174e-05,
618
+ "loss": 0.7866,
619
+ "step": 87
620
+ },
621
+ {
622
+ "epoch": 0.766050054406964,
623
+ "grad_norm": 1.1484375,
624
+ "learning_rate": 4.6217391304347826e-05,
625
+ "loss": 0.804,
626
+ "step": 88
627
+ },
628
+ {
629
+ "epoch": 0.7747551686615887,
630
+ "grad_norm": 0.96484375,
631
+ "learning_rate": 4.617391304347826e-05,
632
+ "loss": 0.681,
633
+ "step": 89
634
+ },
635
+ {
636
+ "epoch": 0.7834602829162133,
637
+ "grad_norm": 1.0078125,
638
+ "learning_rate": 4.61304347826087e-05,
639
+ "loss": 0.9594,
640
+ "step": 90
641
+ },
642
+ {
643
+ "epoch": 0.7921653971708379,
644
+ "grad_norm": 1.0546875,
645
+ "learning_rate": 4.608695652173913e-05,
646
+ "loss": 0.9842,
647
+ "step": 91
648
+ },
649
+ {
650
+ "epoch": 0.8008705114254625,
651
+ "grad_norm": 1.09375,
652
+ "learning_rate": 4.6043478260869567e-05,
653
+ "loss": 0.9142,
654
+ "step": 92
655
+ },
656
+ {
657
+ "epoch": 0.809575625680087,
658
+ "grad_norm": 1.15625,
659
+ "learning_rate": 4.600000000000001e-05,
660
+ "loss": 0.7379,
661
+ "step": 93
662
+ },
663
+ {
664
+ "epoch": 0.8182807399347116,
665
+ "grad_norm": 1.3984375,
666
+ "learning_rate": 4.595652173913044e-05,
667
+ "loss": 1.0656,
668
+ "step": 94
669
+ },
670
+ {
671
+ "epoch": 0.8269858541893362,
672
+ "grad_norm": 1.15625,
673
+ "learning_rate": 4.591304347826087e-05,
674
+ "loss": 0.9404,
675
+ "step": 95
676
+ },
677
+ {
678
+ "epoch": 0.8356909684439608,
679
+ "grad_norm": 1.109375,
680
+ "learning_rate": 4.586956521739131e-05,
681
+ "loss": 1.0705,
682
+ "step": 96
683
+ },
684
+ {
685
+ "epoch": 0.8443960826985855,
686
+ "grad_norm": 1.1171875,
687
+ "learning_rate": 4.582608695652174e-05,
688
+ "loss": 1.1848,
689
+ "step": 97
690
+ },
691
+ {
692
+ "epoch": 0.85310119695321,
693
+ "grad_norm": 1.15625,
694
+ "learning_rate": 4.5782608695652174e-05,
695
+ "loss": 0.8771,
696
+ "step": 98
697
+ },
698
+ {
699
+ "epoch": 0.8618063112078346,
700
+ "grad_norm": 1.265625,
701
+ "learning_rate": 4.5739130434782614e-05,
702
+ "loss": 0.9586,
703
+ "step": 99
704
+ },
705
+ {
706
+ "epoch": 0.8705114254624592,
707
+ "grad_norm": 1.1015625,
708
+ "learning_rate": 4.569565217391305e-05,
709
+ "loss": 0.7516,
710
+ "step": 100
711
+ },
712
+ {
713
+ "epoch": 0.8792165397170838,
714
+ "grad_norm": 1.3125,
715
+ "learning_rate": 4.565217391304348e-05,
716
+ "loss": 0.8131,
717
+ "step": 101
718
+ },
719
+ {
720
+ "epoch": 0.8879216539717084,
721
+ "grad_norm": 1.109375,
722
+ "learning_rate": 4.5608695652173914e-05,
723
+ "loss": 0.9118,
724
+ "step": 102
725
+ },
726
+ {
727
+ "epoch": 0.8966267682263329,
728
+ "grad_norm": 1.4140625,
729
+ "learning_rate": 4.556521739130435e-05,
730
+ "loss": 0.8855,
731
+ "step": 103
732
+ },
733
+ {
734
+ "epoch": 0.9053318824809575,
735
+ "grad_norm": 1.1796875,
736
+ "learning_rate": 4.552173913043479e-05,
737
+ "loss": 0.7721,
738
+ "step": 104
739
+ },
740
+ {
741
+ "epoch": 0.9140369967355821,
742
+ "grad_norm": 1.2265625,
743
+ "learning_rate": 4.547826086956522e-05,
744
+ "loss": 0.7768,
745
+ "step": 105
746
+ },
747
+ {
748
+ "epoch": 0.9227421109902068,
749
+ "grad_norm": 1.0703125,
750
+ "learning_rate": 4.5434782608695655e-05,
751
+ "loss": 0.9557,
752
+ "step": 106
753
+ },
754
+ {
755
+ "epoch": 0.9314472252448314,
756
+ "grad_norm": 1.0625,
757
+ "learning_rate": 4.539130434782609e-05,
758
+ "loss": 0.8868,
759
+ "step": 107
760
+ },
761
+ {
762
+ "epoch": 0.940152339499456,
763
+ "grad_norm": 1.15625,
764
+ "learning_rate": 4.534782608695652e-05,
765
+ "loss": 0.8889,
766
+ "step": 108
767
+ },
768
+ {
769
+ "epoch": 0.9488574537540805,
770
+ "grad_norm": 1.15625,
771
+ "learning_rate": 4.5304347826086955e-05,
772
+ "loss": 1.1122,
773
+ "step": 109
774
+ },
775
+ {
776
+ "epoch": 0.9575625680087051,
777
+ "grad_norm": 1.1875,
778
+ "learning_rate": 4.5260869565217395e-05,
779
+ "loss": 0.7954,
780
+ "step": 110
781
+ },
782
+ {
783
+ "epoch": 0.9662676822633297,
784
+ "grad_norm": 1.1796875,
785
+ "learning_rate": 4.521739130434783e-05,
786
+ "loss": 0.8551,
787
+ "step": 111
788
+ },
789
+ {
790
+ "epoch": 0.9749727965179543,
791
+ "grad_norm": 1.1953125,
792
+ "learning_rate": 4.517391304347826e-05,
793
+ "loss": 1.0285,
794
+ "step": 112
795
+ },
796
+ {
797
+ "epoch": 0.9836779107725789,
798
+ "grad_norm": 1.25,
799
+ "learning_rate": 4.51304347826087e-05,
800
+ "loss": 1.0737,
801
+ "step": 113
802
+ },
803
+ {
804
+ "epoch": 0.9923830250272034,
805
+ "grad_norm": 1.015625,
806
+ "learning_rate": 4.508695652173913e-05,
807
+ "loss": 0.8448,
808
+ "step": 114
809
+ },
810
+ {
811
+ "epoch": 1.0,
812
+ "grad_norm": 1.2421875,
813
+ "learning_rate": 4.504347826086956e-05,
814
+ "loss": 0.9171,
815
+ "step": 115
816
+ },
817
+ {
818
+ "epoch": 1.0087051142546246,
819
+ "grad_norm": 1.234375,
820
+ "learning_rate": 4.5e-05,
821
+ "loss": 0.7003,
822
+ "step": 116
823
+ },
824
+ {
825
+ "epoch": 1.0174102285092492,
826
+ "grad_norm": 1.5078125,
827
+ "learning_rate": 4.4956521739130436e-05,
828
+ "loss": 0.7938,
829
+ "step": 117
830
+ },
831
+ {
832
+ "epoch": 1.0261153427638738,
833
+ "grad_norm": 1.09375,
834
+ "learning_rate": 4.4913043478260876e-05,
835
+ "loss": 0.7323,
836
+ "step": 118
837
+ },
838
+ {
839
+ "epoch": 1.0348204570184985,
840
+ "grad_norm": 1.2421875,
841
+ "learning_rate": 4.486956521739131e-05,
842
+ "loss": 0.5957,
843
+ "step": 119
844
+ },
845
+ {
846
+ "epoch": 1.043525571273123,
847
+ "grad_norm": 1.453125,
848
+ "learning_rate": 4.482608695652174e-05,
849
+ "loss": 0.7519,
850
+ "step": 120
851
+ },
852
+ {
853
+ "epoch": 1.0522306855277475,
854
+ "grad_norm": 1.4453125,
855
+ "learning_rate": 4.478260869565218e-05,
856
+ "loss": 0.7003,
857
+ "step": 121
858
+ },
859
+ {
860
+ "epoch": 1.060935799782372,
861
+ "grad_norm": 1.4296875,
862
+ "learning_rate": 4.473913043478261e-05,
863
+ "loss": 0.8537,
864
+ "step": 122
865
+ },
866
+ {
867
+ "epoch": 1.0696409140369967,
868
+ "grad_norm": 1.2265625,
869
+ "learning_rate": 4.4695652173913044e-05,
870
+ "loss": 0.7305,
871
+ "step": 123
872
+ },
873
+ {
874
+ "epoch": 1.0783460282916213,
875
+ "grad_norm": 1.3203125,
876
+ "learning_rate": 4.4652173913043484e-05,
877
+ "loss": 0.6122,
878
+ "step": 124
879
+ },
880
+ {
881
+ "epoch": 1.087051142546246,
882
+ "grad_norm": 1.296875,
883
+ "learning_rate": 4.460869565217392e-05,
884
+ "loss": 0.9224,
885
+ "step": 125
886
+ },
887
+ {
888
+ "epoch": 1.0957562568008705,
889
+ "grad_norm": 0.94921875,
890
+ "learning_rate": 4.456521739130435e-05,
891
+ "loss": 0.6261,
892
+ "step": 126
893
+ },
894
+ {
895
+ "epoch": 1.1044613710554951,
896
+ "grad_norm": 1.125,
897
+ "learning_rate": 4.4521739130434784e-05,
898
+ "loss": 0.7336,
899
+ "step": 127
900
+ },
901
+ {
902
+ "epoch": 1.1131664853101197,
903
+ "grad_norm": 1.125,
904
+ "learning_rate": 4.447826086956522e-05,
905
+ "loss": 0.6978,
906
+ "step": 128
907
+ },
908
+ {
909
+ "epoch": 1.1218715995647444,
910
+ "grad_norm": 1.125,
911
+ "learning_rate": 4.443478260869565e-05,
912
+ "loss": 0.6352,
913
+ "step": 129
914
+ },
915
+ {
916
+ "epoch": 1.130576713819369,
917
+ "grad_norm": 1.125,
918
+ "learning_rate": 4.439130434782609e-05,
919
+ "loss": 0.4918,
920
+ "step": 130
921
+ },
922
+ {
923
+ "epoch": 1.1392818280739934,
924
+ "grad_norm": 1.25,
925
+ "learning_rate": 4.4347826086956525e-05,
926
+ "loss": 0.9074,
927
+ "step": 131
928
+ },
929
+ {
930
+ "epoch": 1.147986942328618,
931
+ "grad_norm": 1.296875,
932
+ "learning_rate": 4.430434782608696e-05,
933
+ "loss": 0.7653,
934
+ "step": 132
935
+ },
936
+ {
937
+ "epoch": 1.1566920565832426,
938
+ "grad_norm": 1.046875,
939
+ "learning_rate": 4.426086956521739e-05,
940
+ "loss": 0.6195,
941
+ "step": 133
942
+ },
943
+ {
944
+ "epoch": 1.1653971708378672,
945
+ "grad_norm": 1.1953125,
946
+ "learning_rate": 4.4217391304347825e-05,
947
+ "loss": 0.5887,
948
+ "step": 134
949
+ },
950
+ {
951
+ "epoch": 1.1741022850924918,
952
+ "grad_norm": 2.15625,
953
+ "learning_rate": 4.4173913043478265e-05,
954
+ "loss": 0.7727,
955
+ "step": 135
956
+ },
957
+ {
958
+ "epoch": 1.1828073993471164,
959
+ "grad_norm": 1.0703125,
960
+ "learning_rate": 4.41304347826087e-05,
961
+ "loss": 0.598,
962
+ "step": 136
963
+ },
964
+ {
965
+ "epoch": 1.191512513601741,
966
+ "grad_norm": 1.2265625,
967
+ "learning_rate": 4.408695652173913e-05,
968
+ "loss": 0.7572,
969
+ "step": 137
970
+ },
971
+ {
972
+ "epoch": 1.2002176278563657,
973
+ "grad_norm": 1.2265625,
974
+ "learning_rate": 4.404347826086957e-05,
975
+ "loss": 0.6791,
976
+ "step": 138
977
+ },
978
+ {
979
+ "epoch": 1.2089227421109903,
980
+ "grad_norm": 1.125,
981
+ "learning_rate": 4.4000000000000006e-05,
982
+ "loss": 0.7195,
983
+ "step": 139
984
+ },
985
+ {
986
+ "epoch": 1.2176278563656149,
987
+ "grad_norm": 0.984375,
988
+ "learning_rate": 4.395652173913043e-05,
989
+ "loss": 0.5021,
990
+ "step": 140
991
+ },
992
+ {
993
+ "epoch": 1.2263329706202395,
994
+ "grad_norm": 1.2578125,
995
+ "learning_rate": 4.391304347826087e-05,
996
+ "loss": 0.6567,
997
+ "step": 141
998
+ },
999
+ {
1000
+ "epoch": 1.2350380848748639,
1001
+ "grad_norm": 1.1875,
1002
+ "learning_rate": 4.3869565217391306e-05,
1003
+ "loss": 0.8551,
1004
+ "step": 142
1005
+ },
1006
+ {
1007
+ "epoch": 1.2437431991294885,
1008
+ "grad_norm": 1.046875,
1009
+ "learning_rate": 4.382608695652174e-05,
1010
+ "loss": 0.7298,
1011
+ "step": 143
1012
+ },
1013
+ {
1014
+ "epoch": 1.252448313384113,
1015
+ "grad_norm": 1.3203125,
1016
+ "learning_rate": 4.378260869565218e-05,
1017
+ "loss": 0.9639,
1018
+ "step": 144
1019
+ },
1020
+ {
1021
+ "epoch": 1.2611534276387377,
1022
+ "grad_norm": 1.0546875,
1023
+ "learning_rate": 4.373913043478261e-05,
1024
+ "loss": 0.7321,
1025
+ "step": 145
1026
+ },
1027
+ {
1028
+ "epoch": 1.2698585418933623,
1029
+ "grad_norm": 0.9921875,
1030
+ "learning_rate": 4.3695652173913046e-05,
1031
+ "loss": 0.6488,
1032
+ "step": 146
1033
+ },
1034
+ {
1035
+ "epoch": 1.278563656147987,
1036
+ "grad_norm": 1.0703125,
1037
+ "learning_rate": 4.365217391304348e-05,
1038
+ "loss": 0.7783,
1039
+ "step": 147
1040
+ },
1041
+ {
1042
+ "epoch": 1.2872687704026116,
1043
+ "grad_norm": 1.1484375,
1044
+ "learning_rate": 4.360869565217391e-05,
1045
+ "loss": 0.6042,
1046
+ "step": 148
1047
+ },
1048
+ {
1049
+ "epoch": 1.2959738846572362,
1050
+ "grad_norm": 1.09375,
1051
+ "learning_rate": 4.3565217391304353e-05,
1052
+ "loss": 0.7951,
1053
+ "step": 149
1054
+ },
1055
+ {
1056
+ "epoch": 1.3046789989118608,
1057
+ "grad_norm": 1.09375,
1058
+ "learning_rate": 4.352173913043479e-05,
1059
+ "loss": 0.6017,
1060
+ "step": 150
1061
+ },
1062
+ {
1063
+ "epoch": 1.3133841131664852,
1064
+ "grad_norm": 1.2109375,
1065
+ "learning_rate": 4.347826086956522e-05,
1066
+ "loss": 0.6653,
1067
+ "step": 151
1068
+ },
1069
+ {
1070
+ "epoch": 1.32208922742111,
1071
+ "grad_norm": 1.3828125,
1072
+ "learning_rate": 4.3434782608695654e-05,
1073
+ "loss": 0.9719,
1074
+ "step": 152
1075
+ },
1076
+ {
1077
+ "epoch": 1.3307943416757344,
1078
+ "grad_norm": 1.015625,
1079
+ "learning_rate": 4.339130434782609e-05,
1080
+ "loss": 0.6013,
1081
+ "step": 153
1082
+ },
1083
+ {
1084
+ "epoch": 1.339499455930359,
1085
+ "grad_norm": 1.1015625,
1086
+ "learning_rate": 4.334782608695652e-05,
1087
+ "loss": 0.6483,
1088
+ "step": 154
1089
+ },
1090
+ {
1091
+ "epoch": 1.3482045701849836,
1092
+ "grad_norm": 1.0859375,
1093
+ "learning_rate": 4.330434782608696e-05,
1094
+ "loss": 0.6805,
1095
+ "step": 155
1096
+ },
1097
+ {
1098
+ "epoch": 1.3569096844396082,
1099
+ "grad_norm": 1.140625,
1100
+ "learning_rate": 4.3260869565217394e-05,
1101
+ "loss": 0.7598,
1102
+ "step": 156
1103
+ },
1104
+ {
1105
+ "epoch": 1.3656147986942329,
1106
+ "grad_norm": 1.046875,
1107
+ "learning_rate": 4.321739130434783e-05,
1108
+ "loss": 0.569,
1109
+ "step": 157
1110
+ },
1111
+ {
1112
+ "epoch": 1.3743199129488575,
1113
+ "grad_norm": 1.0625,
1114
+ "learning_rate": 4.317391304347827e-05,
1115
+ "loss": 0.7223,
1116
+ "step": 158
1117
+ },
1118
+ {
1119
+ "epoch": 1.383025027203482,
1120
+ "grad_norm": 1.1328125,
1121
+ "learning_rate": 4.3130434782608695e-05,
1122
+ "loss": 0.7309,
1123
+ "step": 159
1124
+ },
1125
+ {
1126
+ "epoch": 1.3917301414581067,
1127
+ "grad_norm": 1.21875,
1128
+ "learning_rate": 4.308695652173913e-05,
1129
+ "loss": 0.9004,
1130
+ "step": 160
1131
+ },
1132
+ {
1133
+ "epoch": 1.4004352557127313,
1134
+ "grad_norm": 1.125,
1135
+ "learning_rate": 4.304347826086957e-05,
1136
+ "loss": 0.6035,
1137
+ "step": 161
1138
+ },
1139
+ {
1140
+ "epoch": 1.4091403699673557,
1141
+ "grad_norm": 1.2578125,
1142
+ "learning_rate": 4.3e-05,
1143
+ "loss": 0.957,
1144
+ "step": 162
1145
+ },
1146
+ {
1147
+ "epoch": 1.4178454842219805,
1148
+ "grad_norm": 2.6875,
1149
+ "learning_rate": 4.2956521739130435e-05,
1150
+ "loss": 0.5841,
1151
+ "step": 163
1152
+ },
1153
+ {
1154
+ "epoch": 1.426550598476605,
1155
+ "grad_norm": 1.1640625,
1156
+ "learning_rate": 4.2913043478260875e-05,
1157
+ "loss": 0.8018,
1158
+ "step": 164
1159
+ },
1160
+ {
1161
+ "epoch": 1.4352557127312295,
1162
+ "grad_norm": 1.078125,
1163
+ "learning_rate": 4.286956521739131e-05,
1164
+ "loss": 0.6603,
1165
+ "step": 165
1166
+ },
1167
+ {
1168
+ "epoch": 1.4439608269858542,
1169
+ "grad_norm": 1.40625,
1170
+ "learning_rate": 4.282608695652174e-05,
1171
+ "loss": 0.7107,
1172
+ "step": 166
1173
+ },
1174
+ {
1175
+ "epoch": 1.4526659412404788,
1176
+ "grad_norm": 1.0390625,
1177
+ "learning_rate": 4.2782608695652176e-05,
1178
+ "loss": 0.7577,
1179
+ "step": 167
1180
+ },
1181
+ {
1182
+ "epoch": 1.4613710554951034,
1183
+ "grad_norm": 1.0859375,
1184
+ "learning_rate": 4.273913043478261e-05,
1185
+ "loss": 0.7573,
1186
+ "step": 168
1187
+ },
1188
+ {
1189
+ "epoch": 1.470076169749728,
1190
+ "grad_norm": 1.265625,
1191
+ "learning_rate": 4.269565217391305e-05,
1192
+ "loss": 0.7008,
1193
+ "step": 169
1194
+ },
1195
+ {
1196
+ "epoch": 1.4787812840043526,
1197
+ "grad_norm": 1.1953125,
1198
+ "learning_rate": 4.265217391304348e-05,
1199
+ "loss": 0.6966,
1200
+ "step": 170
1201
+ },
1202
+ {
1203
+ "epoch": 1.4874863982589772,
1204
+ "grad_norm": 1.1015625,
1205
+ "learning_rate": 4.2608695652173916e-05,
1206
+ "loss": 0.6855,
1207
+ "step": 171
1208
+ },
1209
+ {
1210
+ "epoch": 1.4961915125136018,
1211
+ "grad_norm": 1.109375,
1212
+ "learning_rate": 4.256521739130435e-05,
1213
+ "loss": 0.67,
1214
+ "step": 172
1215
+ },
1216
+ {
1217
+ "epoch": 1.5048966267682262,
1218
+ "grad_norm": 1.15625,
1219
+ "learning_rate": 4.252173913043478e-05,
1220
+ "loss": 0.8798,
1221
+ "step": 173
1222
+ },
1223
+ {
1224
+ "epoch": 1.513601741022851,
1225
+ "grad_norm": 1.15625,
1226
+ "learning_rate": 4.2478260869565216e-05,
1227
+ "loss": 0.7921,
1228
+ "step": 174
1229
+ },
1230
+ {
1231
+ "epoch": 1.5223068552774754,
1232
+ "grad_norm": 1.1171875,
1233
+ "learning_rate": 4.2434782608695657e-05,
1234
+ "loss": 0.6351,
1235
+ "step": 175
1236
+ },
1237
+ {
1238
+ "epoch": 1.5310119695321,
1239
+ "grad_norm": 1.609375,
1240
+ "learning_rate": 4.239130434782609e-05,
1241
+ "loss": 0.6681,
1242
+ "step": 176
1243
+ },
1244
+ {
1245
+ "epoch": 1.5397170837867247,
1246
+ "grad_norm": 1.265625,
1247
+ "learning_rate": 4.2347826086956523e-05,
1248
+ "loss": 0.7897,
1249
+ "step": 177
1250
+ },
1251
+ {
1252
+ "epoch": 1.5484221980413493,
1253
+ "grad_norm": 1.2734375,
1254
+ "learning_rate": 4.230434782608696e-05,
1255
+ "loss": 0.772,
1256
+ "step": 178
1257
+ },
1258
+ {
1259
+ "epoch": 1.557127312295974,
1260
+ "grad_norm": 1.0546875,
1261
+ "learning_rate": 4.226086956521739e-05,
1262
+ "loss": 0.5812,
1263
+ "step": 179
1264
+ },
1265
+ {
1266
+ "epoch": 1.5658324265505985,
1267
+ "grad_norm": 1.0703125,
1268
+ "learning_rate": 4.221739130434783e-05,
1269
+ "loss": 0.5453,
1270
+ "step": 180
1271
+ },
1272
+ {
1273
+ "epoch": 1.5745375408052231,
1274
+ "grad_norm": 1.0078125,
1275
+ "learning_rate": 4.2173913043478264e-05,
1276
+ "loss": 0.6831,
1277
+ "step": 181
1278
+ },
1279
+ {
1280
+ "epoch": 1.5832426550598475,
1281
+ "grad_norm": 1.0625,
1282
+ "learning_rate": 4.21304347826087e-05,
1283
+ "loss": 0.767,
1284
+ "step": 182
1285
+ },
1286
+ {
1287
+ "epoch": 1.5919477693144723,
1288
+ "grad_norm": 1.1484375,
1289
+ "learning_rate": 4.208695652173914e-05,
1290
+ "loss": 0.6145,
1291
+ "step": 183
1292
+ },
1293
+ {
1294
+ "epoch": 1.6006528835690967,
1295
+ "grad_norm": 1.1015625,
1296
+ "learning_rate": 4.204347826086957e-05,
1297
+ "loss": 0.8171,
1298
+ "step": 184
1299
+ },
1300
+ {
1301
+ "epoch": 1.6093579978237216,
1302
+ "grad_norm": 1.4609375,
1303
+ "learning_rate": 4.2e-05,
1304
+ "loss": 0.6039,
1305
+ "step": 185
1306
+ },
1307
+ {
1308
+ "epoch": 1.618063112078346,
1309
+ "grad_norm": 1.1171875,
1310
+ "learning_rate": 4.195652173913044e-05,
1311
+ "loss": 0.7126,
1312
+ "step": 186
1313
+ },
1314
+ {
1315
+ "epoch": 1.6267682263329706,
1316
+ "grad_norm": 1.1328125,
1317
+ "learning_rate": 4.191304347826087e-05,
1318
+ "loss": 0.5621,
1319
+ "step": 187
1320
+ },
1321
+ {
1322
+ "epoch": 1.6354733405875952,
1323
+ "grad_norm": 1.0703125,
1324
+ "learning_rate": 4.1869565217391305e-05,
1325
+ "loss": 0.7605,
1326
+ "step": 188
1327
+ },
1328
+ {
1329
+ "epoch": 1.6441784548422198,
1330
+ "grad_norm": 1.0625,
1331
+ "learning_rate": 4.1826086956521745e-05,
1332
+ "loss": 0.8704,
1333
+ "step": 189
1334
+ },
1335
+ {
1336
+ "epoch": 1.6528835690968444,
1337
+ "grad_norm": 1.359375,
1338
+ "learning_rate": 4.178260869565218e-05,
1339
+ "loss": 0.7108,
1340
+ "step": 190
1341
+ },
1342
+ {
1343
+ "epoch": 1.661588683351469,
1344
+ "grad_norm": 1.0859375,
1345
+ "learning_rate": 4.1739130434782605e-05,
1346
+ "loss": 0.67,
1347
+ "step": 191
1348
+ },
1349
+ {
1350
+ "epoch": 1.6702937976060936,
1351
+ "grad_norm": 1.1171875,
1352
+ "learning_rate": 4.1695652173913045e-05,
1353
+ "loss": 0.9325,
1354
+ "step": 192
1355
+ },
1356
+ {
1357
+ "epoch": 1.678998911860718,
1358
+ "grad_norm": 1.2578125,
1359
+ "learning_rate": 4.165217391304348e-05,
1360
+ "loss": 0.8226,
1361
+ "step": 193
1362
+ },
1363
+ {
1364
+ "epoch": 1.6877040261153429,
1365
+ "grad_norm": 1.03125,
1366
+ "learning_rate": 4.160869565217391e-05,
1367
+ "loss": 0.8652,
1368
+ "step": 194
1369
+ },
1370
+ {
1371
+ "epoch": 1.6964091403699673,
1372
+ "grad_norm": 1.171875,
1373
+ "learning_rate": 4.156521739130435e-05,
1374
+ "loss": 0.6873,
1375
+ "step": 195
1376
+ },
1377
+ {
1378
+ "epoch": 1.705114254624592,
1379
+ "grad_norm": 1.1953125,
1380
+ "learning_rate": 4.1521739130434786e-05,
1381
+ "loss": 0.7681,
1382
+ "step": 196
1383
+ },
1384
+ {
1385
+ "epoch": 1.7138193688792165,
1386
+ "grad_norm": 1.0625,
1387
+ "learning_rate": 4.147826086956522e-05,
1388
+ "loss": 0.6852,
1389
+ "step": 197
1390
+ },
1391
+ {
1392
+ "epoch": 1.722524483133841,
1393
+ "grad_norm": 1.0546875,
1394
+ "learning_rate": 4.143478260869565e-05,
1395
+ "loss": 0.8035,
1396
+ "step": 198
1397
+ },
1398
+ {
1399
+ "epoch": 1.7312295973884657,
1400
+ "grad_norm": 1.140625,
1401
+ "learning_rate": 4.1391304347826086e-05,
1402
+ "loss": 0.6638,
1403
+ "step": 199
1404
+ },
1405
+ {
1406
+ "epoch": 1.7399347116430903,
1407
+ "grad_norm": 1.2265625,
1408
+ "learning_rate": 4.1347826086956526e-05,
1409
+ "loss": 0.8636,
1410
+ "step": 200
1411
+ },
1412
+ {
1413
+ "epoch": 1.748639825897715,
1414
+ "grad_norm": 1.140625,
1415
+ "learning_rate": 4.130434782608696e-05,
1416
+ "loss": 0.9844,
1417
+ "step": 201
1418
+ },
1419
+ {
1420
+ "epoch": 1.7573449401523396,
1421
+ "grad_norm": 1.1796875,
1422
+ "learning_rate": 4.126086956521739e-05,
1423
+ "loss": 0.7461,
1424
+ "step": 202
1425
+ },
1426
+ {
1427
+ "epoch": 1.7660500544069642,
1428
+ "grad_norm": 1.1484375,
1429
+ "learning_rate": 4.1217391304347827e-05,
1430
+ "loss": 0.677,
1431
+ "step": 203
1432
+ },
1433
+ {
1434
+ "epoch": 1.7747551686615886,
1435
+ "grad_norm": 1.234375,
1436
+ "learning_rate": 4.117391304347826e-05,
1437
+ "loss": 0.5882,
1438
+ "step": 204
1439
+ },
1440
+ {
1441
+ "epoch": 1.7834602829162134,
1442
+ "grad_norm": 1.2109375,
1443
+ "learning_rate": 4.1130434782608693e-05,
1444
+ "loss": 0.7051,
1445
+ "step": 205
1446
+ },
1447
+ {
1448
+ "epoch": 1.7921653971708378,
1449
+ "grad_norm": 1.2109375,
1450
+ "learning_rate": 4.1086956521739134e-05,
1451
+ "loss": 0.7658,
1452
+ "step": 206
1453
+ },
1454
+ {
1455
+ "epoch": 1.8008705114254626,
1456
+ "grad_norm": 1.328125,
1457
+ "learning_rate": 4.104347826086957e-05,
1458
+ "loss": 0.8158,
1459
+ "step": 207
1460
+ },
1461
+ {
1462
+ "epoch": 1.809575625680087,
1463
+ "grad_norm": 1.4765625,
1464
+ "learning_rate": 4.1e-05,
1465
+ "loss": 0.7081,
1466
+ "step": 208
1467
+ },
1468
+ {
1469
+ "epoch": 1.8182807399347116,
1470
+ "grad_norm": 1.046875,
1471
+ "learning_rate": 4.095652173913044e-05,
1472
+ "loss": 0.648,
1473
+ "step": 209
1474
+ },
1475
+ {
1476
+ "epoch": 1.8269858541893362,
1477
+ "grad_norm": 1.2734375,
1478
+ "learning_rate": 4.091304347826087e-05,
1479
+ "loss": 0.7721,
1480
+ "step": 210
1481
+ },
1482
+ {
1483
+ "epoch": 1.8356909684439608,
1484
+ "grad_norm": 1.03125,
1485
+ "learning_rate": 4.086956521739131e-05,
1486
+ "loss": 0.5612,
1487
+ "step": 211
1488
+ },
1489
+ {
1490
+ "epoch": 1.8443960826985855,
1491
+ "grad_norm": 1.1015625,
1492
+ "learning_rate": 4.082608695652174e-05,
1493
+ "loss": 0.635,
1494
+ "step": 212
1495
+ },
1496
+ {
1497
+ "epoch": 1.8531011969532099,
1498
+ "grad_norm": 1.1015625,
1499
+ "learning_rate": 4.0782608695652174e-05,
1500
+ "loss": 0.7699,
1501
+ "step": 213
1502
+ },
1503
+ {
1504
+ "epoch": 1.8618063112078347,
1505
+ "grad_norm": 1.234375,
1506
+ "learning_rate": 4.0739130434782615e-05,
1507
+ "loss": 0.7569,
1508
+ "step": 214
1509
+ },
1510
+ {
1511
+ "epoch": 1.870511425462459,
1512
+ "grad_norm": 1.203125,
1513
+ "learning_rate": 4.069565217391305e-05,
1514
+ "loss": 0.7783,
1515
+ "step": 215
1516
+ },
1517
+ {
1518
+ "epoch": 1.879216539717084,
1519
+ "grad_norm": 1.203125,
1520
+ "learning_rate": 4.065217391304348e-05,
1521
+ "loss": 0.7102,
1522
+ "step": 216
1523
+ },
1524
+ {
1525
+ "epoch": 1.8879216539717083,
1526
+ "grad_norm": 2.53125,
1527
+ "learning_rate": 4.0608695652173915e-05,
1528
+ "loss": 1.0202,
1529
+ "step": 217
1530
+ },
1531
+ {
1532
+ "epoch": 1.896626768226333,
1533
+ "grad_norm": 1.2109375,
1534
+ "learning_rate": 4.056521739130435e-05,
1535
+ "loss": 0.682,
1536
+ "step": 218
1537
+ },
1538
+ {
1539
+ "epoch": 1.9053318824809575,
1540
+ "grad_norm": 1.03125,
1541
+ "learning_rate": 4.052173913043478e-05,
1542
+ "loss": 0.648,
1543
+ "step": 219
1544
+ },
1545
+ {
1546
+ "epoch": 1.9140369967355821,
1547
+ "grad_norm": 1.234375,
1548
+ "learning_rate": 4.047826086956522e-05,
1549
+ "loss": 0.8244,
1550
+ "step": 220
1551
+ },
1552
+ {
1553
+ "epoch": 1.9227421109902068,
1554
+ "grad_norm": 1.078125,
1555
+ "learning_rate": 4.0434782608695655e-05,
1556
+ "loss": 0.6392,
1557
+ "step": 221
1558
+ },
1559
+ {
1560
+ "epoch": 1.9314472252448314,
1561
+ "grad_norm": 1.09375,
1562
+ "learning_rate": 4.039130434782609e-05,
1563
+ "loss": 0.7142,
1564
+ "step": 222
1565
+ },
1566
+ {
1567
+ "epoch": 1.940152339499456,
1568
+ "grad_norm": 1.1328125,
1569
+ "learning_rate": 4.034782608695652e-05,
1570
+ "loss": 0.637,
1571
+ "step": 223
1572
+ },
1573
+ {
1574
+ "epoch": 1.9488574537540804,
1575
+ "grad_norm": 1.03125,
1576
+ "learning_rate": 4.0304347826086956e-05,
1577
+ "loss": 0.7092,
1578
+ "step": 224
1579
+ },
1580
+ {
1581
+ "epoch": 1.9575625680087052,
1582
+ "grad_norm": 1.0390625,
1583
+ "learning_rate": 4.026086956521739e-05,
1584
+ "loss": 0.6927,
1585
+ "step": 225
1586
+ },
1587
+ {
1588
+ "epoch": 1.9662676822633296,
1589
+ "grad_norm": 1.125,
1590
+ "learning_rate": 4.021739130434783e-05,
1591
+ "loss": 0.744,
1592
+ "step": 226
1593
+ },
1594
+ {
1595
+ "epoch": 1.9749727965179544,
1596
+ "grad_norm": 1.1484375,
1597
+ "learning_rate": 4.017391304347826e-05,
1598
+ "loss": 0.6604,
1599
+ "step": 227
1600
+ },
1601
+ {
1602
+ "epoch": 1.9836779107725788,
1603
+ "grad_norm": 1.0625,
1604
+ "learning_rate": 4.01304347826087e-05,
1605
+ "loss": 0.7406,
1606
+ "step": 228
1607
+ },
1608
+ {
1609
+ "epoch": 1.9923830250272034,
1610
+ "grad_norm": 1.078125,
1611
+ "learning_rate": 4.008695652173913e-05,
1612
+ "loss": 0.645,
1613
+ "step": 229
1614
+ },
1615
+ {
1616
+ "epoch": 2.0,
1617
+ "grad_norm": 1.4765625,
1618
+ "learning_rate": 4.004347826086956e-05,
1619
+ "loss": 0.7339,
1620
+ "step": 230
1621
+ },
1622
+ {
1623
+ "epoch": 2.0087051142546244,
1624
+ "grad_norm": 1.2578125,
1625
+ "learning_rate": 4e-05,
1626
+ "loss": 0.6646,
1627
+ "step": 231
1628
+ },
1629
+ {
1630
+ "epoch": 2.0174102285092492,
1631
+ "grad_norm": 1.21875,
1632
+ "learning_rate": 3.995652173913044e-05,
1633
+ "loss": 0.4969,
1634
+ "step": 232
1635
+ },
1636
+ {
1637
+ "epoch": 2.0261153427638736,
1638
+ "grad_norm": 1.1953125,
1639
+ "learning_rate": 3.991304347826087e-05,
1640
+ "loss": 0.5202,
1641
+ "step": 233
1642
+ },
1643
+ {
1644
+ "epoch": 2.0348204570184985,
1645
+ "grad_norm": 0.97265625,
1646
+ "learning_rate": 3.986956521739131e-05,
1647
+ "loss": 0.671,
1648
+ "step": 234
1649
+ },
1650
+ {
1651
+ "epoch": 2.043525571273123,
1652
+ "grad_norm": 1.1875,
1653
+ "learning_rate": 3.9826086956521744e-05,
1654
+ "loss": 0.4306,
1655
+ "step": 235
1656
+ },
1657
+ {
1658
+ "epoch": 2.0522306855277477,
1659
+ "grad_norm": 1.75,
1660
+ "learning_rate": 3.978260869565217e-05,
1661
+ "loss": 0.5232,
1662
+ "step": 236
1663
+ },
1664
+ {
1665
+ "epoch": 2.060935799782372,
1666
+ "grad_norm": 1.6953125,
1667
+ "learning_rate": 3.973913043478261e-05,
1668
+ "loss": 0.498,
1669
+ "step": 237
1670
+ },
1671
+ {
1672
+ "epoch": 2.069640914036997,
1673
+ "grad_norm": 1.6484375,
1674
+ "learning_rate": 3.9695652173913044e-05,
1675
+ "loss": 0.6858,
1676
+ "step": 238
1677
+ },
1678
+ {
1679
+ "epoch": 2.0783460282916213,
1680
+ "grad_norm": 1.5703125,
1681
+ "learning_rate": 3.965217391304348e-05,
1682
+ "loss": 0.525,
1683
+ "step": 239
1684
+ },
1685
+ {
1686
+ "epoch": 2.087051142546246,
1687
+ "grad_norm": 1.203125,
1688
+ "learning_rate": 3.960869565217392e-05,
1689
+ "loss": 0.4226,
1690
+ "step": 240
1691
+ },
1692
+ {
1693
+ "epoch": 2.0957562568008705,
1694
+ "grad_norm": 1.1953125,
1695
+ "learning_rate": 3.956521739130435e-05,
1696
+ "loss": 0.5474,
1697
+ "step": 241
1698
+ },
1699
+ {
1700
+ "epoch": 2.104461371055495,
1701
+ "grad_norm": 1.109375,
1702
+ "learning_rate": 3.9521739130434785e-05,
1703
+ "loss": 0.4487,
1704
+ "step": 242
1705
+ },
1706
+ {
1707
+ "epoch": 2.1131664853101197,
1708
+ "grad_norm": 1.125,
1709
+ "learning_rate": 3.947826086956522e-05,
1710
+ "loss": 0.404,
1711
+ "step": 243
1712
+ },
1713
+ {
1714
+ "epoch": 2.121871599564744,
1715
+ "grad_norm": 1.1796875,
1716
+ "learning_rate": 3.943478260869565e-05,
1717
+ "loss": 0.4901,
1718
+ "step": 244
1719
+ },
1720
+ {
1721
+ "epoch": 2.130576713819369,
1722
+ "grad_norm": 1.1875,
1723
+ "learning_rate": 3.939130434782609e-05,
1724
+ "loss": 0.417,
1725
+ "step": 245
1726
+ },
1727
+ {
1728
+ "epoch": 2.1392818280739934,
1729
+ "grad_norm": 1.1171875,
1730
+ "learning_rate": 3.9347826086956525e-05,
1731
+ "loss": 0.7131,
1732
+ "step": 246
1733
+ },
1734
+ {
1735
+ "epoch": 2.147986942328618,
1736
+ "grad_norm": 1.296875,
1737
+ "learning_rate": 3.930434782608696e-05,
1738
+ "loss": 0.5996,
1739
+ "step": 247
1740
+ },
1741
+ {
1742
+ "epoch": 2.1566920565832426,
1743
+ "grad_norm": 1.171875,
1744
+ "learning_rate": 3.926086956521739e-05,
1745
+ "loss": 0.6252,
1746
+ "step": 248
1747
+ },
1748
+ {
1749
+ "epoch": 2.1653971708378674,
1750
+ "grad_norm": 1.28125,
1751
+ "learning_rate": 3.9217391304347825e-05,
1752
+ "loss": 0.4977,
1753
+ "step": 249
1754
+ },
1755
+ {
1756
+ "epoch": 2.174102285092492,
1757
+ "grad_norm": 1.1640625,
1758
+ "learning_rate": 3.917391304347826e-05,
1759
+ "loss": 0.7226,
1760
+ "step": 250
1761
+ },
1762
+ {
1763
+ "epoch": 2.1828073993471167,
1764
+ "grad_norm": 1.40625,
1765
+ "learning_rate": 3.91304347826087e-05,
1766
+ "loss": 0.7221,
1767
+ "step": 251
1768
+ },
1769
+ {
1770
+ "epoch": 2.191512513601741,
1771
+ "grad_norm": 1.171875,
1772
+ "learning_rate": 3.908695652173913e-05,
1773
+ "loss": 0.6118,
1774
+ "step": 252
1775
+ },
1776
+ {
1777
+ "epoch": 2.2002176278563654,
1778
+ "grad_norm": 1.109375,
1779
+ "learning_rate": 3.9043478260869566e-05,
1780
+ "loss": 0.8739,
1781
+ "step": 253
1782
+ },
1783
+ {
1784
+ "epoch": 2.2089227421109903,
1785
+ "grad_norm": 1.2734375,
1786
+ "learning_rate": 3.9000000000000006e-05,
1787
+ "loss": 0.5692,
1788
+ "step": 254
1789
+ },
1790
+ {
1791
+ "epoch": 2.2176278563656147,
1792
+ "grad_norm": 1.2890625,
1793
+ "learning_rate": 3.895652173913043e-05,
1794
+ "loss": 0.527,
1795
+ "step": 255
1796
+ },
1797
+ {
1798
+ "epoch": 2.2263329706202395,
1799
+ "grad_norm": 1.1015625,
1800
+ "learning_rate": 3.8913043478260866e-05,
1801
+ "loss": 0.4093,
1802
+ "step": 256
1803
+ },
1804
+ {
1805
+ "epoch": 2.235038084874864,
1806
+ "grad_norm": 1.3828125,
1807
+ "learning_rate": 3.8869565217391306e-05,
1808
+ "loss": 0.57,
1809
+ "step": 257
1810
+ },
1811
+ {
1812
+ "epoch": 2.2437431991294887,
1813
+ "grad_norm": 1.15625,
1814
+ "learning_rate": 3.882608695652174e-05,
1815
+ "loss": 0.4616,
1816
+ "step": 258
1817
+ },
1818
+ {
1819
+ "epoch": 2.252448313384113,
1820
+ "grad_norm": 1.2265625,
1821
+ "learning_rate": 3.878260869565218e-05,
1822
+ "loss": 0.5554,
1823
+ "step": 259
1824
+ },
1825
+ {
1826
+ "epoch": 2.261153427638738,
1827
+ "grad_norm": 1.0078125,
1828
+ "learning_rate": 3.8739130434782613e-05,
1829
+ "loss": 0.4197,
1830
+ "step": 260
1831
+ },
1832
+ {
1833
+ "epoch": 2.2698585418933623,
1834
+ "grad_norm": 1.140625,
1835
+ "learning_rate": 3.869565217391305e-05,
1836
+ "loss": 0.4527,
1837
+ "step": 261
1838
+ },
1839
+ {
1840
+ "epoch": 2.2785636561479867,
1841
+ "grad_norm": 1.28125,
1842
+ "learning_rate": 3.865217391304348e-05,
1843
+ "loss": 0.6683,
1844
+ "step": 262
1845
+ },
1846
+ {
1847
+ "epoch": 2.2872687704026116,
1848
+ "grad_norm": 1.203125,
1849
+ "learning_rate": 3.8608695652173914e-05,
1850
+ "loss": 0.4808,
1851
+ "step": 263
1852
+ },
1853
+ {
1854
+ "epoch": 2.295973884657236,
1855
+ "grad_norm": 1.109375,
1856
+ "learning_rate": 3.856521739130435e-05,
1857
+ "loss": 0.4871,
1858
+ "step": 264
1859
+ },
1860
+ {
1861
+ "epoch": 2.304678998911861,
1862
+ "grad_norm": 1.2421875,
1863
+ "learning_rate": 3.852173913043479e-05,
1864
+ "loss": 0.6857,
1865
+ "step": 265
1866
+ },
1867
+ {
1868
+ "epoch": 2.313384113166485,
1869
+ "grad_norm": 1.1328125,
1870
+ "learning_rate": 3.847826086956522e-05,
1871
+ "loss": 0.6002,
1872
+ "step": 266
1873
+ },
1874
+ {
1875
+ "epoch": 2.32208922742111,
1876
+ "grad_norm": 1.1796875,
1877
+ "learning_rate": 3.8434782608695654e-05,
1878
+ "loss": 0.5997,
1879
+ "step": 267
1880
+ },
1881
+ {
1882
+ "epoch": 2.3307943416757344,
1883
+ "grad_norm": 1.1640625,
1884
+ "learning_rate": 3.839130434782609e-05,
1885
+ "loss": 0.4839,
1886
+ "step": 268
1887
+ },
1888
+ {
1889
+ "epoch": 2.3394994559303592,
1890
+ "grad_norm": 1.1640625,
1891
+ "learning_rate": 3.834782608695652e-05,
1892
+ "loss": 0.5754,
1893
+ "step": 269
1894
+ },
1895
+ {
1896
+ "epoch": 2.3482045701849836,
1897
+ "grad_norm": 1.21875,
1898
+ "learning_rate": 3.8304347826086955e-05,
1899
+ "loss": 0.5072,
1900
+ "step": 270
1901
+ },
1902
+ {
1903
+ "epoch": 2.356909684439608,
1904
+ "grad_norm": 1.453125,
1905
+ "learning_rate": 3.8260869565217395e-05,
1906
+ "loss": 0.6542,
1907
+ "step": 271
1908
+ },
1909
+ {
1910
+ "epoch": 2.365614798694233,
1911
+ "grad_norm": 1.2265625,
1912
+ "learning_rate": 3.821739130434783e-05,
1913
+ "loss": 0.6175,
1914
+ "step": 272
1915
+ },
1916
+ {
1917
+ "epoch": 2.3743199129488577,
1918
+ "grad_norm": 1.0859375,
1919
+ "learning_rate": 3.817391304347827e-05,
1920
+ "loss": 0.5063,
1921
+ "step": 273
1922
+ },
1923
+ {
1924
+ "epoch": 2.383025027203482,
1925
+ "grad_norm": 1.1484375,
1926
+ "learning_rate": 3.8130434782608695e-05,
1927
+ "loss": 0.5095,
1928
+ "step": 274
1929
+ },
1930
+ {
1931
+ "epoch": 2.3917301414581065,
1932
+ "grad_norm": 1.5078125,
1933
+ "learning_rate": 3.808695652173913e-05,
1934
+ "loss": 0.5236,
1935
+ "step": 275
1936
+ },
1937
+ {
1938
+ "epoch": 2.4004352557127313,
1939
+ "grad_norm": 1.25,
1940
+ "learning_rate": 3.804347826086957e-05,
1941
+ "loss": 0.6417,
1942
+ "step": 276
1943
+ },
1944
+ {
1945
+ "epoch": 2.4091403699673557,
1946
+ "grad_norm": 1.2421875,
1947
+ "learning_rate": 3.8e-05,
1948
+ "loss": 0.5618,
1949
+ "step": 277
1950
+ },
1951
+ {
1952
+ "epoch": 2.4178454842219805,
1953
+ "grad_norm": 1.1796875,
1954
+ "learning_rate": 3.7956521739130436e-05,
1955
+ "loss": 0.409,
1956
+ "step": 278
1957
+ },
1958
+ {
1959
+ "epoch": 2.426550598476605,
1960
+ "grad_norm": 1.2421875,
1961
+ "learning_rate": 3.7913043478260876e-05,
1962
+ "loss": 0.4815,
1963
+ "step": 279
1964
+ },
1965
+ {
1966
+ "epoch": 2.4352557127312298,
1967
+ "grad_norm": 1.1796875,
1968
+ "learning_rate": 3.786956521739131e-05,
1969
+ "loss": 0.5533,
1970
+ "step": 280
1971
+ },
1972
+ {
1973
+ "epoch": 2.443960826985854,
1974
+ "grad_norm": 1.1875,
1975
+ "learning_rate": 3.7826086956521736e-05,
1976
+ "loss": 0.4977,
1977
+ "step": 281
1978
+ },
1979
+ {
1980
+ "epoch": 2.452665941240479,
1981
+ "grad_norm": 1.265625,
1982
+ "learning_rate": 3.7782608695652176e-05,
1983
+ "loss": 0.5402,
1984
+ "step": 282
1985
+ },
1986
+ {
1987
+ "epoch": 2.4613710554951034,
1988
+ "grad_norm": 1.234375,
1989
+ "learning_rate": 3.773913043478261e-05,
1990
+ "loss": 0.4572,
1991
+ "step": 283
1992
+ },
1993
+ {
1994
+ "epoch": 2.4700761697497278,
1995
+ "grad_norm": 1.1953125,
1996
+ "learning_rate": 3.769565217391304e-05,
1997
+ "loss": 0.5009,
1998
+ "step": 284
1999
+ },
2000
+ {
2001
+ "epoch": 2.4787812840043526,
2002
+ "grad_norm": 1.140625,
2003
+ "learning_rate": 3.765217391304348e-05,
2004
+ "loss": 0.508,
2005
+ "step": 285
2006
+ },
2007
+ {
2008
+ "epoch": 2.487486398258977,
2009
+ "grad_norm": 1.3359375,
2010
+ "learning_rate": 3.7608695652173917e-05,
2011
+ "loss": 0.5922,
2012
+ "step": 286
2013
+ },
2014
+ {
2015
+ "epoch": 2.496191512513602,
2016
+ "grad_norm": 1.15625,
2017
+ "learning_rate": 3.756521739130435e-05,
2018
+ "loss": 0.5631,
2019
+ "step": 287
2020
+ },
2021
+ {
2022
+ "epoch": 2.504896626768226,
2023
+ "grad_norm": 1.1953125,
2024
+ "learning_rate": 3.752173913043478e-05,
2025
+ "loss": 0.5335,
2026
+ "step": 288
2027
+ },
2028
+ {
2029
+ "epoch": 2.513601741022851,
2030
+ "grad_norm": 1.28125,
2031
+ "learning_rate": 3.747826086956522e-05,
2032
+ "loss": 0.6231,
2033
+ "step": 289
2034
+ },
2035
+ {
2036
+ "epoch": 2.5223068552774754,
2037
+ "grad_norm": 1.4765625,
2038
+ "learning_rate": 3.743478260869566e-05,
2039
+ "loss": 0.5049,
2040
+ "step": 290
2041
+ },
2042
+ {
2043
+ "epoch": 2.5310119695321003,
2044
+ "grad_norm": 1.25,
2045
+ "learning_rate": 3.739130434782609e-05,
2046
+ "loss": 0.6165,
2047
+ "step": 291
2048
+ },
2049
+ {
2050
+ "epoch": 2.5397170837867247,
2051
+ "grad_norm": 1.2265625,
2052
+ "learning_rate": 3.7347826086956524e-05,
2053
+ "loss": 0.5344,
2054
+ "step": 292
2055
+ },
2056
+ {
2057
+ "epoch": 2.548422198041349,
2058
+ "grad_norm": 1.078125,
2059
+ "learning_rate": 3.730434782608696e-05,
2060
+ "loss": 0.5464,
2061
+ "step": 293
2062
+ },
2063
+ {
2064
+ "epoch": 2.557127312295974,
2065
+ "grad_norm": 1.1640625,
2066
+ "learning_rate": 3.726086956521739e-05,
2067
+ "loss": 0.571,
2068
+ "step": 294
2069
+ },
2070
+ {
2071
+ "epoch": 2.5658324265505987,
2072
+ "grad_norm": 1.1328125,
2073
+ "learning_rate": 3.7217391304347824e-05,
2074
+ "loss": 0.4843,
2075
+ "step": 295
2076
+ },
2077
+ {
2078
+ "epoch": 2.574537540805223,
2079
+ "grad_norm": 1.4453125,
2080
+ "learning_rate": 3.7173913043478264e-05,
2081
+ "loss": 0.4947,
2082
+ "step": 296
2083
+ },
2084
+ {
2085
+ "epoch": 2.5832426550598475,
2086
+ "grad_norm": 1.171875,
2087
+ "learning_rate": 3.71304347826087e-05,
2088
+ "loss": 0.5202,
2089
+ "step": 297
2090
+ },
2091
+ {
2092
+ "epoch": 2.5919477693144723,
2093
+ "grad_norm": 1.1796875,
2094
+ "learning_rate": 3.708695652173913e-05,
2095
+ "loss": 0.4633,
2096
+ "step": 298
2097
+ },
2098
+ {
2099
+ "epoch": 2.6006528835690967,
2100
+ "grad_norm": 1.3125,
2101
+ "learning_rate": 3.704347826086957e-05,
2102
+ "loss": 0.4723,
2103
+ "step": 299
2104
+ },
2105
+ {
2106
+ "epoch": 2.6093579978237216,
2107
+ "grad_norm": 1.265625,
2108
+ "learning_rate": 3.7e-05,
2109
+ "loss": 0.4811,
2110
+ "step": 300
2111
+ },
2112
+ {
2113
+ "epoch": 2.618063112078346,
2114
+ "grad_norm": 1.34375,
2115
+ "learning_rate": 3.695652173913043e-05,
2116
+ "loss": 0.7662,
2117
+ "step": 301
2118
+ },
2119
+ {
2120
+ "epoch": 2.6267682263329704,
2121
+ "grad_norm": 1.3125,
2122
+ "learning_rate": 3.691304347826087e-05,
2123
+ "loss": 0.67,
2124
+ "step": 302
2125
+ },
2126
+ {
2127
+ "epoch": 2.635473340587595,
2128
+ "grad_norm": 1.3671875,
2129
+ "learning_rate": 3.6869565217391305e-05,
2130
+ "loss": 0.6694,
2131
+ "step": 303
2132
+ },
2133
+ {
2134
+ "epoch": 2.64417845484222,
2135
+ "grad_norm": 0.9921875,
2136
+ "learning_rate": 3.682608695652174e-05,
2137
+ "loss": 0.5212,
2138
+ "step": 304
2139
+ },
2140
+ {
2141
+ "epoch": 2.6528835690968444,
2142
+ "grad_norm": 1.09375,
2143
+ "learning_rate": 3.678260869565218e-05,
2144
+ "loss": 0.4885,
2145
+ "step": 305
2146
+ },
2147
+ {
2148
+ "epoch": 2.661588683351469,
2149
+ "grad_norm": 1.15625,
2150
+ "learning_rate": 3.673913043478261e-05,
2151
+ "loss": 0.6483,
2152
+ "step": 306
2153
+ },
2154
+ {
2155
+ "epoch": 2.6702937976060936,
2156
+ "grad_norm": 1.25,
2157
+ "learning_rate": 3.6695652173913046e-05,
2158
+ "loss": 0.557,
2159
+ "step": 307
2160
+ },
2161
+ {
2162
+ "epoch": 2.678998911860718,
2163
+ "grad_norm": 1.109375,
2164
+ "learning_rate": 3.665217391304348e-05,
2165
+ "loss": 0.5384,
2166
+ "step": 308
2167
+ },
2168
+ {
2169
+ "epoch": 2.687704026115343,
2170
+ "grad_norm": 1.1484375,
2171
+ "learning_rate": 3.660869565217391e-05,
2172
+ "loss": 0.3727,
2173
+ "step": 309
2174
+ },
2175
+ {
2176
+ "epoch": 2.6964091403699673,
2177
+ "grad_norm": 1.3515625,
2178
+ "learning_rate": 3.656521739130435e-05,
2179
+ "loss": 0.5383,
2180
+ "step": 310
2181
+ },
2182
+ {
2183
+ "epoch": 2.705114254624592,
2184
+ "grad_norm": 1.328125,
2185
+ "learning_rate": 3.6521739130434786e-05,
2186
+ "loss": 0.6103,
2187
+ "step": 311
2188
+ },
2189
+ {
2190
+ "epoch": 2.7138193688792165,
2191
+ "grad_norm": 1.2421875,
2192
+ "learning_rate": 3.647826086956522e-05,
2193
+ "loss": 0.5469,
2194
+ "step": 312
2195
+ },
2196
+ {
2197
+ "epoch": 2.7225244831338413,
2198
+ "grad_norm": 1.1953125,
2199
+ "learning_rate": 3.643478260869565e-05,
2200
+ "loss": 0.5393,
2201
+ "step": 313
2202
+ },
2203
+ {
2204
+ "epoch": 2.7312295973884657,
2205
+ "grad_norm": 1.1953125,
2206
+ "learning_rate": 3.6391304347826086e-05,
2207
+ "loss": 0.6248,
2208
+ "step": 314
2209
+ },
2210
+ {
2211
+ "epoch": 2.73993471164309,
2212
+ "grad_norm": 1.2421875,
2213
+ "learning_rate": 3.634782608695652e-05,
2214
+ "loss": 0.5939,
2215
+ "step": 315
2216
+ },
2217
+ {
2218
+ "epoch": 2.748639825897715,
2219
+ "grad_norm": 1.1796875,
2220
+ "learning_rate": 3.630434782608696e-05,
2221
+ "loss": 0.5281,
2222
+ "step": 316
2223
+ },
2224
+ {
2225
+ "epoch": 2.7573449401523398,
2226
+ "grad_norm": 1.09375,
2227
+ "learning_rate": 3.6260869565217394e-05,
2228
+ "loss": 0.6094,
2229
+ "step": 317
2230
+ },
2231
+ {
2232
+ "epoch": 2.766050054406964,
2233
+ "grad_norm": 1.625,
2234
+ "learning_rate": 3.621739130434783e-05,
2235
+ "loss": 0.5353,
2236
+ "step": 318
2237
+ },
2238
+ {
2239
+ "epoch": 2.7747551686615886,
2240
+ "grad_norm": 1.296875,
2241
+ "learning_rate": 3.617391304347826e-05,
2242
+ "loss": 0.8688,
2243
+ "step": 319
2244
+ },
2245
+ {
2246
+ "epoch": 2.7834602829162134,
2247
+ "grad_norm": 1.1796875,
2248
+ "learning_rate": 3.6130434782608694e-05,
2249
+ "loss": 0.4359,
2250
+ "step": 320
2251
+ },
2252
+ {
2253
+ "epoch": 2.792165397170838,
2254
+ "grad_norm": 1.34375,
2255
+ "learning_rate": 3.6086956521739134e-05,
2256
+ "loss": 0.5274,
2257
+ "step": 321
2258
+ },
2259
+ {
2260
+ "epoch": 2.8008705114254626,
2261
+ "grad_norm": 1.0703125,
2262
+ "learning_rate": 3.604347826086957e-05,
2263
+ "loss": 0.5489,
2264
+ "step": 322
2265
+ },
2266
+ {
2267
+ "epoch": 2.809575625680087,
2268
+ "grad_norm": 1.4375,
2269
+ "learning_rate": 3.6e-05,
2270
+ "loss": 0.4897,
2271
+ "step": 323
2272
+ },
2273
+ {
2274
+ "epoch": 2.8182807399347114,
2275
+ "grad_norm": 1.2109375,
2276
+ "learning_rate": 3.595652173913044e-05,
2277
+ "loss": 0.4663,
2278
+ "step": 324
2279
+ },
2280
+ {
2281
+ "epoch": 2.8269858541893362,
2282
+ "grad_norm": 1.125,
2283
+ "learning_rate": 3.5913043478260875e-05,
2284
+ "loss": 0.6313,
2285
+ "step": 325
2286
+ },
2287
+ {
2288
+ "epoch": 2.835690968443961,
2289
+ "grad_norm": 1.1875,
2290
+ "learning_rate": 3.58695652173913e-05,
2291
+ "loss": 0.8168,
2292
+ "step": 326
2293
+ },
2294
+ {
2295
+ "epoch": 2.8443960826985855,
2296
+ "grad_norm": 1.203125,
2297
+ "learning_rate": 3.582608695652174e-05,
2298
+ "loss": 0.8116,
2299
+ "step": 327
2300
+ },
2301
+ {
2302
+ "epoch": 2.85310119695321,
2303
+ "grad_norm": 1.1015625,
2304
+ "learning_rate": 3.5782608695652175e-05,
2305
+ "loss": 0.6349,
2306
+ "step": 328
2307
+ },
2308
+ {
2309
+ "epoch": 2.8618063112078347,
2310
+ "grad_norm": 1.0859375,
2311
+ "learning_rate": 3.573913043478261e-05,
2312
+ "loss": 0.5963,
2313
+ "step": 329
2314
+ },
2315
+ {
2316
+ "epoch": 2.870511425462459,
2317
+ "grad_norm": 1.09375,
2318
+ "learning_rate": 3.569565217391305e-05,
2319
+ "loss": 0.6366,
2320
+ "step": 330
2321
+ },
2322
+ {
2323
+ "epoch": 2.879216539717084,
2324
+ "grad_norm": 1.09375,
2325
+ "learning_rate": 3.565217391304348e-05,
2326
+ "loss": 0.5822,
2327
+ "step": 331
2328
+ },
2329
+ {
2330
+ "epoch": 2.8879216539717083,
2331
+ "grad_norm": 1.1796875,
2332
+ "learning_rate": 3.5608695652173915e-05,
2333
+ "loss": 0.4839,
2334
+ "step": 332
2335
+ },
2336
+ {
2337
+ "epoch": 2.8966267682263327,
2338
+ "grad_norm": 1.0859375,
2339
+ "learning_rate": 3.556521739130435e-05,
2340
+ "loss": 0.6158,
2341
+ "step": 333
2342
+ },
2343
+ {
2344
+ "epoch": 2.9053318824809575,
2345
+ "grad_norm": 1.2109375,
2346
+ "learning_rate": 3.552173913043478e-05,
2347
+ "loss": 0.4699,
2348
+ "step": 334
2349
+ },
2350
+ {
2351
+ "epoch": 2.9140369967355824,
2352
+ "grad_norm": 1.4765625,
2353
+ "learning_rate": 3.5478260869565216e-05,
2354
+ "loss": 0.6381,
2355
+ "step": 335
2356
+ },
2357
+ {
2358
+ "epoch": 2.9227421109902068,
2359
+ "grad_norm": 1.2421875,
2360
+ "learning_rate": 3.5434782608695656e-05,
2361
+ "loss": 0.5853,
2362
+ "step": 336
2363
+ },
2364
+ {
2365
+ "epoch": 2.931447225244831,
2366
+ "grad_norm": 1.234375,
2367
+ "learning_rate": 3.539130434782609e-05,
2368
+ "loss": 0.6238,
2369
+ "step": 337
2370
+ },
2371
+ {
2372
+ "epoch": 2.940152339499456,
2373
+ "grad_norm": 1.3046875,
2374
+ "learning_rate": 3.534782608695652e-05,
2375
+ "loss": 0.6795,
2376
+ "step": 338
2377
+ },
2378
+ {
2379
+ "epoch": 2.9488574537540804,
2380
+ "grad_norm": 1.0859375,
2381
+ "learning_rate": 3.5304347826086956e-05,
2382
+ "loss": 0.546,
2383
+ "step": 339
2384
+ },
2385
+ {
2386
+ "epoch": 2.957562568008705,
2387
+ "grad_norm": 1.1171875,
2388
+ "learning_rate": 3.526086956521739e-05,
2389
+ "loss": 0.4245,
2390
+ "step": 340
2391
+ },
2392
+ {
2393
+ "epoch": 2.9662676822633296,
2394
+ "grad_norm": 1.265625,
2395
+ "learning_rate": 3.521739130434783e-05,
2396
+ "loss": 0.5433,
2397
+ "step": 341
2398
+ },
2399
+ {
2400
+ "epoch": 2.9749727965179544,
2401
+ "grad_norm": 1.2578125,
2402
+ "learning_rate": 3.517391304347826e-05,
2403
+ "loss": 0.4675,
2404
+ "step": 342
2405
+ },
2406
+ {
2407
+ "epoch": 2.983677910772579,
2408
+ "grad_norm": 1.28125,
2409
+ "learning_rate": 3.51304347826087e-05,
2410
+ "loss": 0.6289,
2411
+ "step": 343
2412
+ },
2413
+ {
2414
+ "epoch": 2.9923830250272037,
2415
+ "grad_norm": 1.21875,
2416
+ "learning_rate": 3.508695652173914e-05,
2417
+ "loss": 0.6289,
2418
+ "step": 344
2419
+ },
2420
+ {
2421
+ "epoch": 3.0,
2422
+ "grad_norm": 1.28125,
2423
+ "learning_rate": 3.5043478260869564e-05,
2424
+ "loss": 0.4109,
2425
+ "step": 345
2426
+ },
2427
+ {
2428
+ "epoch": 3.0087051142546244,
2429
+ "grad_norm": 1.421875,
2430
+ "learning_rate": 3.5e-05,
2431
+ "loss": 0.4183,
2432
+ "step": 346
2433
+ },
2434
+ {
2435
+ "epoch": 3.0174102285092492,
2436
+ "grad_norm": 1.1953125,
2437
+ "learning_rate": 3.495652173913044e-05,
2438
+ "loss": 0.3592,
2439
+ "step": 347
2440
+ },
2441
+ {
2442
+ "epoch": 3.0261153427638736,
2443
+ "grad_norm": 1.5234375,
2444
+ "learning_rate": 3.491304347826087e-05,
2445
+ "loss": 0.5384,
2446
+ "step": 348
2447
+ },
2448
+ {
2449
+ "epoch": 3.0348204570184985,
2450
+ "grad_norm": 1.125,
2451
+ "learning_rate": 3.4869565217391304e-05,
2452
+ "loss": 0.5139,
2453
+ "step": 349
2454
+ },
2455
+ {
2456
+ "epoch": 3.043525571273123,
2457
+ "grad_norm": 1.7421875,
2458
+ "learning_rate": 3.4826086956521744e-05,
2459
+ "loss": 0.9164,
2460
+ "step": 350
2461
+ },
2462
+ {
2463
+ "epoch": 3.0522306855277477,
2464
+ "grad_norm": 1.8125,
2465
+ "learning_rate": 3.478260869565218e-05,
2466
+ "loss": 0.3615,
2467
+ "step": 351
2468
+ },
2469
+ {
2470
+ "epoch": 3.060935799782372,
2471
+ "grad_norm": 2.046875,
2472
+ "learning_rate": 3.473913043478261e-05,
2473
+ "loss": 0.3857,
2474
+ "step": 352
2475
+ },
2476
+ {
2477
+ "epoch": 3.069640914036997,
2478
+ "grad_norm": 1.84375,
2479
+ "learning_rate": 3.4695652173913045e-05,
2480
+ "loss": 0.4078,
2481
+ "step": 353
2482
+ },
2483
+ {
2484
+ "epoch": 3.0783460282916213,
2485
+ "grad_norm": 1.796875,
2486
+ "learning_rate": 3.465217391304348e-05,
2487
+ "loss": 0.4181,
2488
+ "step": 354
2489
+ },
2490
+ {
2491
+ "epoch": 3.087051142546246,
2492
+ "grad_norm": 1.6015625,
2493
+ "learning_rate": 3.460869565217392e-05,
2494
+ "loss": 0.3059,
2495
+ "step": 355
2496
+ },
2497
+ {
2498
+ "epoch": 3.0957562568008705,
2499
+ "grad_norm": 1.390625,
2500
+ "learning_rate": 3.456521739130435e-05,
2501
+ "loss": 0.4606,
2502
+ "step": 356
2503
+ },
2504
+ {
2505
+ "epoch": 3.104461371055495,
2506
+ "grad_norm": 1.28125,
2507
+ "learning_rate": 3.4521739130434785e-05,
2508
+ "loss": 0.4711,
2509
+ "step": 357
2510
+ },
2511
+ {
2512
+ "epoch": 3.1131664853101197,
2513
+ "grad_norm": 1.28125,
2514
+ "learning_rate": 3.447826086956522e-05,
2515
+ "loss": 0.4096,
2516
+ "step": 358
2517
+ },
2518
+ {
2519
+ "epoch": 3.121871599564744,
2520
+ "grad_norm": 1.25,
2521
+ "learning_rate": 3.443478260869565e-05,
2522
+ "loss": 0.6088,
2523
+ "step": 359
2524
+ },
2525
+ {
2526
+ "epoch": 3.130576713819369,
2527
+ "grad_norm": 1.171875,
2528
+ "learning_rate": 3.4391304347826085e-05,
2529
+ "loss": 0.3165,
2530
+ "step": 360
2531
+ },
2532
+ {
2533
+ "epoch": 3.1392818280739934,
2534
+ "grad_norm": 1.2421875,
2535
+ "learning_rate": 3.4347826086956526e-05,
2536
+ "loss": 0.306,
2537
+ "step": 361
2538
+ },
2539
+ {
2540
+ "epoch": 3.147986942328618,
2541
+ "grad_norm": 1.2890625,
2542
+ "learning_rate": 3.430434782608696e-05,
2543
+ "loss": 0.3672,
2544
+ "step": 362
2545
+ },
2546
+ {
2547
+ "epoch": 3.1566920565832426,
2548
+ "grad_norm": 1.1875,
2549
+ "learning_rate": 3.426086956521739e-05,
2550
+ "loss": 0.4125,
2551
+ "step": 363
2552
+ },
2553
+ {
2554
+ "epoch": 3.1653971708378674,
2555
+ "grad_norm": 1.5,
2556
+ "learning_rate": 3.4217391304347826e-05,
2557
+ "loss": 0.4383,
2558
+ "step": 364
2559
+ },
2560
+ {
2561
+ "epoch": 3.174102285092492,
2562
+ "grad_norm": 1.40625,
2563
+ "learning_rate": 3.417391304347826e-05,
2564
+ "loss": 0.3679,
2565
+ "step": 365
2566
+ },
2567
+ {
2568
+ "epoch": 3.1828073993471167,
2569
+ "grad_norm": 1.453125,
2570
+ "learning_rate": 3.413043478260869e-05,
2571
+ "loss": 0.4468,
2572
+ "step": 366
2573
+ },
2574
+ {
2575
+ "epoch": 3.191512513601741,
2576
+ "grad_norm": 1.34375,
2577
+ "learning_rate": 3.408695652173913e-05,
2578
+ "loss": 0.3335,
2579
+ "step": 367
2580
+ },
2581
+ {
2582
+ "epoch": 3.2002176278563654,
2583
+ "grad_norm": 1.359375,
2584
+ "learning_rate": 3.4043478260869566e-05,
2585
+ "loss": 0.3439,
2586
+ "step": 368
2587
+ },
2588
+ {
2589
+ "epoch": 3.2089227421109903,
2590
+ "grad_norm": 1.4765625,
2591
+ "learning_rate": 3.4000000000000007e-05,
2592
+ "loss": 0.4858,
2593
+ "step": 369
2594
+ },
2595
+ {
2596
+ "epoch": 3.2176278563656147,
2597
+ "grad_norm": 1.453125,
2598
+ "learning_rate": 3.395652173913044e-05,
2599
+ "loss": 0.3231,
2600
+ "step": 370
2601
+ },
2602
+ {
2603
+ "epoch": 3.2263329706202395,
2604
+ "grad_norm": 1.6171875,
2605
+ "learning_rate": 3.3913043478260867e-05,
2606
+ "loss": 0.5224,
2607
+ "step": 371
2608
+ },
2609
+ {
2610
+ "epoch": 3.235038084874864,
2611
+ "grad_norm": 1.328125,
2612
+ "learning_rate": 3.386956521739131e-05,
2613
+ "loss": 0.3573,
2614
+ "step": 372
2615
+ },
2616
+ {
2617
+ "epoch": 3.2437431991294887,
2618
+ "grad_norm": 1.359375,
2619
+ "learning_rate": 3.382608695652174e-05,
2620
+ "loss": 0.3516,
2621
+ "step": 373
2622
+ },
2623
+ {
2624
+ "epoch": 3.252448313384113,
2625
+ "grad_norm": 1.3515625,
2626
+ "learning_rate": 3.3782608695652174e-05,
2627
+ "loss": 0.3656,
2628
+ "step": 374
2629
+ },
2630
+ {
2631
+ "epoch": 3.261153427638738,
2632
+ "grad_norm": 1.4453125,
2633
+ "learning_rate": 3.3739130434782614e-05,
2634
+ "loss": 0.5266,
2635
+ "step": 375
2636
+ },
2637
+ {
2638
+ "epoch": 3.2698585418933623,
2639
+ "grad_norm": 1.3515625,
2640
+ "learning_rate": 3.369565217391305e-05,
2641
+ "loss": 0.3544,
2642
+ "step": 376
2643
+ },
2644
+ {
2645
+ "epoch": 3.2785636561479867,
2646
+ "grad_norm": 1.2890625,
2647
+ "learning_rate": 3.365217391304348e-05,
2648
+ "loss": 0.3623,
2649
+ "step": 377
2650
+ },
2651
+ {
2652
+ "epoch": 3.2872687704026116,
2653
+ "grad_norm": 1.3203125,
2654
+ "learning_rate": 3.3608695652173914e-05,
2655
+ "loss": 0.3814,
2656
+ "step": 378
2657
+ },
2658
+ {
2659
+ "epoch": 3.295973884657236,
2660
+ "grad_norm": 1.390625,
2661
+ "learning_rate": 3.356521739130435e-05,
2662
+ "loss": 0.42,
2663
+ "step": 379
2664
+ },
2665
+ {
2666
+ "epoch": 3.304678998911861,
2667
+ "grad_norm": 1.2890625,
2668
+ "learning_rate": 3.352173913043478e-05,
2669
+ "loss": 0.4972,
2670
+ "step": 380
2671
+ },
2672
+ {
2673
+ "epoch": 3.313384113166485,
2674
+ "grad_norm": 1.6015625,
2675
+ "learning_rate": 3.347826086956522e-05,
2676
+ "loss": 0.5218,
2677
+ "step": 381
2678
+ },
2679
+ {
2680
+ "epoch": 3.32208922742111,
2681
+ "grad_norm": 1.1953125,
2682
+ "learning_rate": 3.3434782608695655e-05,
2683
+ "loss": 0.3364,
2684
+ "step": 382
2685
+ },
2686
+ {
2687
+ "epoch": 3.3307943416757344,
2688
+ "grad_norm": 1.2890625,
2689
+ "learning_rate": 3.339130434782609e-05,
2690
+ "loss": 0.363,
2691
+ "step": 383
2692
+ },
2693
+ {
2694
+ "epoch": 3.3394994559303592,
2695
+ "grad_norm": 1.40625,
2696
+ "learning_rate": 3.334782608695652e-05,
2697
+ "loss": 0.3453,
2698
+ "step": 384
2699
+ },
2700
+ {
2701
+ "epoch": 3.3482045701849836,
2702
+ "grad_norm": 1.4765625,
2703
+ "learning_rate": 3.3304347826086955e-05,
2704
+ "loss": 0.3866,
2705
+ "step": 385
2706
+ },
2707
+ {
2708
+ "epoch": 3.356909684439608,
2709
+ "grad_norm": 1.34375,
2710
+ "learning_rate": 3.3260869565217395e-05,
2711
+ "loss": 0.3148,
2712
+ "step": 386
2713
+ },
2714
+ {
2715
+ "epoch": 3.365614798694233,
2716
+ "grad_norm": 1.2578125,
2717
+ "learning_rate": 3.321739130434783e-05,
2718
+ "loss": 0.6014,
2719
+ "step": 387
2720
+ },
2721
+ {
2722
+ "epoch": 3.3743199129488577,
2723
+ "grad_norm": 1.4609375,
2724
+ "learning_rate": 3.317391304347826e-05,
2725
+ "loss": 0.3306,
2726
+ "step": 388
2727
+ },
2728
+ {
2729
+ "epoch": 3.383025027203482,
2730
+ "grad_norm": 1.3125,
2731
+ "learning_rate": 3.31304347826087e-05,
2732
+ "loss": 0.4507,
2733
+ "step": 389
2734
+ },
2735
+ {
2736
+ "epoch": 3.3917301414581065,
2737
+ "grad_norm": 1.3984375,
2738
+ "learning_rate": 3.308695652173913e-05,
2739
+ "loss": 0.4535,
2740
+ "step": 390
2741
+ },
2742
+ {
2743
+ "epoch": 3.4004352557127313,
2744
+ "grad_norm": 1.296875,
2745
+ "learning_rate": 3.304347826086956e-05,
2746
+ "loss": 0.4736,
2747
+ "step": 391
2748
+ },
2749
+ {
2750
+ "epoch": 3.4091403699673557,
2751
+ "grad_norm": 1.2890625,
2752
+ "learning_rate": 3.3e-05,
2753
+ "loss": 0.3137,
2754
+ "step": 392
2755
+ },
2756
+ {
2757
+ "epoch": 3.4178454842219805,
2758
+ "grad_norm": 1.2109375,
2759
+ "learning_rate": 3.2956521739130436e-05,
2760
+ "loss": 0.335,
2761
+ "step": 393
2762
+ },
2763
+ {
2764
+ "epoch": 3.426550598476605,
2765
+ "grad_norm": 1.328125,
2766
+ "learning_rate": 3.291304347826087e-05,
2767
+ "loss": 0.3644,
2768
+ "step": 394
2769
+ },
2770
+ {
2771
+ "epoch": 3.4352557127312298,
2772
+ "grad_norm": 1.40625,
2773
+ "learning_rate": 3.286956521739131e-05,
2774
+ "loss": 0.3236,
2775
+ "step": 395
2776
+ },
2777
+ {
2778
+ "epoch": 3.443960826985854,
2779
+ "grad_norm": 1.40625,
2780
+ "learning_rate": 3.282608695652174e-05,
2781
+ "loss": 0.6159,
2782
+ "step": 396
2783
+ },
2784
+ {
2785
+ "epoch": 3.452665941240479,
2786
+ "grad_norm": 1.4140625,
2787
+ "learning_rate": 3.278260869565217e-05,
2788
+ "loss": 0.4811,
2789
+ "step": 397
2790
+ },
2791
+ {
2792
+ "epoch": 3.4613710554951034,
2793
+ "grad_norm": 1.1640625,
2794
+ "learning_rate": 3.273913043478261e-05,
2795
+ "loss": 0.3592,
2796
+ "step": 398
2797
+ },
2798
+ {
2799
+ "epoch": 3.4700761697497278,
2800
+ "grad_norm": 1.203125,
2801
+ "learning_rate": 3.269565217391304e-05,
2802
+ "loss": 0.31,
2803
+ "step": 399
2804
+ },
2805
+ {
2806
+ "epoch": 3.4787812840043526,
2807
+ "grad_norm": 1.46875,
2808
+ "learning_rate": 3.2652173913043484e-05,
2809
+ "loss": 0.3305,
2810
+ "step": 400
2811
+ },
2812
+ {
2813
+ "epoch": 3.487486398258977,
2814
+ "grad_norm": 1.4609375,
2815
+ "learning_rate": 3.260869565217392e-05,
2816
+ "loss": 0.4076,
2817
+ "step": 401
2818
+ },
2819
+ {
2820
+ "epoch": 3.496191512513602,
2821
+ "grad_norm": 1.46875,
2822
+ "learning_rate": 3.256521739130435e-05,
2823
+ "loss": 0.4706,
2824
+ "step": 402
2825
+ },
2826
+ {
2827
+ "epoch": 3.504896626768226,
2828
+ "grad_norm": 1.2578125,
2829
+ "learning_rate": 3.2521739130434784e-05,
2830
+ "loss": 0.2767,
2831
+ "step": 403
2832
+ },
2833
+ {
2834
+ "epoch": 3.513601741022851,
2835
+ "grad_norm": 1.28125,
2836
+ "learning_rate": 3.247826086956522e-05,
2837
+ "loss": 0.3916,
2838
+ "step": 404
2839
+ },
2840
+ {
2841
+ "epoch": 3.5223068552774754,
2842
+ "grad_norm": 1.234375,
2843
+ "learning_rate": 3.243478260869565e-05,
2844
+ "loss": 0.3626,
2845
+ "step": 405
2846
+ },
2847
+ {
2848
+ "epoch": 3.5310119695321003,
2849
+ "grad_norm": 1.265625,
2850
+ "learning_rate": 3.239130434782609e-05,
2851
+ "loss": 0.3815,
2852
+ "step": 406
2853
+ },
2854
+ {
2855
+ "epoch": 3.5397170837867247,
2856
+ "grad_norm": 1.1484375,
2857
+ "learning_rate": 3.2347826086956524e-05,
2858
+ "loss": 0.3034,
2859
+ "step": 407
2860
+ },
2861
+ {
2862
+ "epoch": 3.548422198041349,
2863
+ "grad_norm": 1.3046875,
2864
+ "learning_rate": 3.230434782608696e-05,
2865
+ "loss": 0.374,
2866
+ "step": 408
2867
+ },
2868
+ {
2869
+ "epoch": 3.557127312295974,
2870
+ "grad_norm": 1.3828125,
2871
+ "learning_rate": 3.226086956521739e-05,
2872
+ "loss": 0.3537,
2873
+ "step": 409
2874
+ },
2875
+ {
2876
+ "epoch": 3.5658324265505987,
2877
+ "grad_norm": 1.4609375,
2878
+ "learning_rate": 3.2217391304347825e-05,
2879
+ "loss": 0.5427,
2880
+ "step": 410
2881
+ },
2882
+ {
2883
+ "epoch": 3.574537540805223,
2884
+ "grad_norm": 1.3359375,
2885
+ "learning_rate": 3.217391304347826e-05,
2886
+ "loss": 0.3466,
2887
+ "step": 411
2888
+ },
2889
+ {
2890
+ "epoch": 3.5832426550598475,
2891
+ "grad_norm": 1.4453125,
2892
+ "learning_rate": 3.21304347826087e-05,
2893
+ "loss": 0.5072,
2894
+ "step": 412
2895
+ },
2896
+ {
2897
+ "epoch": 3.5919477693144723,
2898
+ "grad_norm": 1.3984375,
2899
+ "learning_rate": 3.208695652173913e-05,
2900
+ "loss": 0.3162,
2901
+ "step": 413
2902
+ },
2903
+ {
2904
+ "epoch": 3.6006528835690967,
2905
+ "grad_norm": 1.1953125,
2906
+ "learning_rate": 3.204347826086957e-05,
2907
+ "loss": 0.51,
2908
+ "step": 414
2909
+ },
2910
+ {
2911
+ "epoch": 3.6093579978237216,
2912
+ "grad_norm": 1.25,
2913
+ "learning_rate": 3.2000000000000005e-05,
2914
+ "loss": 0.4698,
2915
+ "step": 415
2916
+ },
2917
+ {
2918
+ "epoch": 3.618063112078346,
2919
+ "grad_norm": 1.296875,
2920
+ "learning_rate": 3.195652173913043e-05,
2921
+ "loss": 0.3544,
2922
+ "step": 416
2923
+ },
2924
+ {
2925
+ "epoch": 3.6267682263329704,
2926
+ "grad_norm": 1.453125,
2927
+ "learning_rate": 3.191304347826087e-05,
2928
+ "loss": 0.3747,
2929
+ "step": 417
2930
+ },
2931
+ {
2932
+ "epoch": 3.635473340587595,
2933
+ "grad_norm": 1.328125,
2934
+ "learning_rate": 3.1869565217391306e-05,
2935
+ "loss": 0.4795,
2936
+ "step": 418
2937
+ },
2938
+ {
2939
+ "epoch": 3.64417845484222,
2940
+ "grad_norm": 1.2734375,
2941
+ "learning_rate": 3.182608695652174e-05,
2942
+ "loss": 0.5804,
2943
+ "step": 419
2944
+ },
2945
+ {
2946
+ "epoch": 3.6528835690968444,
2947
+ "grad_norm": 1.3125,
2948
+ "learning_rate": 3.178260869565218e-05,
2949
+ "loss": 0.4708,
2950
+ "step": 420
2951
+ },
2952
+ {
2953
+ "epoch": 3.661588683351469,
2954
+ "grad_norm": 1.1796875,
2955
+ "learning_rate": 3.173913043478261e-05,
2956
+ "loss": 0.4494,
2957
+ "step": 421
2958
+ },
2959
+ {
2960
+ "epoch": 3.6702937976060936,
2961
+ "grad_norm": 1.34375,
2962
+ "learning_rate": 3.1695652173913046e-05,
2963
+ "loss": 0.4712,
2964
+ "step": 422
2965
+ },
2966
+ {
2967
+ "epoch": 3.678998911860718,
2968
+ "grad_norm": 1.6953125,
2969
+ "learning_rate": 3.165217391304348e-05,
2970
+ "loss": 0.5315,
2971
+ "step": 423
2972
+ },
2973
+ {
2974
+ "epoch": 3.687704026115343,
2975
+ "grad_norm": 1.2578125,
2976
+ "learning_rate": 3.160869565217391e-05,
2977
+ "loss": 0.3474,
2978
+ "step": 424
2979
+ },
2980
+ {
2981
+ "epoch": 3.6964091403699673,
2982
+ "grad_norm": 1.328125,
2983
+ "learning_rate": 3.1565217391304346e-05,
2984
+ "loss": 0.408,
2985
+ "step": 425
2986
+ },
2987
+ {
2988
+ "epoch": 3.705114254624592,
2989
+ "grad_norm": 1.3828125,
2990
+ "learning_rate": 3.152173913043479e-05,
2991
+ "loss": 0.3594,
2992
+ "step": 426
2993
+ },
2994
+ {
2995
+ "epoch": 3.7138193688792165,
2996
+ "grad_norm": 1.296875,
2997
+ "learning_rate": 3.147826086956522e-05,
2998
+ "loss": 0.3591,
2999
+ "step": 427
3000
+ },
3001
+ {
3002
+ "epoch": 3.7225244831338413,
3003
+ "grad_norm": 1.359375,
3004
+ "learning_rate": 3.1434782608695653e-05,
3005
+ "loss": 0.4131,
3006
+ "step": 428
3007
+ },
3008
+ {
3009
+ "epoch": 3.7312295973884657,
3010
+ "grad_norm": 1.2265625,
3011
+ "learning_rate": 3.139130434782609e-05,
3012
+ "loss": 0.59,
3013
+ "step": 429
3014
+ },
3015
+ {
3016
+ "epoch": 3.73993471164309,
3017
+ "grad_norm": 1.515625,
3018
+ "learning_rate": 3.134782608695652e-05,
3019
+ "loss": 0.5576,
3020
+ "step": 430
3021
+ },
3022
+ {
3023
+ "epoch": 3.748639825897715,
3024
+ "grad_norm": 1.3828125,
3025
+ "learning_rate": 3.130434782608696e-05,
3026
+ "loss": 0.333,
3027
+ "step": 431
3028
+ },
3029
+ {
3030
+ "epoch": 3.7573449401523398,
3031
+ "grad_norm": 1.25,
3032
+ "learning_rate": 3.1260869565217394e-05,
3033
+ "loss": 0.3218,
3034
+ "step": 432
3035
+ },
3036
+ {
3037
+ "epoch": 3.766050054406964,
3038
+ "grad_norm": 1.328125,
3039
+ "learning_rate": 3.121739130434783e-05,
3040
+ "loss": 0.4542,
3041
+ "step": 433
3042
+ },
3043
+ {
3044
+ "epoch": 3.7747551686615886,
3045
+ "grad_norm": 1.5234375,
3046
+ "learning_rate": 3.117391304347827e-05,
3047
+ "loss": 0.497,
3048
+ "step": 434
3049
+ },
3050
+ {
3051
+ "epoch": 3.7834602829162134,
3052
+ "grad_norm": 1.40625,
3053
+ "learning_rate": 3.1130434782608694e-05,
3054
+ "loss": 0.3721,
3055
+ "step": 435
3056
+ },
3057
+ {
3058
+ "epoch": 3.792165397170838,
3059
+ "grad_norm": 1.34375,
3060
+ "learning_rate": 3.108695652173913e-05,
3061
+ "loss": 0.4833,
3062
+ "step": 436
3063
+ },
3064
+ {
3065
+ "epoch": 3.8008705114254626,
3066
+ "grad_norm": 1.484375,
3067
+ "learning_rate": 3.104347826086957e-05,
3068
+ "loss": 0.3413,
3069
+ "step": 437
3070
+ },
3071
+ {
3072
+ "epoch": 3.809575625680087,
3073
+ "grad_norm": 1.1640625,
3074
+ "learning_rate": 3.1e-05,
3075
+ "loss": 0.3731,
3076
+ "step": 438
3077
+ },
3078
+ {
3079
+ "epoch": 3.8182807399347114,
3080
+ "grad_norm": 1.2734375,
3081
+ "learning_rate": 3.0956521739130435e-05,
3082
+ "loss": 0.3438,
3083
+ "step": 439
3084
+ },
3085
+ {
3086
+ "epoch": 3.8269858541893362,
3087
+ "grad_norm": 1.359375,
3088
+ "learning_rate": 3.0913043478260875e-05,
3089
+ "loss": 0.4249,
3090
+ "step": 440
3091
+ },
3092
+ {
3093
+ "epoch": 3.835690968443961,
3094
+ "grad_norm": 1.25,
3095
+ "learning_rate": 3.086956521739131e-05,
3096
+ "loss": 0.4995,
3097
+ "step": 441
3098
+ },
3099
+ {
3100
+ "epoch": 3.8443960826985855,
3101
+ "grad_norm": 1.484375,
3102
+ "learning_rate": 3.0826086956521735e-05,
3103
+ "loss": 0.4477,
3104
+ "step": 442
3105
+ },
3106
+ {
3107
+ "epoch": 3.85310119695321,
3108
+ "grad_norm": 1.25,
3109
+ "learning_rate": 3.0782608695652175e-05,
3110
+ "loss": 0.4823,
3111
+ "step": 443
3112
+ },
3113
+ {
3114
+ "epoch": 3.8618063112078347,
3115
+ "grad_norm": 1.3828125,
3116
+ "learning_rate": 3.073913043478261e-05,
3117
+ "loss": 0.5226,
3118
+ "step": 444
3119
+ },
3120
+ {
3121
+ "epoch": 3.870511425462459,
3122
+ "grad_norm": 1.359375,
3123
+ "learning_rate": 3.069565217391305e-05,
3124
+ "loss": 0.4088,
3125
+ "step": 445
3126
+ },
3127
+ {
3128
+ "epoch": 3.879216539717084,
3129
+ "grad_norm": 1.5625,
3130
+ "learning_rate": 3.065217391304348e-05,
3131
+ "loss": 0.4232,
3132
+ "step": 446
3133
+ },
3134
+ {
3135
+ "epoch": 3.8879216539717083,
3136
+ "grad_norm": 1.4375,
3137
+ "learning_rate": 3.0608695652173916e-05,
3138
+ "loss": 0.3985,
3139
+ "step": 447
3140
+ },
3141
+ {
3142
+ "epoch": 3.8966267682263327,
3143
+ "grad_norm": 1.421875,
3144
+ "learning_rate": 3.056521739130435e-05,
3145
+ "loss": 0.4992,
3146
+ "step": 448
3147
+ },
3148
+ {
3149
+ "epoch": 3.9053318824809575,
3150
+ "grad_norm": 1.4296875,
3151
+ "learning_rate": 3.052173913043478e-05,
3152
+ "loss": 0.4681,
3153
+ "step": 449
3154
+ },
3155
+ {
3156
+ "epoch": 3.9140369967355824,
3157
+ "grad_norm": 1.34375,
3158
+ "learning_rate": 3.0478260869565216e-05,
3159
+ "loss": 0.3519,
3160
+ "step": 450
3161
+ },
3162
+ {
3163
+ "epoch": 3.9227421109902068,
3164
+ "grad_norm": 1.15625,
3165
+ "learning_rate": 3.0434782608695656e-05,
3166
+ "loss": 0.38,
3167
+ "step": 451
3168
+ },
3169
+ {
3170
+ "epoch": 3.931447225244831,
3171
+ "grad_norm": 1.421875,
3172
+ "learning_rate": 3.039130434782609e-05,
3173
+ "loss": 0.5022,
3174
+ "step": 452
3175
+ },
3176
+ {
3177
+ "epoch": 3.940152339499456,
3178
+ "grad_norm": 1.2890625,
3179
+ "learning_rate": 3.034782608695652e-05,
3180
+ "loss": 0.4629,
3181
+ "step": 453
3182
+ },
3183
+ {
3184
+ "epoch": 3.9488574537540804,
3185
+ "grad_norm": 1.328125,
3186
+ "learning_rate": 3.030434782608696e-05,
3187
+ "loss": 0.7028,
3188
+ "step": 454
3189
+ },
3190
+ {
3191
+ "epoch": 3.957562568008705,
3192
+ "grad_norm": 1.15625,
3193
+ "learning_rate": 3.0260869565217393e-05,
3194
+ "loss": 0.3148,
3195
+ "step": 455
3196
+ },
3197
+ {
3198
+ "epoch": 3.9662676822633296,
3199
+ "grad_norm": 1.3046875,
3200
+ "learning_rate": 3.0217391304347827e-05,
3201
+ "loss": 0.5369,
3202
+ "step": 456
3203
+ },
3204
+ {
3205
+ "epoch": 3.9749727965179544,
3206
+ "grad_norm": 1.125,
3207
+ "learning_rate": 3.0173913043478264e-05,
3208
+ "loss": 0.3866,
3209
+ "step": 457
3210
+ },
3211
+ {
3212
+ "epoch": 3.983677910772579,
3213
+ "grad_norm": 1.3203125,
3214
+ "learning_rate": 3.0130434782608697e-05,
3215
+ "loss": 0.5832,
3216
+ "step": 458
3217
+ },
3218
+ {
3219
+ "epoch": 3.9923830250272037,
3220
+ "grad_norm": 1.359375,
3221
+ "learning_rate": 3.008695652173913e-05,
3222
+ "loss": 0.4359,
3223
+ "step": 459
3224
+ },
3225
+ {
3226
+ "epoch": 4.0,
3227
+ "grad_norm": 1.7265625,
3228
+ "learning_rate": 3.0043478260869567e-05,
3229
+ "loss": 0.3593,
3230
+ "step": 460
3231
+ },
3232
+ {
3233
+ "epoch": 4.008705114254624,
3234
+ "grad_norm": 1.4140625,
3235
+ "learning_rate": 3e-05,
3236
+ "loss": 0.3197,
3237
+ "step": 461
3238
+ },
3239
+ {
3240
+ "epoch": 4.017410228509249,
3241
+ "grad_norm": 1.15625,
3242
+ "learning_rate": 2.9956521739130438e-05,
3243
+ "loss": 0.3831,
3244
+ "step": 462
3245
+ },
3246
+ {
3247
+ "epoch": 4.026115342763874,
3248
+ "grad_norm": 1.3828125,
3249
+ "learning_rate": 2.991304347826087e-05,
3250
+ "loss": 0.2362,
3251
+ "step": 463
3252
+ },
3253
+ {
3254
+ "epoch": 4.0348204570184985,
3255
+ "grad_norm": 1.328125,
3256
+ "learning_rate": 2.9869565217391304e-05,
3257
+ "loss": 0.2624,
3258
+ "step": 464
3259
+ },
3260
+ {
3261
+ "epoch": 4.043525571273123,
3262
+ "grad_norm": 1.234375,
3263
+ "learning_rate": 2.982608695652174e-05,
3264
+ "loss": 0.2093,
3265
+ "step": 465
3266
+ },
3267
+ {
3268
+ "epoch": 4.052230685527747,
3269
+ "grad_norm": 1.9296875,
3270
+ "learning_rate": 2.9782608695652175e-05,
3271
+ "loss": 0.4162,
3272
+ "step": 466
3273
+ },
3274
+ {
3275
+ "epoch": 4.0609357997823725,
3276
+ "grad_norm": 1.9921875,
3277
+ "learning_rate": 2.9739130434782608e-05,
3278
+ "loss": 0.286,
3279
+ "step": 467
3280
+ },
3281
+ {
3282
+ "epoch": 4.069640914036997,
3283
+ "grad_norm": 2.28125,
3284
+ "learning_rate": 2.969565217391305e-05,
3285
+ "loss": 0.3429,
3286
+ "step": 468
3287
+ },
3288
+ {
3289
+ "epoch": 4.078346028291621,
3290
+ "grad_norm": 1.7421875,
3291
+ "learning_rate": 2.965217391304348e-05,
3292
+ "loss": 0.2248,
3293
+ "step": 469
3294
+ },
3295
+ {
3296
+ "epoch": 4.087051142546246,
3297
+ "grad_norm": 1.6484375,
3298
+ "learning_rate": 2.9608695652173912e-05,
3299
+ "loss": 0.3577,
3300
+ "step": 470
3301
+ },
3302
+ {
3303
+ "epoch": 4.09575625680087,
3304
+ "grad_norm": 1.453125,
3305
+ "learning_rate": 2.9565217391304352e-05,
3306
+ "loss": 0.2521,
3307
+ "step": 471
3308
+ },
3309
+ {
3310
+ "epoch": 4.104461371055495,
3311
+ "grad_norm": 1.4921875,
3312
+ "learning_rate": 2.9521739130434782e-05,
3313
+ "loss": 0.2246,
3314
+ "step": 472
3315
+ },
3316
+ {
3317
+ "epoch": 4.11316648531012,
3318
+ "grad_norm": 1.7578125,
3319
+ "learning_rate": 2.9478260869565215e-05,
3320
+ "loss": 0.4472,
3321
+ "step": 473
3322
+ },
3323
+ {
3324
+ "epoch": 4.121871599564744,
3325
+ "grad_norm": 1.515625,
3326
+ "learning_rate": 2.9434782608695656e-05,
3327
+ "loss": 0.2405,
3328
+ "step": 474
3329
+ },
3330
+ {
3331
+ "epoch": 4.1305767138193685,
3332
+ "grad_norm": 1.171875,
3333
+ "learning_rate": 2.939130434782609e-05,
3334
+ "loss": 0.2111,
3335
+ "step": 475
3336
+ },
3337
+ {
3338
+ "epoch": 4.139281828073994,
3339
+ "grad_norm": 1.171875,
3340
+ "learning_rate": 2.9347826086956526e-05,
3341
+ "loss": 0.3421,
3342
+ "step": 476
3343
+ },
3344
+ {
3345
+ "epoch": 4.147986942328618,
3346
+ "grad_norm": 1.359375,
3347
+ "learning_rate": 2.930434782608696e-05,
3348
+ "loss": 0.343,
3349
+ "step": 477
3350
+ },
3351
+ {
3352
+ "epoch": 4.156692056583243,
3353
+ "grad_norm": 1.390625,
3354
+ "learning_rate": 2.9260869565217393e-05,
3355
+ "loss": 0.1912,
3356
+ "step": 478
3357
+ },
3358
+ {
3359
+ "epoch": 4.165397170837867,
3360
+ "grad_norm": 1.8203125,
3361
+ "learning_rate": 2.921739130434783e-05,
3362
+ "loss": 0.3702,
3363
+ "step": 479
3364
+ },
3365
+ {
3366
+ "epoch": 4.174102285092492,
3367
+ "grad_norm": 1.828125,
3368
+ "learning_rate": 2.9173913043478263e-05,
3369
+ "loss": 0.2608,
3370
+ "step": 480
3371
+ },
3372
+ {
3373
+ "epoch": 4.182807399347117,
3374
+ "grad_norm": 1.4375,
3375
+ "learning_rate": 2.9130434782608696e-05,
3376
+ "loss": 0.2425,
3377
+ "step": 481
3378
+ },
3379
+ {
3380
+ "epoch": 4.191512513601741,
3381
+ "grad_norm": 1.5,
3382
+ "learning_rate": 2.9086956521739133e-05,
3383
+ "loss": 0.2814,
3384
+ "step": 482
3385
+ },
3386
+ {
3387
+ "epoch": 4.200217627856365,
3388
+ "grad_norm": 1.65625,
3389
+ "learning_rate": 2.9043478260869567e-05,
3390
+ "loss": 0.2904,
3391
+ "step": 483
3392
+ },
3393
+ {
3394
+ "epoch": 4.20892274211099,
3395
+ "grad_norm": 1.4765625,
3396
+ "learning_rate": 2.9e-05,
3397
+ "loss": 0.275,
3398
+ "step": 484
3399
+ },
3400
+ {
3401
+ "epoch": 4.217627856365615,
3402
+ "grad_norm": 1.390625,
3403
+ "learning_rate": 2.8956521739130437e-05,
3404
+ "loss": 0.2462,
3405
+ "step": 485
3406
+ },
3407
+ {
3408
+ "epoch": 4.2263329706202395,
3409
+ "grad_norm": 1.5234375,
3410
+ "learning_rate": 2.891304347826087e-05,
3411
+ "loss": 0.262,
3412
+ "step": 486
3413
+ },
3414
+ {
3415
+ "epoch": 4.235038084874864,
3416
+ "grad_norm": 1.25,
3417
+ "learning_rate": 2.8869565217391304e-05,
3418
+ "loss": 0.2443,
3419
+ "step": 487
3420
+ },
3421
+ {
3422
+ "epoch": 4.243743199129488,
3423
+ "grad_norm": 1.4140625,
3424
+ "learning_rate": 2.882608695652174e-05,
3425
+ "loss": 0.3418,
3426
+ "step": 488
3427
+ },
3428
+ {
3429
+ "epoch": 4.2524483133841136,
3430
+ "grad_norm": 1.4140625,
3431
+ "learning_rate": 2.8782608695652174e-05,
3432
+ "loss": 0.2845,
3433
+ "step": 489
3434
+ },
3435
+ {
3436
+ "epoch": 4.261153427638738,
3437
+ "grad_norm": 1.671875,
3438
+ "learning_rate": 2.8739130434782608e-05,
3439
+ "loss": 0.3489,
3440
+ "step": 490
3441
+ },
3442
+ {
3443
+ "epoch": 4.269858541893362,
3444
+ "grad_norm": 1.421875,
3445
+ "learning_rate": 2.8695652173913044e-05,
3446
+ "loss": 0.2591,
3447
+ "step": 491
3448
+ },
3449
+ {
3450
+ "epoch": 4.278563656147987,
3451
+ "grad_norm": 1.5390625,
3452
+ "learning_rate": 2.8652173913043478e-05,
3453
+ "loss": 0.5734,
3454
+ "step": 492
3455
+ },
3456
+ {
3457
+ "epoch": 4.287268770402612,
3458
+ "grad_norm": 1.4296875,
3459
+ "learning_rate": 2.8608695652173918e-05,
3460
+ "loss": 0.2109,
3461
+ "step": 493
3462
+ },
3463
+ {
3464
+ "epoch": 4.295973884657236,
3465
+ "grad_norm": 1.5,
3466
+ "learning_rate": 2.8565217391304348e-05,
3467
+ "loss": 0.2421,
3468
+ "step": 494
3469
+ },
3470
+ {
3471
+ "epoch": 4.304678998911861,
3472
+ "grad_norm": 1.5390625,
3473
+ "learning_rate": 2.852173913043478e-05,
3474
+ "loss": 0.4073,
3475
+ "step": 495
3476
+ },
3477
+ {
3478
+ "epoch": 4.313384113166485,
3479
+ "grad_norm": 1.5546875,
3480
+ "learning_rate": 2.847826086956522e-05,
3481
+ "loss": 0.2709,
3482
+ "step": 496
3483
+ },
3484
+ {
3485
+ "epoch": 4.32208922742111,
3486
+ "grad_norm": 1.453125,
3487
+ "learning_rate": 2.8434782608695655e-05,
3488
+ "loss": 0.2862,
3489
+ "step": 497
3490
+ },
3491
+ {
3492
+ "epoch": 4.330794341675735,
3493
+ "grad_norm": 1.390625,
3494
+ "learning_rate": 2.8391304347826085e-05,
3495
+ "loss": 0.5668,
3496
+ "step": 498
3497
+ },
3498
+ {
3499
+ "epoch": 4.339499455930359,
3500
+ "grad_norm": 1.7421875,
3501
+ "learning_rate": 2.8347826086956525e-05,
3502
+ "loss": 0.53,
3503
+ "step": 499
3504
+ },
3505
+ {
3506
+ "epoch": 4.348204570184984,
3507
+ "grad_norm": 1.40625,
3508
+ "learning_rate": 2.830434782608696e-05,
3509
+ "loss": 0.2408,
3510
+ "step": 500
3511
+ },
3512
+ {
3513
+ "epoch": 4.356909684439608,
3514
+ "grad_norm": 1.265625,
3515
+ "learning_rate": 2.826086956521739e-05,
3516
+ "loss": 0.2201,
3517
+ "step": 501
3518
+ },
3519
+ {
3520
+ "epoch": 4.365614798694233,
3521
+ "grad_norm": 1.2890625,
3522
+ "learning_rate": 2.821739130434783e-05,
3523
+ "loss": 0.2811,
3524
+ "step": 502
3525
+ },
3526
+ {
3527
+ "epoch": 4.374319912948858,
3528
+ "grad_norm": 1.484375,
3529
+ "learning_rate": 2.8173913043478262e-05,
3530
+ "loss": 0.2138,
3531
+ "step": 503
3532
+ },
3533
+ {
3534
+ "epoch": 4.383025027203482,
3535
+ "grad_norm": 1.3203125,
3536
+ "learning_rate": 2.8130434782608696e-05,
3537
+ "loss": 0.2542,
3538
+ "step": 504
3539
+ },
3540
+ {
3541
+ "epoch": 4.3917301414581065,
3542
+ "grad_norm": 1.640625,
3543
+ "learning_rate": 2.8086956521739133e-05,
3544
+ "loss": 0.289,
3545
+ "step": 505
3546
+ },
3547
+ {
3548
+ "epoch": 4.400435255712731,
3549
+ "grad_norm": 1.5078125,
3550
+ "learning_rate": 2.8043478260869566e-05,
3551
+ "loss": 0.2872,
3552
+ "step": 506
3553
+ },
3554
+ {
3555
+ "epoch": 4.409140369967356,
3556
+ "grad_norm": 1.3515625,
3557
+ "learning_rate": 2.8000000000000003e-05,
3558
+ "loss": 0.3546,
3559
+ "step": 507
3560
+ },
3561
+ {
3562
+ "epoch": 4.4178454842219805,
3563
+ "grad_norm": 1.546875,
3564
+ "learning_rate": 2.7956521739130436e-05,
3565
+ "loss": 0.2645,
3566
+ "step": 508
3567
+ },
3568
+ {
3569
+ "epoch": 4.426550598476605,
3570
+ "grad_norm": 1.2890625,
3571
+ "learning_rate": 2.791304347826087e-05,
3572
+ "loss": 0.3288,
3573
+ "step": 509
3574
+ },
3575
+ {
3576
+ "epoch": 4.435255712731229,
3577
+ "grad_norm": 1.6015625,
3578
+ "learning_rate": 2.7869565217391307e-05,
3579
+ "loss": 0.3309,
3580
+ "step": 510
3581
+ },
3582
+ {
3583
+ "epoch": 4.443960826985855,
3584
+ "grad_norm": 1.390625,
3585
+ "learning_rate": 2.782608695652174e-05,
3586
+ "loss": 0.2221,
3587
+ "step": 511
3588
+ },
3589
+ {
3590
+ "epoch": 4.452665941240479,
3591
+ "grad_norm": 1.1796875,
3592
+ "learning_rate": 2.7782608695652174e-05,
3593
+ "loss": 0.3022,
3594
+ "step": 512
3595
+ },
3596
+ {
3597
+ "epoch": 4.461371055495103,
3598
+ "grad_norm": 1.484375,
3599
+ "learning_rate": 2.773913043478261e-05,
3600
+ "loss": 0.2131,
3601
+ "step": 513
3602
+ },
3603
+ {
3604
+ "epoch": 4.470076169749728,
3605
+ "grad_norm": 1.3046875,
3606
+ "learning_rate": 2.7695652173913044e-05,
3607
+ "loss": 0.3289,
3608
+ "step": 514
3609
+ },
3610
+ {
3611
+ "epoch": 4.478781284004352,
3612
+ "grad_norm": 1.2890625,
3613
+ "learning_rate": 2.7652173913043477e-05,
3614
+ "loss": 0.1957,
3615
+ "step": 515
3616
+ },
3617
+ {
3618
+ "epoch": 4.487486398258977,
3619
+ "grad_norm": 1.578125,
3620
+ "learning_rate": 2.7608695652173917e-05,
3621
+ "loss": 0.3898,
3622
+ "step": 516
3623
+ },
3624
+ {
3625
+ "epoch": 4.496191512513602,
3626
+ "grad_norm": 1.2890625,
3627
+ "learning_rate": 2.7565217391304347e-05,
3628
+ "loss": 0.28,
3629
+ "step": 517
3630
+ },
3631
+ {
3632
+ "epoch": 4.504896626768226,
3633
+ "grad_norm": 1.328125,
3634
+ "learning_rate": 2.752173913043478e-05,
3635
+ "loss": 0.348,
3636
+ "step": 518
3637
+ },
3638
+ {
3639
+ "epoch": 4.513601741022851,
3640
+ "grad_norm": 1.546875,
3641
+ "learning_rate": 2.747826086956522e-05,
3642
+ "loss": 0.4128,
3643
+ "step": 519
3644
+ },
3645
+ {
3646
+ "epoch": 4.522306855277476,
3647
+ "grad_norm": 1.2421875,
3648
+ "learning_rate": 2.743478260869565e-05,
3649
+ "loss": 0.294,
3650
+ "step": 520
3651
+ },
3652
+ {
3653
+ "epoch": 4.5310119695321,
3654
+ "grad_norm": 1.609375,
3655
+ "learning_rate": 2.7391304347826085e-05,
3656
+ "loss": 0.1959,
3657
+ "step": 521
3658
+ },
3659
+ {
3660
+ "epoch": 4.539717083786725,
3661
+ "grad_norm": 1.4609375,
3662
+ "learning_rate": 2.7347826086956525e-05,
3663
+ "loss": 0.2295,
3664
+ "step": 522
3665
+ },
3666
+ {
3667
+ "epoch": 4.548422198041349,
3668
+ "grad_norm": 1.6796875,
3669
+ "learning_rate": 2.7304347826086958e-05,
3670
+ "loss": 0.5485,
3671
+ "step": 523
3672
+ },
3673
+ {
3674
+ "epoch": 4.5571273122959735,
3675
+ "grad_norm": 1.4765625,
3676
+ "learning_rate": 2.7260869565217395e-05,
3677
+ "loss": 0.2739,
3678
+ "step": 524
3679
+ },
3680
+ {
3681
+ "epoch": 4.565832426550599,
3682
+ "grad_norm": 1.421875,
3683
+ "learning_rate": 2.721739130434783e-05,
3684
+ "loss": 0.3562,
3685
+ "step": 525
3686
+ },
3687
+ {
3688
+ "epoch": 4.574537540805223,
3689
+ "grad_norm": 1.4921875,
3690
+ "learning_rate": 2.7173913043478262e-05,
3691
+ "loss": 0.4189,
3692
+ "step": 526
3693
+ },
3694
+ {
3695
+ "epoch": 4.5832426550598475,
3696
+ "grad_norm": 1.453125,
3697
+ "learning_rate": 2.71304347826087e-05,
3698
+ "loss": 0.4434,
3699
+ "step": 527
3700
+ },
3701
+ {
3702
+ "epoch": 4.591947769314472,
3703
+ "grad_norm": 1.453125,
3704
+ "learning_rate": 2.7086956521739132e-05,
3705
+ "loss": 0.3395,
3706
+ "step": 528
3707
+ },
3708
+ {
3709
+ "epoch": 4.600652883569097,
3710
+ "grad_norm": 1.5078125,
3711
+ "learning_rate": 2.7043478260869566e-05,
3712
+ "loss": 0.3857,
3713
+ "step": 529
3714
+ },
3715
+ {
3716
+ "epoch": 4.609357997823722,
3717
+ "grad_norm": 1.5703125,
3718
+ "learning_rate": 2.7000000000000002e-05,
3719
+ "loss": 0.2454,
3720
+ "step": 530
3721
+ },
3722
+ {
3723
+ "epoch": 4.618063112078346,
3724
+ "grad_norm": 1.546875,
3725
+ "learning_rate": 2.6956521739130436e-05,
3726
+ "loss": 0.2801,
3727
+ "step": 531
3728
+ },
3729
+ {
3730
+ "epoch": 4.62676822633297,
3731
+ "grad_norm": 1.796875,
3732
+ "learning_rate": 2.691304347826087e-05,
3733
+ "loss": 0.4674,
3734
+ "step": 532
3735
+ },
3736
+ {
3737
+ "epoch": 4.635473340587595,
3738
+ "grad_norm": 1.4296875,
3739
+ "learning_rate": 2.6869565217391306e-05,
3740
+ "loss": 0.2464,
3741
+ "step": 533
3742
+ },
3743
+ {
3744
+ "epoch": 4.64417845484222,
3745
+ "grad_norm": 1.3828125,
3746
+ "learning_rate": 2.682608695652174e-05,
3747
+ "loss": 0.2842,
3748
+ "step": 534
3749
+ },
3750
+ {
3751
+ "epoch": 4.652883569096844,
3752
+ "grad_norm": 1.3515625,
3753
+ "learning_rate": 2.6782608695652173e-05,
3754
+ "loss": 0.3065,
3755
+ "step": 535
3756
+ },
3757
+ {
3758
+ "epoch": 4.661588683351469,
3759
+ "grad_norm": 1.5078125,
3760
+ "learning_rate": 2.673913043478261e-05,
3761
+ "loss": 0.2565,
3762
+ "step": 536
3763
+ },
3764
+ {
3765
+ "epoch": 4.670293797606093,
3766
+ "grad_norm": 1.46875,
3767
+ "learning_rate": 2.6695652173913043e-05,
3768
+ "loss": 0.2242,
3769
+ "step": 537
3770
+ },
3771
+ {
3772
+ "epoch": 4.6789989118607185,
3773
+ "grad_norm": 1.3125,
3774
+ "learning_rate": 2.6652173913043483e-05,
3775
+ "loss": 0.3663,
3776
+ "step": 538
3777
+ },
3778
+ {
3779
+ "epoch": 4.687704026115343,
3780
+ "grad_norm": 1.46875,
3781
+ "learning_rate": 2.6608695652173913e-05,
3782
+ "loss": 0.3065,
3783
+ "step": 539
3784
+ },
3785
+ {
3786
+ "epoch": 4.696409140369967,
3787
+ "grad_norm": 1.5859375,
3788
+ "learning_rate": 2.6565217391304347e-05,
3789
+ "loss": 0.3606,
3790
+ "step": 540
3791
+ },
3792
+ {
3793
+ "epoch": 4.705114254624592,
3794
+ "grad_norm": 1.65625,
3795
+ "learning_rate": 2.6521739130434787e-05,
3796
+ "loss": 0.4462,
3797
+ "step": 541
3798
+ },
3799
+ {
3800
+ "epoch": 4.713819368879216,
3801
+ "grad_norm": 1.390625,
3802
+ "learning_rate": 2.647826086956522e-05,
3803
+ "loss": 0.257,
3804
+ "step": 542
3805
+ },
3806
+ {
3807
+ "epoch": 4.722524483133841,
3808
+ "grad_norm": 1.7421875,
3809
+ "learning_rate": 2.643478260869565e-05,
3810
+ "loss": 0.4824,
3811
+ "step": 543
3812
+ },
3813
+ {
3814
+ "epoch": 4.731229597388466,
3815
+ "grad_norm": 1.359375,
3816
+ "learning_rate": 2.639130434782609e-05,
3817
+ "loss": 0.3473,
3818
+ "step": 544
3819
+ },
3820
+ {
3821
+ "epoch": 4.73993471164309,
3822
+ "grad_norm": 1.6171875,
3823
+ "learning_rate": 2.6347826086956524e-05,
3824
+ "loss": 0.4204,
3825
+ "step": 545
3826
+ },
3827
+ {
3828
+ "epoch": 4.748639825897715,
3829
+ "grad_norm": 1.6484375,
3830
+ "learning_rate": 2.6304347826086954e-05,
3831
+ "loss": 0.3148,
3832
+ "step": 546
3833
+ },
3834
+ {
3835
+ "epoch": 4.75734494015234,
3836
+ "grad_norm": 1.4921875,
3837
+ "learning_rate": 2.6260869565217394e-05,
3838
+ "loss": 0.368,
3839
+ "step": 547
3840
+ },
3841
+ {
3842
+ "epoch": 4.766050054406964,
3843
+ "grad_norm": 1.6015625,
3844
+ "learning_rate": 2.6217391304347828e-05,
3845
+ "loss": 0.3937,
3846
+ "step": 548
3847
+ },
3848
+ {
3849
+ "epoch": 4.7747551686615886,
3850
+ "grad_norm": 1.4375,
3851
+ "learning_rate": 2.617391304347826e-05,
3852
+ "loss": 0.4529,
3853
+ "step": 549
3854
+ },
3855
+ {
3856
+ "epoch": 4.783460282916213,
3857
+ "grad_norm": 1.2890625,
3858
+ "learning_rate": 2.6130434782608698e-05,
3859
+ "loss": 0.2479,
3860
+ "step": 550
3861
+ },
3862
+ {
3863
+ "epoch": 4.792165397170838,
3864
+ "grad_norm": 1.5,
3865
+ "learning_rate": 2.608695652173913e-05,
3866
+ "loss": 0.348,
3867
+ "step": 551
3868
+ },
3869
+ {
3870
+ "epoch": 4.800870511425463,
3871
+ "grad_norm": 1.390625,
3872
+ "learning_rate": 2.6043478260869565e-05,
3873
+ "loss": 0.2708,
3874
+ "step": 552
3875
+ },
3876
+ {
3877
+ "epoch": 4.809575625680087,
3878
+ "grad_norm": 1.359375,
3879
+ "learning_rate": 2.6000000000000002e-05,
3880
+ "loss": 0.3292,
3881
+ "step": 553
3882
+ },
3883
+ {
3884
+ "epoch": 4.818280739934711,
3885
+ "grad_norm": 1.3359375,
3886
+ "learning_rate": 2.5956521739130435e-05,
3887
+ "loss": 0.2596,
3888
+ "step": 554
3889
+ },
3890
+ {
3891
+ "epoch": 4.826985854189337,
3892
+ "grad_norm": 1.4921875,
3893
+ "learning_rate": 2.5913043478260872e-05,
3894
+ "loss": 0.4254,
3895
+ "step": 555
3896
+ },
3897
+ {
3898
+ "epoch": 4.835690968443961,
3899
+ "grad_norm": 1.1484375,
3900
+ "learning_rate": 2.5869565217391305e-05,
3901
+ "loss": 0.2204,
3902
+ "step": 556
3903
+ },
3904
+ {
3905
+ "epoch": 4.8443960826985855,
3906
+ "grad_norm": 1.7109375,
3907
+ "learning_rate": 2.582608695652174e-05,
3908
+ "loss": 0.3869,
3909
+ "step": 557
3910
+ },
3911
+ {
3912
+ "epoch": 4.85310119695321,
3913
+ "grad_norm": 1.796875,
3914
+ "learning_rate": 2.5782608695652176e-05,
3915
+ "loss": 0.3907,
3916
+ "step": 558
3917
+ },
3918
+ {
3919
+ "epoch": 4.861806311207834,
3920
+ "grad_norm": 1.390625,
3921
+ "learning_rate": 2.573913043478261e-05,
3922
+ "loss": 0.2937,
3923
+ "step": 559
3924
+ },
3925
+ {
3926
+ "epoch": 4.8705114254624595,
3927
+ "grad_norm": 1.796875,
3928
+ "learning_rate": 2.5695652173913043e-05,
3929
+ "loss": 0.4582,
3930
+ "step": 560
3931
+ },
3932
+ {
3933
+ "epoch": 4.879216539717084,
3934
+ "grad_norm": 1.5625,
3935
+ "learning_rate": 2.5652173913043483e-05,
3936
+ "loss": 0.5466,
3937
+ "step": 561
3938
+ },
3939
+ {
3940
+ "epoch": 4.887921653971708,
3941
+ "grad_norm": 1.7265625,
3942
+ "learning_rate": 2.5608695652173913e-05,
3943
+ "loss": 0.2614,
3944
+ "step": 562
3945
+ },
3946
+ {
3947
+ "epoch": 4.896626768226333,
3948
+ "grad_norm": 1.5390625,
3949
+ "learning_rate": 2.5565217391304346e-05,
3950
+ "loss": 0.3748,
3951
+ "step": 563
3952
+ },
3953
+ {
3954
+ "epoch": 4.905331882480958,
3955
+ "grad_norm": 1.7578125,
3956
+ "learning_rate": 2.5521739130434786e-05,
3957
+ "loss": 0.5146,
3958
+ "step": 564
3959
+ },
3960
+ {
3961
+ "epoch": 4.914036996735582,
3962
+ "grad_norm": 1.5234375,
3963
+ "learning_rate": 2.5478260869565217e-05,
3964
+ "loss": 0.2761,
3965
+ "step": 565
3966
+ },
3967
+ {
3968
+ "epoch": 4.922742110990207,
3969
+ "grad_norm": 1.265625,
3970
+ "learning_rate": 2.543478260869565e-05,
3971
+ "loss": 0.2113,
3972
+ "step": 566
3973
+ },
3974
+ {
3975
+ "epoch": 4.931447225244831,
3976
+ "grad_norm": 1.4453125,
3977
+ "learning_rate": 2.539130434782609e-05,
3978
+ "loss": 0.2445,
3979
+ "step": 567
3980
+ },
3981
+ {
3982
+ "epoch": 4.9401523394994555,
3983
+ "grad_norm": 1.3515625,
3984
+ "learning_rate": 2.5347826086956524e-05,
3985
+ "loss": 0.2419,
3986
+ "step": 568
3987
+ },
3988
+ {
3989
+ "epoch": 4.948857453754081,
3990
+ "grad_norm": 1.3359375,
3991
+ "learning_rate": 2.530434782608696e-05,
3992
+ "loss": 0.205,
3993
+ "step": 569
3994
+ },
3995
+ {
3996
+ "epoch": 4.957562568008705,
3997
+ "grad_norm": 1.5546875,
3998
+ "learning_rate": 2.5260869565217394e-05,
3999
+ "loss": 0.2996,
4000
+ "step": 570
4001
+ },
4002
+ {
4003
+ "epoch": 4.96626768226333,
4004
+ "grad_norm": 1.3671875,
4005
+ "learning_rate": 2.5217391304347827e-05,
4006
+ "loss": 0.4251,
4007
+ "step": 571
4008
+ },
4009
+ {
4010
+ "epoch": 4.974972796517954,
4011
+ "grad_norm": 1.5546875,
4012
+ "learning_rate": 2.5173913043478264e-05,
4013
+ "loss": 0.3702,
4014
+ "step": 572
4015
+ },
4016
+ {
4017
+ "epoch": 4.983677910772579,
4018
+ "grad_norm": 2.015625,
4019
+ "learning_rate": 2.5130434782608698e-05,
4020
+ "loss": 0.3779,
4021
+ "step": 573
4022
+ },
4023
+ {
4024
+ "epoch": 4.992383025027204,
4025
+ "grad_norm": 1.28125,
4026
+ "learning_rate": 2.508695652173913e-05,
4027
+ "loss": 0.4117,
4028
+ "step": 574
4029
+ },
4030
+ {
4031
+ "epoch": 5.0,
4032
+ "grad_norm": 1.5234375,
4033
+ "learning_rate": 2.5043478260869568e-05,
4034
+ "loss": 0.3301,
4035
+ "step": 575
4036
+ }
4037
+ ],
4038
+ "logging_steps": 1,
4039
+ "max_steps": 1150,
4040
+ "num_input_tokens_seen": 0,
4041
+ "num_train_epochs": 10,
4042
+ "save_steps": 500,
4043
+ "stateful_callbacks": {
4044
+ "TrainerControl": {
4045
+ "args": {
4046
+ "should_epoch_stop": false,
4047
+ "should_evaluate": false,
4048
+ "should_log": false,
4049
+ "should_save": true,
4050
+ "should_training_stop": false
4051
+ },
4052
+ "attributes": {}
4053
+ }
4054
+ },
4055
+ "total_flos": 2.959449338884915e+17,
4056
+ "train_batch_size": 4,
4057
+ "trial_name": null,
4058
+ "trial_params": null
4059
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9eb33b570a2d53cc411ba4cb3dd8961f6d08bd5390b1b0d416c95bd11ce4b6b
3
+ size 5905
vocab.json ADDED
The diff for this file is too large to render. See raw diff