Hugging Face's logo

Devy1
/
Qwen2.5-Coder-CONTROL-checkpoints_multi_language_2k-1.5B-Base-3

Safetensors
qwen2
Model card Files
xet
 Community
Devy1 commited on
Commit
9a130d2
·
verified ·
1 Parent(s): f61cbdb

Upload checkpoint Qwen2.5-Coder-CONTROL-checkpoints_multi_language_2k-1.5B-Base-3 at epoch 3

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 +2848 -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 sha256:8edd2d43bafee4a73ddf47b0006372190ed4b615be4c2431ab69d530b1b34c11
3
+ size 3087467144
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0757a07d1ee2bbb42bbc5562feb58c8fdc9c1e446231b632af72c0edea51a6b
3
+ size 6175153765
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4a9f217e852f439efa6bd32fde98d6867f11aa6ea13ddc021ba10af6a0b0934
3
+ size 14645
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c556fecee0c9f34b42faca8f8e424bfae608ed5090d6e7ff15c1d7387bb2f477
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 +1,2848 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 3.0,
6
+ "eval_steps": 500,
7
+ "global_step": 402,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.007511737089201878,
14
+ "grad_norm": 1.5546875,
15
+ "learning_rate": 5e-05,
16
+ "loss": 0.7059,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.015023474178403756,
21
+ "grad_norm": 2.046875,
22
+ "learning_rate": 4.996268656716418e-05,
23
+ "loss": 0.8614,
24
+ "step": 2
25
+ },
26
+ {
27
+ "epoch": 0.022535211267605635,
28
+ "grad_norm": 14.375,
29
+ "learning_rate": 4.992537313432836e-05,
30
+ "loss": 0.799,
31
+ "step": 3
32
+ },
33
+ {
34
+ "epoch": 0.03004694835680751,
35
+ "grad_norm": 2.421875,
36
+ "learning_rate": 4.988805970149254e-05,
37
+ "loss": 0.7599,
38
+ "step": 4
39
+ },
40
+ {
41
+ "epoch": 0.03755868544600939,
42
+ "grad_norm": 1.7890625,
43
+ "learning_rate": 4.985074626865672e-05,
44
+ "loss": 0.7254,
45
+ "step": 5
46
+ },
47
+ {
48
+ "epoch": 0.04507042253521127,
49
+ "grad_norm": 1.4375,
50
+ "learning_rate": 4.98134328358209e-05,
51
+ "loss": 0.7982,
52
+ "step": 6
53
+ },
54
+ {
55
+ "epoch": 0.05258215962441314,
56
+ "grad_norm": 1.546875,
57
+ "learning_rate": 4.977611940298508e-05,
58
+ "loss": 0.7192,
59
+ "step": 7
60
+ },
61
+ {
62
+ "epoch": 0.06009389671361502,
63
+ "grad_norm": 2.71875,
64
+ "learning_rate": 4.973880597014925e-05,
65
+ "loss": 0.8359,
66
+ "step": 8
67
+ },
68
+ {
69
+ "epoch": 0.0676056338028169,
70
+ "grad_norm": 1.4140625,
71
+ "learning_rate": 4.9701492537313436e-05,
72
+ "loss": 0.6135,
73
+ "step": 9
74
+ },
75
+ {
76
+ "epoch": 0.07511737089201878,
77
+ "grad_norm": 3.96875,
78
+ "learning_rate": 4.966417910447762e-05,
79
+ "loss": 1.0456,
80
+ "step": 10
81
+ },
82
+ {
83
+ "epoch": 0.08262910798122065,
84
+ "grad_norm": 1.6171875,
85
+ "learning_rate": 4.9626865671641794e-05,
86
+ "loss": 0.6937,
87
+ "step": 11
88
+ },
89
+ {
90
+ "epoch": 0.09014084507042254,
91
+ "grad_norm": 1.3828125,
92
+ "learning_rate": 4.9589552238805977e-05,
93
+ "loss": 0.6935,
94
+ "step": 12
95
+ },
96
+ {
97
+ "epoch": 0.09765258215962441,
98
+ "grad_norm": 1.7265625,
99
+ "learning_rate": 4.955223880597015e-05,
100
+ "loss": 0.9582,
101
+ "step": 13
102
+ },
103
+ {
104
+ "epoch": 0.10516431924882629,
105
+ "grad_norm": 1.6015625,
106
+ "learning_rate": 4.951492537313433e-05,
107
+ "loss": 0.6369,
108
+ "step": 14
109
+ },
110
+ {
111
+ "epoch": 0.11267605633802817,
112
+ "grad_norm": 1.4453125,
113
+ "learning_rate": 4.9477611940298504e-05,
114
+ "loss": 0.8523,
115
+ "step": 15
116
+ },
117
+ {
118
+ "epoch": 0.12018779342723004,
119
+ "grad_norm": 1.234375,
120
+ "learning_rate": 4.944029850746269e-05,
121
+ "loss": 0.7572,
122
+ "step": 16
123
+ },
124
+ {
125
+ "epoch": 0.12769953051643193,
126
+ "grad_norm": 1.6171875,
127
+ "learning_rate": 4.940298507462687e-05,
128
+ "loss": 0.8729,
129
+ "step": 17
130
+ },
131
+ {
132
+ "epoch": 0.1352112676056338,
133
+ "grad_norm": 1.4296875,
134
+ "learning_rate": 4.9365671641791045e-05,
135
+ "loss": 0.7657,
136
+ "step": 18
137
+ },
138
+ {
139
+ "epoch": 0.14272300469483568,
140
+ "grad_norm": 1.5546875,
141
+ "learning_rate": 4.932835820895523e-05,
142
+ "loss": 0.7701,
143
+ "step": 19
144
+ },
145
+ {
146
+ "epoch": 0.15023474178403756,
147
+ "grad_norm": 1.421875,
148
+ "learning_rate": 4.92910447761194e-05,
149
+ "loss": 0.8063,
150
+ "step": 20
151
+ },
152
+ {
153
+ "epoch": 0.15774647887323945,
154
+ "grad_norm": 1.203125,
155
+ "learning_rate": 4.9253731343283586e-05,
156
+ "loss": 0.6274,
157
+ "step": 21
158
+ },
159
+ {
160
+ "epoch": 0.1652582159624413,
161
+ "grad_norm": 1.0234375,
162
+ "learning_rate": 4.921641791044777e-05,
163
+ "loss": 0.6862,
164
+ "step": 22
165
+ },
166
+ {
167
+ "epoch": 0.1727699530516432,
168
+ "grad_norm": 1.3203125,
169
+ "learning_rate": 4.9179104477611944e-05,
170
+ "loss": 0.8265,
171
+ "step": 23
172
+ },
173
+ {
174
+ "epoch": 0.18028169014084508,
175
+ "grad_norm": 1.3671875,
176
+ "learning_rate": 4.914179104477612e-05,
177
+ "loss": 0.6659,
178
+ "step": 24
179
+ },
180
+ {
181
+ "epoch": 0.18779342723004694,
182
+ "grad_norm": 1.34375,
183
+ "learning_rate": 4.91044776119403e-05,
184
+ "loss": 0.6795,
185
+ "step": 25
186
+ },
187
+ {
188
+ "epoch": 0.19530516431924883,
189
+ "grad_norm": 3.734375,
190
+ "learning_rate": 4.906716417910448e-05,
191
+ "loss": 0.7914,
192
+ "step": 26
193
+ },
194
+ {
195
+ "epoch": 0.2028169014084507,
196
+ "grad_norm": 1.578125,
197
+ "learning_rate": 4.902985074626866e-05,
198
+ "loss": 0.7915,
199
+ "step": 27
200
+ },
201
+ {
202
+ "epoch": 0.21032863849765257,
203
+ "grad_norm": 1.078125,
204
+ "learning_rate": 4.899253731343284e-05,
205
+ "loss": 0.6433,
206
+ "step": 28
207
+ },
208
+ {
209
+ "epoch": 0.21784037558685446,
210
+ "grad_norm": 1.2265625,
211
+ "learning_rate": 4.895522388059702e-05,
212
+ "loss": 0.8444,
213
+ "step": 29
214
+ },
215
+ {
216
+ "epoch": 0.22535211267605634,
217
+ "grad_norm": 1.328125,
218
+ "learning_rate": 4.8917910447761195e-05,
219
+ "loss": 0.7353,
220
+ "step": 30
221
+ },
222
+ {
223
+ "epoch": 0.23286384976525823,
224
+ "grad_norm": 1.0859375,
225
+ "learning_rate": 4.888059701492538e-05,
226
+ "loss": 0.6133,
227
+ "step": 31
228
+ },
229
+ {
230
+ "epoch": 0.2403755868544601,
231
+ "grad_norm": 1.4375,
232
+ "learning_rate": 4.884328358208955e-05,
233
+ "loss": 0.671,
234
+ "step": 32
235
+ },
236
+ {
237
+ "epoch": 0.24788732394366197,
238
+ "grad_norm": 1.1484375,
239
+ "learning_rate": 4.8805970149253735e-05,
240
+ "loss": 0.7712,
241
+ "step": 33
242
+ },
243
+ {
244
+ "epoch": 0.25539906103286386,
245
+ "grad_norm": 1.4765625,
246
+ "learning_rate": 4.876865671641792e-05,
247
+ "loss": 0.8207,
248
+ "step": 34
249
+ },
250
+ {
251
+ "epoch": 0.26291079812206575,
252
+ "grad_norm": 1.28125,
253
+ "learning_rate": 4.8731343283582094e-05,
254
+ "loss": 0.7435,
255
+ "step": 35
256
+ },
257
+ {
258
+ "epoch": 0.2704225352112676,
259
+ "grad_norm": 1.171875,
260
+ "learning_rate": 4.869402985074627e-05,
261
+ "loss": 0.7014,
262
+ "step": 36
263
+ },
264
+ {
265
+ "epoch": 0.27793427230046946,
266
+ "grad_norm": 1.1875,
267
+ "learning_rate": 4.8656716417910445e-05,
268
+ "loss": 0.7946,
269
+ "step": 37
270
+ },
271
+ {
272
+ "epoch": 0.28544600938967135,
273
+ "grad_norm": 1.6328125,
274
+ "learning_rate": 4.861940298507463e-05,
275
+ "loss": 0.5345,
276
+ "step": 38
277
+ },
278
+ {
279
+ "epoch": 0.29295774647887324,
280
+ "grad_norm": 2.859375,
281
+ "learning_rate": 4.858208955223881e-05,
282
+ "loss": 0.8832,
283
+ "step": 39
284
+ },
285
+ {
286
+ "epoch": 0.3004694835680751,
287
+ "grad_norm": 2.09375,
288
+ "learning_rate": 4.8544776119402986e-05,
289
+ "loss": 0.8591,
290
+ "step": 40
291
+ },
292
+ {
293
+ "epoch": 0.307981220657277,
294
+ "grad_norm": 1.21875,
295
+ "learning_rate": 4.850746268656717e-05,
296
+ "loss": 0.6557,
297
+ "step": 41
298
+ },
299
+ {
300
+ "epoch": 0.3154929577464789,
301
+ "grad_norm": 0.984375,
302
+ "learning_rate": 4.8470149253731344e-05,
303
+ "loss": 0.715,
304
+ "step": 42
305
+ },
306
+ {
307
+ "epoch": 0.32300469483568073,
308
+ "grad_norm": 0.9453125,
309
+ "learning_rate": 4.843283582089552e-05,
310
+ "loss": 0.6608,
311
+ "step": 43
312
+ },
313
+ {
314
+ "epoch": 0.3305164319248826,
315
+ "grad_norm": 1.5703125,
316
+ "learning_rate": 4.83955223880597e-05,
317
+ "loss": 0.5904,
318
+ "step": 44
319
+ },
320
+ {
321
+ "epoch": 0.3380281690140845,
322
+ "grad_norm": 2.703125,
323
+ "learning_rate": 4.8358208955223885e-05,
324
+ "loss": 0.8645,
325
+ "step": 45
326
+ },
327
+ {
328
+ "epoch": 0.3455399061032864,
329
+ "grad_norm": 5.84375,
330
+ "learning_rate": 4.832089552238806e-05,
331
+ "loss": 0.9412,
332
+ "step": 46
333
+ },
334
+ {
335
+ "epoch": 0.3530516431924883,
336
+ "grad_norm": 1.2734375,
337
+ "learning_rate": 4.8283582089552244e-05,
338
+ "loss": 0.8767,
339
+ "step": 47
340
+ },
341
+ {
342
+ "epoch": 0.36056338028169016,
343
+ "grad_norm": 1.296875,
344
+ "learning_rate": 4.824626865671642e-05,
345
+ "loss": 0.8197,
346
+ "step": 48
347
+ },
348
+ {
349
+ "epoch": 0.36807511737089205,
350
+ "grad_norm": 1.0625,
351
+ "learning_rate": 4.8208955223880595e-05,
352
+ "loss": 0.6598,
353
+ "step": 49
354
+ },
355
+ {
356
+ "epoch": 0.3755868544600939,
357
+ "grad_norm": 4.0,
358
+ "learning_rate": 4.817164179104478e-05,
359
+ "loss": 1.1535,
360
+ "step": 50
361
+ },
362
+ {
363
+ "epoch": 0.38309859154929576,
364
+ "grad_norm": 1.09375,
365
+ "learning_rate": 4.813432835820896e-05,
366
+ "loss": 0.6635,
367
+ "step": 51
368
+ },
369
+ {
370
+ "epoch": 0.39061032863849765,
371
+ "grad_norm": 1.3125,
372
+ "learning_rate": 4.8097014925373136e-05,
373
+ "loss": 0.6979,
374
+ "step": 52
375
+ },
376
+ {
377
+ "epoch": 0.39812206572769954,
378
+ "grad_norm": 1.2734375,
379
+ "learning_rate": 4.805970149253732e-05,
380
+ "loss": 0.7167,
381
+ "step": 53
382
+ },
383
+ {
384
+ "epoch": 0.4056338028169014,
385
+ "grad_norm": 1.1640625,
386
+ "learning_rate": 4.8022388059701494e-05,
387
+ "loss": 0.8033,
388
+ "step": 54
389
+ },
390
+ {
391
+ "epoch": 0.4131455399061033,
392
+ "grad_norm": 1.421875,
393
+ "learning_rate": 4.798507462686567e-05,
394
+ "loss": 0.8736,
395
+ "step": 55
396
+ },
397
+ {
398
+ "epoch": 0.42065727699530514,
399
+ "grad_norm": 1.2890625,
400
+ "learning_rate": 4.794776119402985e-05,
401
+ "loss": 0.8361,
402
+ "step": 56
403
+ },
404
+ {
405
+ "epoch": 0.428169014084507,
406
+ "grad_norm": 1.1328125,
407
+ "learning_rate": 4.7910447761194035e-05,
408
+ "loss": 0.6386,
409
+ "step": 57
410
+ },
411
+ {
412
+ "epoch": 0.4356807511737089,
413
+ "grad_norm": 1.2421875,
414
+ "learning_rate": 4.787313432835821e-05,
415
+ "loss": 0.7318,
416
+ "step": 58
417
+ },
418
+ {
419
+ "epoch": 0.4431924882629108,
420
+ "grad_norm": 1.4375,
421
+ "learning_rate": 4.7835820895522394e-05,
422
+ "loss": 0.7141,
423
+ "step": 59
424
+ },
425
+ {
426
+ "epoch": 0.4507042253521127,
427
+ "grad_norm": 1.0859375,
428
+ "learning_rate": 4.779850746268657e-05,
429
+ "loss": 0.6612,
430
+ "step": 60
431
+ },
432
+ {
433
+ "epoch": 0.4582159624413146,
434
+ "grad_norm": 1.109375,
435
+ "learning_rate": 4.7761194029850745e-05,
436
+ "loss": 0.8155,
437
+ "step": 61
438
+ },
439
+ {
440
+ "epoch": 0.46572769953051646,
441
+ "grad_norm": 1.0546875,
442
+ "learning_rate": 4.772388059701493e-05,
443
+ "loss": 0.8123,
444
+ "step": 62
445
+ },
446
+ {
447
+ "epoch": 0.4732394366197183,
448
+ "grad_norm": 1.1640625,
449
+ "learning_rate": 4.768656716417911e-05,
450
+ "loss": 0.6706,
451
+ "step": 63
452
+ },
453
+ {
454
+ "epoch": 0.4807511737089202,
455
+ "grad_norm": 1.25,
456
+ "learning_rate": 4.7649253731343286e-05,
457
+ "loss": 0.8376,
458
+ "step": 64
459
+ },
460
+ {
461
+ "epoch": 0.48826291079812206,
462
+ "grad_norm": 0.98828125,
463
+ "learning_rate": 4.761194029850746e-05,
464
+ "loss": 0.7736,
465
+ "step": 65
466
+ },
467
+ {
468
+ "epoch": 0.49577464788732395,
469
+ "grad_norm": 1.296875,
470
+ "learning_rate": 4.7574626865671644e-05,
471
+ "loss": 0.777,
472
+ "step": 66
473
+ },
474
+ {
475
+ "epoch": 0.5032863849765258,
476
+ "grad_norm": 1.1328125,
477
+ "learning_rate": 4.753731343283582e-05,
478
+ "loss": 0.7517,
479
+ "step": 67
480
+ },
481
+ {
482
+ "epoch": 0.5107981220657277,
483
+ "grad_norm": 1.2578125,
484
+ "learning_rate": 4.75e-05,
485
+ "loss": 0.8422,
486
+ "step": 68
487
+ },
488
+ {
489
+ "epoch": 0.5183098591549296,
490
+ "grad_norm": 1.2109375,
491
+ "learning_rate": 4.7462686567164185e-05,
492
+ "loss": 0.6369,
493
+ "step": 69
494
+ },
495
+ {
496
+ "epoch": 0.5258215962441315,
497
+ "grad_norm": 1.0,
498
+ "learning_rate": 4.742537313432836e-05,
499
+ "loss": 0.6823,
500
+ "step": 70
501
+ },
502
+ {
503
+ "epoch": 0.5333333333333333,
504
+ "grad_norm": 1.1953125,
505
+ "learning_rate": 4.738805970149254e-05,
506
+ "loss": 0.6799,
507
+ "step": 71
508
+ },
509
+ {
510
+ "epoch": 0.5408450704225352,
511
+ "grad_norm": 1.1640625,
512
+ "learning_rate": 4.735074626865672e-05,
513
+ "loss": 0.6477,
514
+ "step": 72
515
+ },
516
+ {
517
+ "epoch": 0.5483568075117371,
518
+ "grad_norm": 1.125,
519
+ "learning_rate": 4.7313432835820895e-05,
520
+ "loss": 0.7116,
521
+ "step": 73
522
+ },
523
+ {
524
+ "epoch": 0.5558685446009389,
525
+ "grad_norm": 1.0390625,
526
+ "learning_rate": 4.727611940298508e-05,
527
+ "loss": 0.6501,
528
+ "step": 74
529
+ },
530
+ {
531
+ "epoch": 0.5633802816901409,
532
+ "grad_norm": 1.203125,
533
+ "learning_rate": 4.723880597014926e-05,
534
+ "loss": 0.8061,
535
+ "step": 75
536
+ },
537
+ {
538
+ "epoch": 0.5708920187793427,
539
+ "grad_norm": 1.203125,
540
+ "learning_rate": 4.7201492537313436e-05,
541
+ "loss": 0.7363,
542
+ "step": 76
543
+ },
544
+ {
545
+ "epoch": 0.5784037558685446,
546
+ "grad_norm": 1.15625,
547
+ "learning_rate": 4.716417910447761e-05,
548
+ "loss": 0.7949,
549
+ "step": 77
550
+ },
551
+ {
552
+ "epoch": 0.5859154929577465,
553
+ "grad_norm": 1.25,
554
+ "learning_rate": 4.7126865671641794e-05,
555
+ "loss": 0.7815,
556
+ "step": 78
557
+ },
558
+ {
559
+ "epoch": 0.5934272300469483,
560
+ "grad_norm": 1.40625,
561
+ "learning_rate": 4.708955223880597e-05,
562
+ "loss": 0.7815,
563
+ "step": 79
564
+ },
565
+ {
566
+ "epoch": 0.6009389671361502,
567
+ "grad_norm": 1.1796875,
568
+ "learning_rate": 4.705223880597015e-05,
569
+ "loss": 0.7299,
570
+ "step": 80
571
+ },
572
+ {
573
+ "epoch": 0.6084507042253521,
574
+ "grad_norm": 1.25,
575
+ "learning_rate": 4.7014925373134335e-05,
576
+ "loss": 0.5975,
577
+ "step": 81
578
+ },
579
+ {
580
+ "epoch": 0.615962441314554,
581
+ "grad_norm": 1.8984375,
582
+ "learning_rate": 4.697761194029851e-05,
583
+ "loss": 0.8056,
584
+ "step": 82
585
+ },
586
+ {
587
+ "epoch": 0.6234741784037559,
588
+ "grad_norm": 1.3671875,
589
+ "learning_rate": 4.6940298507462687e-05,
590
+ "loss": 0.8034,
591
+ "step": 83
592
+ },
593
+ {
594
+ "epoch": 0.6309859154929578,
595
+ "grad_norm": 1.078125,
596
+ "learning_rate": 4.690298507462687e-05,
597
+ "loss": 0.6472,
598
+ "step": 84
599
+ },
600
+ {
601
+ "epoch": 0.6384976525821596,
602
+ "grad_norm": 1.328125,
603
+ "learning_rate": 4.6865671641791045e-05,
604
+ "loss": 0.8561,
605
+ "step": 85
606
+ },
607
+ {
608
+ "epoch": 0.6460093896713615,
609
+ "grad_norm": 2.90625,
610
+ "learning_rate": 4.682835820895523e-05,
611
+ "loss": 0.8774,
612
+ "step": 86
613
+ },
614
+ {
615
+ "epoch": 0.6535211267605634,
616
+ "grad_norm": 1.1484375,
617
+ "learning_rate": 4.67910447761194e-05,
618
+ "loss": 0.6865,
619
+ "step": 87
620
+ },
621
+ {
622
+ "epoch": 0.6610328638497652,
623
+ "grad_norm": 1.53125,
624
+ "learning_rate": 4.6753731343283586e-05,
625
+ "loss": 0.798,
626
+ "step": 88
627
+ },
628
+ {
629
+ "epoch": 0.6685446009389672,
630
+ "grad_norm": 1.1875,
631
+ "learning_rate": 4.671641791044776e-05,
632
+ "loss": 0.7587,
633
+ "step": 89
634
+ },
635
+ {
636
+ "epoch": 0.676056338028169,
637
+ "grad_norm": 1.25,
638
+ "learning_rate": 4.667910447761194e-05,
639
+ "loss": 0.6989,
640
+ "step": 90
641
+ },
642
+ {
643
+ "epoch": 0.6835680751173709,
644
+ "grad_norm": 1.0546875,
645
+ "learning_rate": 4.664179104477612e-05,
646
+ "loss": 0.639,
647
+ "step": 91
648
+ },
649
+ {
650
+ "epoch": 0.6910798122065728,
651
+ "grad_norm": 1.453125,
652
+ "learning_rate": 4.66044776119403e-05,
653
+ "loss": 0.6844,
654
+ "step": 92
655
+ },
656
+ {
657
+ "epoch": 0.6985915492957746,
658
+ "grad_norm": 1.03125,
659
+ "learning_rate": 4.656716417910448e-05,
660
+ "loss": 0.7513,
661
+ "step": 93
662
+ },
663
+ {
664
+ "epoch": 0.7061032863849765,
665
+ "grad_norm": 1.015625,
666
+ "learning_rate": 4.652985074626866e-05,
667
+ "loss": 0.6264,
668
+ "step": 94
669
+ },
670
+ {
671
+ "epoch": 0.7136150234741784,
672
+ "grad_norm": 1.21875,
673
+ "learning_rate": 4.6492537313432837e-05,
674
+ "loss": 0.7228,
675
+ "step": 95
676
+ },
677
+ {
678
+ "epoch": 0.7211267605633803,
679
+ "grad_norm": 1.0078125,
680
+ "learning_rate": 4.645522388059701e-05,
681
+ "loss": 0.5355,
682
+ "step": 96
683
+ },
684
+ {
685
+ "epoch": 0.7286384976525822,
686
+ "grad_norm": 1.359375,
687
+ "learning_rate": 4.6417910447761195e-05,
688
+ "loss": 0.7717,
689
+ "step": 97
690
+ },
691
+ {
692
+ "epoch": 0.7361502347417841,
693
+ "grad_norm": 1.1328125,
694
+ "learning_rate": 4.638059701492538e-05,
695
+ "loss": 0.7365,
696
+ "step": 98
697
+ },
698
+ {
699
+ "epoch": 0.7436619718309859,
700
+ "grad_norm": 3.375,
701
+ "learning_rate": 4.634328358208955e-05,
702
+ "loss": 0.8382,
703
+ "step": 99
704
+ },
705
+ {
706
+ "epoch": 0.7511737089201878,
707
+ "grad_norm": 1.421875,
708
+ "learning_rate": 4.6305970149253736e-05,
709
+ "loss": 0.7674,
710
+ "step": 100
711
+ },
712
+ {
713
+ "epoch": 0.7586854460093897,
714
+ "grad_norm": 1.7265625,
715
+ "learning_rate": 4.626865671641791e-05,
716
+ "loss": 0.554,
717
+ "step": 101
718
+ },
719
+ {
720
+ "epoch": 0.7661971830985915,
721
+ "grad_norm": 1.25,
722
+ "learning_rate": 4.623134328358209e-05,
723
+ "loss": 0.7921,
724
+ "step": 102
725
+ },
726
+ {
727
+ "epoch": 0.7737089201877935,
728
+ "grad_norm": 1.2734375,
729
+ "learning_rate": 4.6194029850746277e-05,
730
+ "loss": 0.6282,
731
+ "step": 103
732
+ },
733
+ {
734
+ "epoch": 0.7812206572769953,
735
+ "grad_norm": 1.171875,
736
+ "learning_rate": 4.615671641791045e-05,
737
+ "loss": 0.7992,
738
+ "step": 104
739
+ },
740
+ {
741
+ "epoch": 0.7887323943661971,
742
+ "grad_norm": 1.0859375,
743
+ "learning_rate": 4.611940298507463e-05,
744
+ "loss": 0.7313,
745
+ "step": 105
746
+ },
747
+ {
748
+ "epoch": 0.7962441314553991,
749
+ "grad_norm": 0.9921875,
750
+ "learning_rate": 4.608208955223881e-05,
751
+ "loss": 0.7154,
752
+ "step": 106
753
+ },
754
+ {
755
+ "epoch": 0.8037558685446009,
756
+ "grad_norm": 1.0859375,
757
+ "learning_rate": 4.6044776119402986e-05,
758
+ "loss": 0.6904,
759
+ "step": 107
760
+ },
761
+ {
762
+ "epoch": 0.8112676056338028,
763
+ "grad_norm": 1.125,
764
+ "learning_rate": 4.600746268656716e-05,
765
+ "loss": 0.7075,
766
+ "step": 108
767
+ },
768
+ {
769
+ "epoch": 0.8187793427230047,
770
+ "grad_norm": 2.609375,
771
+ "learning_rate": 4.5970149253731345e-05,
772
+ "loss": 0.7362,
773
+ "step": 109
774
+ },
775
+ {
776
+ "epoch": 0.8262910798122066,
777
+ "grad_norm": 1.125,
778
+ "learning_rate": 4.593283582089553e-05,
779
+ "loss": 0.6568,
780
+ "step": 110
781
+ },
782
+ {
783
+ "epoch": 0.8338028169014085,
784
+ "grad_norm": 1.3984375,
785
+ "learning_rate": 4.58955223880597e-05,
786
+ "loss": 0.7508,
787
+ "step": 111
788
+ },
789
+ {
790
+ "epoch": 0.8413145539906103,
791
+ "grad_norm": 1.078125,
792
+ "learning_rate": 4.585820895522388e-05,
793
+ "loss": 0.8402,
794
+ "step": 112
795
+ },
796
+ {
797
+ "epoch": 0.8488262910798122,
798
+ "grad_norm": 1.2734375,
799
+ "learning_rate": 4.582089552238806e-05,
800
+ "loss": 0.6946,
801
+ "step": 113
802
+ },
803
+ {
804
+ "epoch": 0.856338028169014,
805
+ "grad_norm": 1.03125,
806
+ "learning_rate": 4.578358208955224e-05,
807
+ "loss": 0.6877,
808
+ "step": 114
809
+ },
810
+ {
811
+ "epoch": 0.863849765258216,
812
+ "grad_norm": 1.2890625,
813
+ "learning_rate": 4.574626865671642e-05,
814
+ "loss": 0.6703,
815
+ "step": 115
816
+ },
817
+ {
818
+ "epoch": 0.8713615023474178,
819
+ "grad_norm": 1.2578125,
820
+ "learning_rate": 4.57089552238806e-05,
821
+ "loss": 0.6673,
822
+ "step": 116
823
+ },
824
+ {
825
+ "epoch": 0.8788732394366198,
826
+ "grad_norm": 1.078125,
827
+ "learning_rate": 4.567164179104478e-05,
828
+ "loss": 0.8,
829
+ "step": 117
830
+ },
831
+ {
832
+ "epoch": 0.8863849765258216,
833
+ "grad_norm": 1.7734375,
834
+ "learning_rate": 4.5634328358208954e-05,
835
+ "loss": 0.7146,
836
+ "step": 118
837
+ },
838
+ {
839
+ "epoch": 0.8938967136150234,
840
+ "grad_norm": 1.1953125,
841
+ "learning_rate": 4.5597014925373136e-05,
842
+ "loss": 0.7342,
843
+ "step": 119
844
+ },
845
+ {
846
+ "epoch": 0.9014084507042254,
847
+ "grad_norm": 1.390625,
848
+ "learning_rate": 4.555970149253732e-05,
849
+ "loss": 0.865,
850
+ "step": 120
851
+ },
852
+ {
853
+ "epoch": 0.9089201877934272,
854
+ "grad_norm": 1.171875,
855
+ "learning_rate": 4.5522388059701495e-05,
856
+ "loss": 0.8743,
857
+ "step": 121
858
+ },
859
+ {
860
+ "epoch": 0.9164319248826291,
861
+ "grad_norm": 1.0625,
862
+ "learning_rate": 4.548507462686568e-05,
863
+ "loss": 0.8124,
864
+ "step": 122
865
+ },
866
+ {
867
+ "epoch": 0.923943661971831,
868
+ "grad_norm": 1.0703125,
869
+ "learning_rate": 4.544776119402985e-05,
870
+ "loss": 0.739,
871
+ "step": 123
872
+ },
873
+ {
874
+ "epoch": 0.9314553990610329,
875
+ "grad_norm": 1.109375,
876
+ "learning_rate": 4.541044776119403e-05,
877
+ "loss": 0.9343,
878
+ "step": 124
879
+ },
880
+ {
881
+ "epoch": 0.9389671361502347,
882
+ "grad_norm": 1.21875,
883
+ "learning_rate": 4.537313432835821e-05,
884
+ "loss": 0.8239,
885
+ "step": 125
886
+ },
887
+ {
888
+ "epoch": 0.9464788732394366,
889
+ "grad_norm": 1.125,
890
+ "learning_rate": 4.5335820895522394e-05,
891
+ "loss": 0.5928,
892
+ "step": 126
893
+ },
894
+ {
895
+ "epoch": 0.9539906103286385,
896
+ "grad_norm": 1.140625,
897
+ "learning_rate": 4.529850746268657e-05,
898
+ "loss": 0.6705,
899
+ "step": 127
900
+ },
901
+ {
902
+ "epoch": 0.9615023474178404,
903
+ "grad_norm": 1.5390625,
904
+ "learning_rate": 4.526119402985075e-05,
905
+ "loss": 0.7937,
906
+ "step": 128
907
+ },
908
+ {
909
+ "epoch": 0.9690140845070423,
910
+ "grad_norm": 1.0703125,
911
+ "learning_rate": 4.522388059701493e-05,
912
+ "loss": 0.7819,
913
+ "step": 129
914
+ },
915
+ {
916
+ "epoch": 0.9765258215962441,
917
+ "grad_norm": 1.2265625,
918
+ "learning_rate": 4.5186567164179104e-05,
919
+ "loss": 0.7217,
920
+ "step": 130
921
+ },
922
+ {
923
+ "epoch": 0.984037558685446,
924
+ "grad_norm": 0.97265625,
925
+ "learning_rate": 4.5149253731343286e-05,
926
+ "loss": 0.5954,
927
+ "step": 131
928
+ },
929
+ {
930
+ "epoch": 0.9915492957746479,
931
+ "grad_norm": 1.171875,
932
+ "learning_rate": 4.511194029850747e-05,
933
+ "loss": 0.6472,
934
+ "step": 132
935
+ },
936
+ {
937
+ "epoch": 0.9990610328638497,
938
+ "grad_norm": 1.078125,
939
+ "learning_rate": 4.5074626865671645e-05,
940
+ "loss": 0.6715,
941
+ "step": 133
942
+ },
943
+ {
944
+ "epoch": 1.0,
945
+ "grad_norm": 2.765625,
946
+ "learning_rate": 4.503731343283582e-05,
947
+ "loss": 0.7686,
948
+ "step": 134
949
+ },
950
+ {
951
+ "epoch": 1.007511737089202,
952
+ "grad_norm": 1.625,
953
+ "learning_rate": 4.5e-05,
954
+ "loss": 0.6494,
955
+ "step": 135
956
+ },
957
+ {
958
+ "epoch": 1.0150234741784037,
959
+ "grad_norm": 1.46875,
960
+ "learning_rate": 4.496268656716418e-05,
961
+ "loss": 0.5652,
962
+ "step": 136
963
+ },
964
+ {
965
+ "epoch": 1.0225352112676056,
966
+ "grad_norm": 0.9296875,
967
+ "learning_rate": 4.492537313432836e-05,
968
+ "loss": 0.5451,
969
+ "step": 137
970
+ },
971
+ {
972
+ "epoch": 1.0300469483568075,
973
+ "grad_norm": 1.1171875,
974
+ "learning_rate": 4.4888059701492544e-05,
975
+ "loss": 0.7535,
976
+ "step": 138
977
+ },
978
+ {
979
+ "epoch": 1.0375586854460095,
980
+ "grad_norm": 1.328125,
981
+ "learning_rate": 4.485074626865672e-05,
982
+ "loss": 0.6217,
983
+ "step": 139
984
+ },
985
+ {
986
+ "epoch": 1.0450704225352112,
987
+ "grad_norm": 1.1953125,
988
+ "learning_rate": 4.4813432835820895e-05,
989
+ "loss": 0.6696,
990
+ "step": 140
991
+ },
992
+ {
993
+ "epoch": 1.0525821596244131,
994
+ "grad_norm": 1.59375,
995
+ "learning_rate": 4.477611940298508e-05,
996
+ "loss": 0.5071,
997
+ "step": 141
998
+ },
999
+ {
1000
+ "epoch": 1.060093896713615,
1001
+ "grad_norm": 1.4921875,
1002
+ "learning_rate": 4.4738805970149254e-05,
1003
+ "loss": 0.5824,
1004
+ "step": 142
1005
+ },
1006
+ {
1007
+ "epoch": 1.0676056338028168,
1008
+ "grad_norm": 1.1796875,
1009
+ "learning_rate": 4.4701492537313436e-05,
1010
+ "loss": 0.4513,
1011
+ "step": 143
1012
+ },
1013
+ {
1014
+ "epoch": 1.0751173708920188,
1015
+ "grad_norm": 1.3671875,
1016
+ "learning_rate": 4.466417910447762e-05,
1017
+ "loss": 0.6003,
1018
+ "step": 144
1019
+ },
1020
+ {
1021
+ "epoch": 1.0826291079812207,
1022
+ "grad_norm": 1.3515625,
1023
+ "learning_rate": 4.4626865671641794e-05,
1024
+ "loss": 0.6236,
1025
+ "step": 145
1026
+ },
1027
+ {
1028
+ "epoch": 1.0901408450704226,
1029
+ "grad_norm": 1.0859375,
1030
+ "learning_rate": 4.458955223880597e-05,
1031
+ "loss": 0.5397,
1032
+ "step": 146
1033
+ },
1034
+ {
1035
+ "epoch": 1.0976525821596244,
1036
+ "grad_norm": 2.0625,
1037
+ "learning_rate": 4.455223880597015e-05,
1038
+ "loss": 0.8009,
1039
+ "step": 147
1040
+ },
1041
+ {
1042
+ "epoch": 1.1051643192488263,
1043
+ "grad_norm": 1.3125,
1044
+ "learning_rate": 4.451492537313433e-05,
1045
+ "loss": 0.7748,
1046
+ "step": 148
1047
+ },
1048
+ {
1049
+ "epoch": 1.1126760563380282,
1050
+ "grad_norm": 3.421875,
1051
+ "learning_rate": 4.447761194029851e-05,
1052
+ "loss": 0.7256,
1053
+ "step": 149
1054
+ },
1055
+ {
1056
+ "epoch": 1.12018779342723,
1057
+ "grad_norm": 1.296875,
1058
+ "learning_rate": 4.4440298507462694e-05,
1059
+ "loss": 0.6101,
1060
+ "step": 150
1061
+ },
1062
+ {
1063
+ "epoch": 1.127699530516432,
1064
+ "grad_norm": 1.109375,
1065
+ "learning_rate": 4.440298507462687e-05,
1066
+ "loss": 0.4897,
1067
+ "step": 151
1068
+ },
1069
+ {
1070
+ "epoch": 1.1352112676056338,
1071
+ "grad_norm": 1.0703125,
1072
+ "learning_rate": 4.4365671641791045e-05,
1073
+ "loss": 0.7608,
1074
+ "step": 152
1075
+ },
1076
+ {
1077
+ "epoch": 1.1427230046948358,
1078
+ "grad_norm": 1.1171875,
1079
+ "learning_rate": 4.432835820895523e-05,
1080
+ "loss": 0.6389,
1081
+ "step": 153
1082
+ },
1083
+ {
1084
+ "epoch": 1.1502347417840375,
1085
+ "grad_norm": 0.98828125,
1086
+ "learning_rate": 4.4291044776119403e-05,
1087
+ "loss": 0.4912,
1088
+ "step": 154
1089
+ },
1090
+ {
1091
+ "epoch": 1.1577464788732394,
1092
+ "grad_norm": 1.1171875,
1093
+ "learning_rate": 4.4253731343283586e-05,
1094
+ "loss": 0.6865,
1095
+ "step": 155
1096
+ },
1097
+ {
1098
+ "epoch": 1.1652582159624414,
1099
+ "grad_norm": 1.0078125,
1100
+ "learning_rate": 4.421641791044777e-05,
1101
+ "loss": 0.4687,
1102
+ "step": 156
1103
+ },
1104
+ {
1105
+ "epoch": 1.172769953051643,
1106
+ "grad_norm": 1.6484375,
1107
+ "learning_rate": 4.4179104477611944e-05,
1108
+ "loss": 0.6835,
1109
+ "step": 157
1110
+ },
1111
+ {
1112
+ "epoch": 1.180281690140845,
1113
+ "grad_norm": 1.015625,
1114
+ "learning_rate": 4.414179104477612e-05,
1115
+ "loss": 0.4748,
1116
+ "step": 158
1117
+ },
1118
+ {
1119
+ "epoch": 1.187793427230047,
1120
+ "grad_norm": 1.1015625,
1121
+ "learning_rate": 4.4104477611940296e-05,
1122
+ "loss": 0.556,
1123
+ "step": 159
1124
+ },
1125
+ {
1126
+ "epoch": 1.1953051643192487,
1127
+ "grad_norm": 1.0390625,
1128
+ "learning_rate": 4.406716417910448e-05,
1129
+ "loss": 0.4781,
1130
+ "step": 160
1131
+ },
1132
+ {
1133
+ "epoch": 1.2028169014084507,
1134
+ "grad_norm": 0.87890625,
1135
+ "learning_rate": 4.402985074626866e-05,
1136
+ "loss": 0.4693,
1137
+ "step": 161
1138
+ },
1139
+ {
1140
+ "epoch": 1.2103286384976526,
1141
+ "grad_norm": 1.1171875,
1142
+ "learning_rate": 4.399253731343284e-05,
1143
+ "loss": 0.5578,
1144
+ "step": 162
1145
+ },
1146
+ {
1147
+ "epoch": 1.2178403755868545,
1148
+ "grad_norm": 1.2890625,
1149
+ "learning_rate": 4.395522388059702e-05,
1150
+ "loss": 0.6094,
1151
+ "step": 163
1152
+ },
1153
+ {
1154
+ "epoch": 1.2253521126760563,
1155
+ "grad_norm": 1.671875,
1156
+ "learning_rate": 4.3917910447761195e-05,
1157
+ "loss": 0.6457,
1158
+ "step": 164
1159
+ },
1160
+ {
1161
+ "epoch": 1.2328638497652582,
1162
+ "grad_norm": 0.96875,
1163
+ "learning_rate": 4.388059701492537e-05,
1164
+ "loss": 0.5279,
1165
+ "step": 165
1166
+ },
1167
+ {
1168
+ "epoch": 1.2403755868544601,
1169
+ "grad_norm": 1.0234375,
1170
+ "learning_rate": 4.384328358208955e-05,
1171
+ "loss": 0.5519,
1172
+ "step": 166
1173
+ },
1174
+ {
1175
+ "epoch": 1.247887323943662,
1176
+ "grad_norm": 0.97265625,
1177
+ "learning_rate": 4.3805970149253736e-05,
1178
+ "loss": 0.4328,
1179
+ "step": 167
1180
+ },
1181
+ {
1182
+ "epoch": 1.2553990610328638,
1183
+ "grad_norm": 1.6484375,
1184
+ "learning_rate": 4.376865671641791e-05,
1185
+ "loss": 0.6765,
1186
+ "step": 168
1187
+ },
1188
+ {
1189
+ "epoch": 1.2629107981220657,
1190
+ "grad_norm": 0.953125,
1191
+ "learning_rate": 4.3731343283582094e-05,
1192
+ "loss": 0.4776,
1193
+ "step": 169
1194
+ },
1195
+ {
1196
+ "epoch": 1.2704225352112677,
1197
+ "grad_norm": 1.125,
1198
+ "learning_rate": 4.369402985074627e-05,
1199
+ "loss": 0.4864,
1200
+ "step": 170
1201
+ },
1202
+ {
1203
+ "epoch": 1.2779342723004694,
1204
+ "grad_norm": 1.2421875,
1205
+ "learning_rate": 4.3656716417910446e-05,
1206
+ "loss": 0.5458,
1207
+ "step": 171
1208
+ },
1209
+ {
1210
+ "epoch": 1.2854460093896714,
1211
+ "grad_norm": 1.046875,
1212
+ "learning_rate": 4.361940298507463e-05,
1213
+ "loss": 0.5282,
1214
+ "step": 172
1215
+ },
1216
+ {
1217
+ "epoch": 1.2929577464788733,
1218
+ "grad_norm": 1.0703125,
1219
+ "learning_rate": 4.358208955223881e-05,
1220
+ "loss": 0.5868,
1221
+ "step": 173
1222
+ },
1223
+ {
1224
+ "epoch": 1.300469483568075,
1225
+ "grad_norm": 0.94921875,
1226
+ "learning_rate": 4.354477611940299e-05,
1227
+ "loss": 0.5913,
1228
+ "step": 174
1229
+ },
1230
+ {
1231
+ "epoch": 1.307981220657277,
1232
+ "grad_norm": 1.1953125,
1233
+ "learning_rate": 4.350746268656717e-05,
1234
+ "loss": 0.6012,
1235
+ "step": 175
1236
+ },
1237
+ {
1238
+ "epoch": 1.315492957746479,
1239
+ "grad_norm": 1.578125,
1240
+ "learning_rate": 4.3470149253731345e-05,
1241
+ "loss": 0.7648,
1242
+ "step": 176
1243
+ },
1244
+ {
1245
+ "epoch": 1.3230046948356806,
1246
+ "grad_norm": 1.09375,
1247
+ "learning_rate": 4.343283582089552e-05,
1248
+ "loss": 0.5035,
1249
+ "step": 177
1250
+ },
1251
+ {
1252
+ "epoch": 1.3305164319248826,
1253
+ "grad_norm": 0.98046875,
1254
+ "learning_rate": 4.33955223880597e-05,
1255
+ "loss": 0.6096,
1256
+ "step": 178
1257
+ },
1258
+ {
1259
+ "epoch": 1.3380281690140845,
1260
+ "grad_norm": 1.1875,
1261
+ "learning_rate": 4.3358208955223886e-05,
1262
+ "loss": 0.6028,
1263
+ "step": 179
1264
+ },
1265
+ {
1266
+ "epoch": 1.3455399061032864,
1267
+ "grad_norm": 1.046875,
1268
+ "learning_rate": 4.332089552238806e-05,
1269
+ "loss": 0.4301,
1270
+ "step": 180
1271
+ },
1272
+ {
1273
+ "epoch": 1.3530516431924884,
1274
+ "grad_norm": 0.875,
1275
+ "learning_rate": 4.328358208955224e-05,
1276
+ "loss": 0.4578,
1277
+ "step": 181
1278
+ },
1279
+ {
1280
+ "epoch": 1.36056338028169,
1281
+ "grad_norm": 1.2734375,
1282
+ "learning_rate": 4.324626865671642e-05,
1283
+ "loss": 0.6293,
1284
+ "step": 182
1285
+ },
1286
+ {
1287
+ "epoch": 1.368075117370892,
1288
+ "grad_norm": 1.09375,
1289
+ "learning_rate": 4.3208955223880596e-05,
1290
+ "loss": 0.6577,
1291
+ "step": 183
1292
+ },
1293
+ {
1294
+ "epoch": 1.375586854460094,
1295
+ "grad_norm": 1.0,
1296
+ "learning_rate": 4.317164179104478e-05,
1297
+ "loss": 0.5545,
1298
+ "step": 184
1299
+ },
1300
+ {
1301
+ "epoch": 1.3830985915492957,
1302
+ "grad_norm": 1.28125,
1303
+ "learning_rate": 4.313432835820896e-05,
1304
+ "loss": 0.6445,
1305
+ "step": 185
1306
+ },
1307
+ {
1308
+ "epoch": 1.3906103286384977,
1309
+ "grad_norm": 0.9609375,
1310
+ "learning_rate": 4.3097014925373137e-05,
1311
+ "loss": 0.3414,
1312
+ "step": 186
1313
+ },
1314
+ {
1315
+ "epoch": 1.3981220657276996,
1316
+ "grad_norm": 1.1328125,
1317
+ "learning_rate": 4.305970149253731e-05,
1318
+ "loss": 0.6948,
1319
+ "step": 187
1320
+ },
1321
+ {
1322
+ "epoch": 1.4056338028169013,
1323
+ "grad_norm": 1.15625,
1324
+ "learning_rate": 4.3022388059701495e-05,
1325
+ "loss": 0.4784,
1326
+ "step": 188
1327
+ },
1328
+ {
1329
+ "epoch": 1.4131455399061033,
1330
+ "grad_norm": 1.140625,
1331
+ "learning_rate": 4.298507462686567e-05,
1332
+ "loss": 0.665,
1333
+ "step": 189
1334
+ },
1335
+ {
1336
+ "epoch": 1.4206572769953052,
1337
+ "grad_norm": 1.0703125,
1338
+ "learning_rate": 4.294776119402985e-05,
1339
+ "loss": 0.5881,
1340
+ "step": 190
1341
+ },
1342
+ {
1343
+ "epoch": 1.428169014084507,
1344
+ "grad_norm": 0.98046875,
1345
+ "learning_rate": 4.2910447761194036e-05,
1346
+ "loss": 0.4298,
1347
+ "step": 191
1348
+ },
1349
+ {
1350
+ "epoch": 1.4356807511737089,
1351
+ "grad_norm": 1.2578125,
1352
+ "learning_rate": 4.287313432835821e-05,
1353
+ "loss": 0.6543,
1354
+ "step": 192
1355
+ },
1356
+ {
1357
+ "epoch": 1.4431924882629108,
1358
+ "grad_norm": 1.09375,
1359
+ "learning_rate": 4.283582089552239e-05,
1360
+ "loss": 0.4931,
1361
+ "step": 193
1362
+ },
1363
+ {
1364
+ "epoch": 1.4507042253521127,
1365
+ "grad_norm": 1.2265625,
1366
+ "learning_rate": 4.279850746268657e-05,
1367
+ "loss": 0.5279,
1368
+ "step": 194
1369
+ },
1370
+ {
1371
+ "epoch": 1.4582159624413147,
1372
+ "grad_norm": 1.015625,
1373
+ "learning_rate": 4.2761194029850746e-05,
1374
+ "loss": 0.5347,
1375
+ "step": 195
1376
+ },
1377
+ {
1378
+ "epoch": 1.4657276995305164,
1379
+ "grad_norm": 1.15625,
1380
+ "learning_rate": 4.272388059701493e-05,
1381
+ "loss": 0.48,
1382
+ "step": 196
1383
+ },
1384
+ {
1385
+ "epoch": 1.4732394366197183,
1386
+ "grad_norm": 1.25,
1387
+ "learning_rate": 4.268656716417911e-05,
1388
+ "loss": 0.6988,
1389
+ "step": 197
1390
+ },
1391
+ {
1392
+ "epoch": 1.4807511737089203,
1393
+ "grad_norm": 1.421875,
1394
+ "learning_rate": 4.2649253731343286e-05,
1395
+ "loss": 0.5859,
1396
+ "step": 198
1397
+ },
1398
+ {
1399
+ "epoch": 1.488262910798122,
1400
+ "grad_norm": 1.1796875,
1401
+ "learning_rate": 4.261194029850746e-05,
1402
+ "loss": 0.5283,
1403
+ "step": 199
1404
+ },
1405
+ {
1406
+ "epoch": 1.495774647887324,
1407
+ "grad_norm": 1.0703125,
1408
+ "learning_rate": 4.2574626865671645e-05,
1409
+ "loss": 0.4631,
1410
+ "step": 200
1411
+ },
1412
+ {
1413
+ "epoch": 1.503286384976526,
1414
+ "grad_norm": 0.9921875,
1415
+ "learning_rate": 4.253731343283582e-05,
1416
+ "loss": 0.5231,
1417
+ "step": 201
1418
+ },
1419
+ {
1420
+ "epoch": 1.5107981220657276,
1421
+ "grad_norm": 1.0625,
1422
+ "learning_rate": 4.25e-05,
1423
+ "loss": 0.5583,
1424
+ "step": 202
1425
+ },
1426
+ {
1427
+ "epoch": 1.5183098591549296,
1428
+ "grad_norm": 0.9453125,
1429
+ "learning_rate": 4.2462686567164186e-05,
1430
+ "loss": 0.4876,
1431
+ "step": 203
1432
+ },
1433
+ {
1434
+ "epoch": 1.5258215962441315,
1435
+ "grad_norm": 1.2265625,
1436
+ "learning_rate": 4.242537313432836e-05,
1437
+ "loss": 0.4504,
1438
+ "step": 204
1439
+ },
1440
+ {
1441
+ "epoch": 1.5333333333333332,
1442
+ "grad_norm": 0.93359375,
1443
+ "learning_rate": 4.238805970149254e-05,
1444
+ "loss": 0.4757,
1445
+ "step": 205
1446
+ },
1447
+ {
1448
+ "epoch": 1.5408450704225352,
1449
+ "grad_norm": 1.0703125,
1450
+ "learning_rate": 4.235074626865671e-05,
1451
+ "loss": 0.5933,
1452
+ "step": 206
1453
+ },
1454
+ {
1455
+ "epoch": 1.548356807511737,
1456
+ "grad_norm": 1.0078125,
1457
+ "learning_rate": 4.2313432835820895e-05,
1458
+ "loss": 0.5415,
1459
+ "step": 207
1460
+ },
1461
+ {
1462
+ "epoch": 1.5558685446009388,
1463
+ "grad_norm": 1.1953125,
1464
+ "learning_rate": 4.227611940298508e-05,
1465
+ "loss": 0.7041,
1466
+ "step": 208
1467
+ },
1468
+ {
1469
+ "epoch": 1.563380281690141,
1470
+ "grad_norm": 1.109375,
1471
+ "learning_rate": 4.2238805970149254e-05,
1472
+ "loss": 0.5753,
1473
+ "step": 209
1474
+ },
1475
+ {
1476
+ "epoch": 1.5708920187793427,
1477
+ "grad_norm": 1.1640625,
1478
+ "learning_rate": 4.2201492537313436e-05,
1479
+ "loss": 0.552,
1480
+ "step": 210
1481
+ },
1482
+ {
1483
+ "epoch": 1.5784037558685446,
1484
+ "grad_norm": 1.0,
1485
+ "learning_rate": 4.216417910447761e-05,
1486
+ "loss": 0.5549,
1487
+ "step": 211
1488
+ },
1489
+ {
1490
+ "epoch": 1.5859154929577466,
1491
+ "grad_norm": 1.1328125,
1492
+ "learning_rate": 4.212686567164179e-05,
1493
+ "loss": 0.6679,
1494
+ "step": 212
1495
+ },
1496
+ {
1497
+ "epoch": 1.5934272300469483,
1498
+ "grad_norm": 2.015625,
1499
+ "learning_rate": 4.208955223880597e-05,
1500
+ "loss": 0.796,
1501
+ "step": 213
1502
+ },
1503
+ {
1504
+ "epoch": 1.6009389671361502,
1505
+ "grad_norm": 2.953125,
1506
+ "learning_rate": 4.205223880597015e-05,
1507
+ "loss": 0.6427,
1508
+ "step": 214
1509
+ },
1510
+ {
1511
+ "epoch": 1.6084507042253522,
1512
+ "grad_norm": 0.90625,
1513
+ "learning_rate": 4.201492537313433e-05,
1514
+ "loss": 0.466,
1515
+ "step": 215
1516
+ },
1517
+ {
1518
+ "epoch": 1.615962441314554,
1519
+ "grad_norm": 0.94140625,
1520
+ "learning_rate": 4.197761194029851e-05,
1521
+ "loss": 0.4242,
1522
+ "step": 216
1523
+ },
1524
+ {
1525
+ "epoch": 1.6234741784037559,
1526
+ "grad_norm": 3.0,
1527
+ "learning_rate": 4.194029850746269e-05,
1528
+ "loss": 0.6968,
1529
+ "step": 217
1530
+ },
1531
+ {
1532
+ "epoch": 1.6309859154929578,
1533
+ "grad_norm": 0.9765625,
1534
+ "learning_rate": 4.190298507462686e-05,
1535
+ "loss": 0.5988,
1536
+ "step": 218
1537
+ },
1538
+ {
1539
+ "epoch": 1.6384976525821595,
1540
+ "grad_norm": 1.203125,
1541
+ "learning_rate": 4.186567164179105e-05,
1542
+ "loss": 0.4979,
1543
+ "step": 219
1544
+ },
1545
+ {
1546
+ "epoch": 1.6460093896713615,
1547
+ "grad_norm": 0.83203125,
1548
+ "learning_rate": 4.182835820895523e-05,
1549
+ "loss": 0.3527,
1550
+ "step": 220
1551
+ },
1552
+ {
1553
+ "epoch": 1.6535211267605634,
1554
+ "grad_norm": 1.09375,
1555
+ "learning_rate": 4.1791044776119404e-05,
1556
+ "loss": 0.5743,
1557
+ "step": 221
1558
+ },
1559
+ {
1560
+ "epoch": 1.6610328638497651,
1561
+ "grad_norm": 1.3203125,
1562
+ "learning_rate": 4.1753731343283586e-05,
1563
+ "loss": 0.6427,
1564
+ "step": 222
1565
+ },
1566
+ {
1567
+ "epoch": 1.6685446009389673,
1568
+ "grad_norm": 1.2109375,
1569
+ "learning_rate": 4.171641791044776e-05,
1570
+ "loss": 0.6496,
1571
+ "step": 223
1572
+ },
1573
+ {
1574
+ "epoch": 1.676056338028169,
1575
+ "grad_norm": 0.8515625,
1576
+ "learning_rate": 4.167910447761194e-05,
1577
+ "loss": 0.3463,
1578
+ "step": 224
1579
+ },
1580
+ {
1581
+ "epoch": 1.683568075117371,
1582
+ "grad_norm": 1.34375,
1583
+ "learning_rate": 4.164179104477613e-05,
1584
+ "loss": 0.6243,
1585
+ "step": 225
1586
+ },
1587
+ {
1588
+ "epoch": 1.6910798122065729,
1589
+ "grad_norm": 0.98828125,
1590
+ "learning_rate": 4.16044776119403e-05,
1591
+ "loss": 0.5389,
1592
+ "step": 226
1593
+ },
1594
+ {
1595
+ "epoch": 1.6985915492957746,
1596
+ "grad_norm": 1.0,
1597
+ "learning_rate": 4.156716417910448e-05,
1598
+ "loss": 0.5515,
1599
+ "step": 227
1600
+ },
1601
+ {
1602
+ "epoch": 1.7061032863849765,
1603
+ "grad_norm": 1.2890625,
1604
+ "learning_rate": 4.152985074626866e-05,
1605
+ "loss": 0.6783,
1606
+ "step": 228
1607
+ },
1608
+ {
1609
+ "epoch": 1.7136150234741785,
1610
+ "grad_norm": 1.0390625,
1611
+ "learning_rate": 4.149253731343284e-05,
1612
+ "loss": 0.5099,
1613
+ "step": 229
1614
+ },
1615
+ {
1616
+ "epoch": 1.7211267605633802,
1617
+ "grad_norm": 1.078125,
1618
+ "learning_rate": 4.145522388059702e-05,
1619
+ "loss": 0.461,
1620
+ "step": 230
1621
+ },
1622
+ {
1623
+ "epoch": 1.7286384976525822,
1624
+ "grad_norm": 1.0703125,
1625
+ "learning_rate": 4.1417910447761195e-05,
1626
+ "loss": 0.4953,
1627
+ "step": 231
1628
+ },
1629
+ {
1630
+ "epoch": 1.736150234741784,
1631
+ "grad_norm": 1.1484375,
1632
+ "learning_rate": 4.138059701492538e-05,
1633
+ "loss": 0.5662,
1634
+ "step": 232
1635
+ },
1636
+ {
1637
+ "epoch": 1.7436619718309858,
1638
+ "grad_norm": 1.1171875,
1639
+ "learning_rate": 4.1343283582089554e-05,
1640
+ "loss": 0.524,
1641
+ "step": 233
1642
+ },
1643
+ {
1644
+ "epoch": 1.7511737089201878,
1645
+ "grad_norm": 1.1328125,
1646
+ "learning_rate": 4.130597014925373e-05,
1647
+ "loss": 0.424,
1648
+ "step": 234
1649
+ },
1650
+ {
1651
+ "epoch": 1.7586854460093897,
1652
+ "grad_norm": 1.5625,
1653
+ "learning_rate": 4.126865671641791e-05,
1654
+ "loss": 0.5042,
1655
+ "step": 235
1656
+ },
1657
+ {
1658
+ "epoch": 1.7661971830985914,
1659
+ "grad_norm": 0.9765625,
1660
+ "learning_rate": 4.1231343283582094e-05,
1661
+ "loss": 0.5515,
1662
+ "step": 236
1663
+ },
1664
+ {
1665
+ "epoch": 1.7737089201877936,
1666
+ "grad_norm": 1.0703125,
1667
+ "learning_rate": 4.119402985074627e-05,
1668
+ "loss": 0.6026,
1669
+ "step": 237
1670
+ },
1671
+ {
1672
+ "epoch": 1.7812206572769953,
1673
+ "grad_norm": 1.171875,
1674
+ "learning_rate": 4.115671641791045e-05,
1675
+ "loss": 0.6206,
1676
+ "step": 238
1677
+ },
1678
+ {
1679
+ "epoch": 1.788732394366197,
1680
+ "grad_norm": 1.96875,
1681
+ "learning_rate": 4.111940298507463e-05,
1682
+ "loss": 0.6904,
1683
+ "step": 239
1684
+ },
1685
+ {
1686
+ "epoch": 1.7962441314553992,
1687
+ "grad_norm": 1.0703125,
1688
+ "learning_rate": 4.1082089552238804e-05,
1689
+ "loss": 0.4693,
1690
+ "step": 240
1691
+ },
1692
+ {
1693
+ "epoch": 1.803755868544601,
1694
+ "grad_norm": 0.8984375,
1695
+ "learning_rate": 4.104477611940299e-05,
1696
+ "loss": 0.3648,
1697
+ "step": 241
1698
+ },
1699
+ {
1700
+ "epoch": 1.8112676056338028,
1701
+ "grad_norm": 1.2734375,
1702
+ "learning_rate": 4.100746268656717e-05,
1703
+ "loss": 0.6071,
1704
+ "step": 242
1705
+ },
1706
+ {
1707
+ "epoch": 1.8187793427230048,
1708
+ "grad_norm": 1.1015625,
1709
+ "learning_rate": 4.0970149253731345e-05,
1710
+ "loss": 0.4936,
1711
+ "step": 243
1712
+ },
1713
+ {
1714
+ "epoch": 1.8262910798122065,
1715
+ "grad_norm": 1.0625,
1716
+ "learning_rate": 4.093283582089553e-05,
1717
+ "loss": 0.6627,
1718
+ "step": 244
1719
+ },
1720
+ {
1721
+ "epoch": 1.8338028169014085,
1722
+ "grad_norm": 1.0625,
1723
+ "learning_rate": 4.0895522388059703e-05,
1724
+ "loss": 0.4625,
1725
+ "step": 245
1726
+ },
1727
+ {
1728
+ "epoch": 1.8413145539906104,
1729
+ "grad_norm": 1.2890625,
1730
+ "learning_rate": 4.085820895522388e-05,
1731
+ "loss": 0.5132,
1732
+ "step": 246
1733
+ },
1734
+ {
1735
+ "epoch": 1.8488262910798121,
1736
+ "grad_norm": 0.91796875,
1737
+ "learning_rate": 4.082089552238806e-05,
1738
+ "loss": 0.5464,
1739
+ "step": 247
1740
+ },
1741
+ {
1742
+ "epoch": 1.856338028169014,
1743
+ "grad_norm": 1.5390625,
1744
+ "learning_rate": 4.0783582089552244e-05,
1745
+ "loss": 0.6268,
1746
+ "step": 248
1747
+ },
1748
+ {
1749
+ "epoch": 1.863849765258216,
1750
+ "grad_norm": 1.1015625,
1751
+ "learning_rate": 4.074626865671642e-05,
1752
+ "loss": 0.4702,
1753
+ "step": 249
1754
+ },
1755
+ {
1756
+ "epoch": 1.8713615023474177,
1757
+ "grad_norm": 1.21875,
1758
+ "learning_rate": 4.07089552238806e-05,
1759
+ "loss": 0.6875,
1760
+ "step": 250
1761
+ },
1762
+ {
1763
+ "epoch": 1.8788732394366199,
1764
+ "grad_norm": 1.828125,
1765
+ "learning_rate": 4.067164179104478e-05,
1766
+ "loss": 0.5994,
1767
+ "step": 251
1768
+ },
1769
+ {
1770
+ "epoch": 1.8863849765258216,
1771
+ "grad_norm": 1.1015625,
1772
+ "learning_rate": 4.0634328358208954e-05,
1773
+ "loss": 0.513,
1774
+ "step": 252
1775
+ },
1776
+ {
1777
+ "epoch": 1.8938967136150233,
1778
+ "grad_norm": 0.99609375,
1779
+ "learning_rate": 4.059701492537314e-05,
1780
+ "loss": 0.5234,
1781
+ "step": 253
1782
+ },
1783
+ {
1784
+ "epoch": 1.9014084507042255,
1785
+ "grad_norm": 1.03125,
1786
+ "learning_rate": 4.055970149253732e-05,
1787
+ "loss": 0.3845,
1788
+ "step": 254
1789
+ },
1790
+ {
1791
+ "epoch": 1.9089201877934272,
1792
+ "grad_norm": 0.8515625,
1793
+ "learning_rate": 4.0522388059701495e-05,
1794
+ "loss": 0.5088,
1795
+ "step": 255
1796
+ },
1797
+ {
1798
+ "epoch": 1.9164319248826291,
1799
+ "grad_norm": 1.3359375,
1800
+ "learning_rate": 4.048507462686567e-05,
1801
+ "loss": 0.5853,
1802
+ "step": 256
1803
+ },
1804
+ {
1805
+ "epoch": 1.923943661971831,
1806
+ "grad_norm": 1.0859375,
1807
+ "learning_rate": 4.044776119402985e-05,
1808
+ "loss": 0.5691,
1809
+ "step": 257
1810
+ },
1811
+ {
1812
+ "epoch": 1.9314553990610328,
1813
+ "grad_norm": 1.2890625,
1814
+ "learning_rate": 4.041044776119403e-05,
1815
+ "loss": 0.5251,
1816
+ "step": 258
1817
+ },
1818
+ {
1819
+ "epoch": 1.9389671361502347,
1820
+ "grad_norm": 1.0390625,
1821
+ "learning_rate": 4.037313432835821e-05,
1822
+ "loss": 0.5245,
1823
+ "step": 259
1824
+ },
1825
+ {
1826
+ "epoch": 1.9464788732394367,
1827
+ "grad_norm": 1.078125,
1828
+ "learning_rate": 4.0335820895522394e-05,
1829
+ "loss": 0.5217,
1830
+ "step": 260
1831
+ },
1832
+ {
1833
+ "epoch": 1.9539906103286384,
1834
+ "grad_norm": 1.078125,
1835
+ "learning_rate": 4.029850746268657e-05,
1836
+ "loss": 0.4795,
1837
+ "step": 261
1838
+ },
1839
+ {
1840
+ "epoch": 1.9615023474178404,
1841
+ "grad_norm": 1.109375,
1842
+ "learning_rate": 4.0261194029850746e-05,
1843
+ "loss": 0.5037,
1844
+ "step": 262
1845
+ },
1846
+ {
1847
+ "epoch": 1.9690140845070423,
1848
+ "grad_norm": 1.0,
1849
+ "learning_rate": 4.022388059701493e-05,
1850
+ "loss": 0.5489,
1851
+ "step": 263
1852
+ },
1853
+ {
1854
+ "epoch": 1.976525821596244,
1855
+ "grad_norm": 1.078125,
1856
+ "learning_rate": 4.0186567164179104e-05,
1857
+ "loss": 0.6005,
1858
+ "step": 264
1859
+ },
1860
+ {
1861
+ "epoch": 1.984037558685446,
1862
+ "grad_norm": 0.96875,
1863
+ "learning_rate": 4.014925373134329e-05,
1864
+ "loss": 0.4849,
1865
+ "step": 265
1866
+ },
1867
+ {
1868
+ "epoch": 1.991549295774648,
1869
+ "grad_norm": 1.1171875,
1870
+ "learning_rate": 4.011194029850747e-05,
1871
+ "loss": 0.541,
1872
+ "step": 266
1873
+ },
1874
+ {
1875
+ "epoch": 1.9990610328638496,
1876
+ "grad_norm": 1.109375,
1877
+ "learning_rate": 4.0074626865671645e-05,
1878
+ "loss": 0.5955,
1879
+ "step": 267
1880
+ },
1881
+ {
1882
+ "epoch": 2.0,
1883
+ "grad_norm": 3.390625,
1884
+ "learning_rate": 4.003731343283582e-05,
1885
+ "loss": 0.4322,
1886
+ "step": 268
1887
+ },
1888
+ {
1889
+ "epoch": 2.0075117370892017,
1890
+ "grad_norm": 1.1171875,
1891
+ "learning_rate": 4e-05,
1892
+ "loss": 0.4255,
1893
+ "step": 269
1894
+ },
1895
+ {
1896
+ "epoch": 2.015023474178404,
1897
+ "grad_norm": 1.140625,
1898
+ "learning_rate": 3.996268656716418e-05,
1899
+ "loss": 0.4638,
1900
+ "step": 270
1901
+ },
1902
+ {
1903
+ "epoch": 2.0225352112676056,
1904
+ "grad_norm": 1.046875,
1905
+ "learning_rate": 3.992537313432836e-05,
1906
+ "loss": 0.4727,
1907
+ "step": 271
1908
+ },
1909
+ {
1910
+ "epoch": 2.0300469483568073,
1911
+ "grad_norm": 0.94921875,
1912
+ "learning_rate": 3.9888059701492544e-05,
1913
+ "loss": 0.3631,
1914
+ "step": 272
1915
+ },
1916
+ {
1917
+ "epoch": 2.0375586854460095,
1918
+ "grad_norm": 1.234375,
1919
+ "learning_rate": 3.985074626865672e-05,
1920
+ "loss": 0.4291,
1921
+ "step": 273
1922
+ },
1923
+ {
1924
+ "epoch": 2.045070422535211,
1925
+ "grad_norm": 1.3125,
1926
+ "learning_rate": 3.9813432835820896e-05,
1927
+ "loss": 0.4012,
1928
+ "step": 274
1929
+ },
1930
+ {
1931
+ "epoch": 2.052582159624413,
1932
+ "grad_norm": 1.453125,
1933
+ "learning_rate": 3.977611940298508e-05,
1934
+ "loss": 0.5699,
1935
+ "step": 275
1936
+ },
1937
+ {
1938
+ "epoch": 2.060093896713615,
1939
+ "grad_norm": 1.546875,
1940
+ "learning_rate": 3.9738805970149254e-05,
1941
+ "loss": 0.4192,
1942
+ "step": 276
1943
+ },
1944
+ {
1945
+ "epoch": 2.067605633802817,
1946
+ "grad_norm": 1.5078125,
1947
+ "learning_rate": 3.9701492537313437e-05,
1948
+ "loss": 0.442,
1949
+ "step": 277
1950
+ },
1951
+ {
1952
+ "epoch": 2.075117370892019,
1953
+ "grad_norm": 1.078125,
1954
+ "learning_rate": 3.966417910447761e-05,
1955
+ "loss": 0.3812,
1956
+ "step": 278
1957
+ },
1958
+ {
1959
+ "epoch": 2.0826291079812207,
1960
+ "grad_norm": 0.9609375,
1961
+ "learning_rate": 3.9626865671641795e-05,
1962
+ "loss": 0.3047,
1963
+ "step": 279
1964
+ },
1965
+ {
1966
+ "epoch": 2.0901408450704224,
1967
+ "grad_norm": 1.1640625,
1968
+ "learning_rate": 3.958955223880597e-05,
1969
+ "loss": 0.474,
1970
+ "step": 280
1971
+ },
1972
+ {
1973
+ "epoch": 2.0976525821596246,
1974
+ "grad_norm": 1.1484375,
1975
+ "learning_rate": 3.9552238805970146e-05,
1976
+ "loss": 0.4265,
1977
+ "step": 281
1978
+ },
1979
+ {
1980
+ "epoch": 2.1051643192488263,
1981
+ "grad_norm": 1.0234375,
1982
+ "learning_rate": 3.951492537313433e-05,
1983
+ "loss": 0.4247,
1984
+ "step": 282
1985
+ },
1986
+ {
1987
+ "epoch": 2.112676056338028,
1988
+ "grad_norm": 1.125,
1989
+ "learning_rate": 3.947761194029851e-05,
1990
+ "loss": 0.4229,
1991
+ "step": 283
1992
+ },
1993
+ {
1994
+ "epoch": 2.12018779342723,
1995
+ "grad_norm": 1.03125,
1996
+ "learning_rate": 3.944029850746269e-05,
1997
+ "loss": 0.585,
1998
+ "step": 284
1999
+ },
2000
+ {
2001
+ "epoch": 2.127699530516432,
2002
+ "grad_norm": 0.9921875,
2003
+ "learning_rate": 3.940298507462687e-05,
2004
+ "loss": 0.4583,
2005
+ "step": 285
2006
+ },
2007
+ {
2008
+ "epoch": 2.1352112676056336,
2009
+ "grad_norm": 1.0859375,
2010
+ "learning_rate": 3.9365671641791046e-05,
2011
+ "loss": 0.4564,
2012
+ "step": 286
2013
+ },
2014
+ {
2015
+ "epoch": 2.142723004694836,
2016
+ "grad_norm": 1.1484375,
2017
+ "learning_rate": 3.932835820895522e-05,
2018
+ "loss": 0.4242,
2019
+ "step": 287
2020
+ },
2021
+ {
2022
+ "epoch": 2.1502347417840375,
2023
+ "grad_norm": 1.0625,
2024
+ "learning_rate": 3.9291044776119404e-05,
2025
+ "loss": 0.492,
2026
+ "step": 288
2027
+ },
2028
+ {
2029
+ "epoch": 2.1577464788732392,
2030
+ "grad_norm": 1.109375,
2031
+ "learning_rate": 3.9253731343283586e-05,
2032
+ "loss": 0.389,
2033
+ "step": 289
2034
+ },
2035
+ {
2036
+ "epoch": 2.1652582159624414,
2037
+ "grad_norm": 0.9765625,
2038
+ "learning_rate": 3.921641791044776e-05,
2039
+ "loss": 0.382,
2040
+ "step": 290
2041
+ },
2042
+ {
2043
+ "epoch": 2.172769953051643,
2044
+ "grad_norm": 1.21875,
2045
+ "learning_rate": 3.9179104477611945e-05,
2046
+ "loss": 0.422,
2047
+ "step": 291
2048
+ },
2049
+ {
2050
+ "epoch": 2.1802816901408453,
2051
+ "grad_norm": 1.2734375,
2052
+ "learning_rate": 3.914179104477612e-05,
2053
+ "loss": 0.4392,
2054
+ "step": 292
2055
+ },
2056
+ {
2057
+ "epoch": 2.187793427230047,
2058
+ "grad_norm": 1.234375,
2059
+ "learning_rate": 3.9104477611940296e-05,
2060
+ "loss": 0.5264,
2061
+ "step": 293
2062
+ },
2063
+ {
2064
+ "epoch": 2.1953051643192487,
2065
+ "grad_norm": 1.046875,
2066
+ "learning_rate": 3.906716417910448e-05,
2067
+ "loss": 0.3301,
2068
+ "step": 294
2069
+ },
2070
+ {
2071
+ "epoch": 2.202816901408451,
2072
+ "grad_norm": 0.96875,
2073
+ "learning_rate": 3.902985074626866e-05,
2074
+ "loss": 0.4047,
2075
+ "step": 295
2076
+ },
2077
+ {
2078
+ "epoch": 2.2103286384976526,
2079
+ "grad_norm": 1.328125,
2080
+ "learning_rate": 3.899253731343284e-05,
2081
+ "loss": 0.4317,
2082
+ "step": 296
2083
+ },
2084
+ {
2085
+ "epoch": 2.2178403755868543,
2086
+ "grad_norm": 1.3359375,
2087
+ "learning_rate": 3.895522388059702e-05,
2088
+ "loss": 0.4581,
2089
+ "step": 297
2090
+ },
2091
+ {
2092
+ "epoch": 2.2253521126760565,
2093
+ "grad_norm": 1.296875,
2094
+ "learning_rate": 3.8917910447761195e-05,
2095
+ "loss": 0.4516,
2096
+ "step": 298
2097
+ },
2098
+ {
2099
+ "epoch": 2.232863849765258,
2100
+ "grad_norm": 0.9140625,
2101
+ "learning_rate": 3.888059701492537e-05,
2102
+ "loss": 0.2895,
2103
+ "step": 299
2104
+ },
2105
+ {
2106
+ "epoch": 2.24037558685446,
2107
+ "grad_norm": 1.1640625,
2108
+ "learning_rate": 3.8843283582089554e-05,
2109
+ "loss": 0.4794,
2110
+ "step": 300
2111
+ },
2112
+ {
2113
+ "epoch": 2.247887323943662,
2114
+ "grad_norm": 1.0859375,
2115
+ "learning_rate": 3.8805970149253736e-05,
2116
+ "loss": 0.5919,
2117
+ "step": 301
2118
+ },
2119
+ {
2120
+ "epoch": 2.255399061032864,
2121
+ "grad_norm": 1.015625,
2122
+ "learning_rate": 3.876865671641791e-05,
2123
+ "loss": 0.4234,
2124
+ "step": 302
2125
+ },
2126
+ {
2127
+ "epoch": 2.262910798122066,
2128
+ "grad_norm": 1.8203125,
2129
+ "learning_rate": 3.873134328358209e-05,
2130
+ "loss": 0.4109,
2131
+ "step": 303
2132
+ },
2133
+ {
2134
+ "epoch": 2.2704225352112677,
2135
+ "grad_norm": 1.109375,
2136
+ "learning_rate": 3.869402985074627e-05,
2137
+ "loss": 0.465,
2138
+ "step": 304
2139
+ },
2140
+ {
2141
+ "epoch": 2.2779342723004694,
2142
+ "grad_norm": 1.21875,
2143
+ "learning_rate": 3.8656716417910446e-05,
2144
+ "loss": 0.4098,
2145
+ "step": 305
2146
+ },
2147
+ {
2148
+ "epoch": 2.2854460093896716,
2149
+ "grad_norm": 1.109375,
2150
+ "learning_rate": 3.861940298507463e-05,
2151
+ "loss": 0.4484,
2152
+ "step": 306
2153
+ },
2154
+ {
2155
+ "epoch": 2.2929577464788733,
2156
+ "grad_norm": 1.171875,
2157
+ "learning_rate": 3.858208955223881e-05,
2158
+ "loss": 0.4225,
2159
+ "step": 307
2160
+ },
2161
+ {
2162
+ "epoch": 2.300469483568075,
2163
+ "grad_norm": 1.4296875,
2164
+ "learning_rate": 3.854477611940299e-05,
2165
+ "loss": 0.5249,
2166
+ "step": 308
2167
+ },
2168
+ {
2169
+ "epoch": 2.307981220657277,
2170
+ "grad_norm": 0.96484375,
2171
+ "learning_rate": 3.850746268656716e-05,
2172
+ "loss": 0.376,
2173
+ "step": 309
2174
+ },
2175
+ {
2176
+ "epoch": 2.315492957746479,
2177
+ "grad_norm": 1.28125,
2178
+ "learning_rate": 3.8470149253731345e-05,
2179
+ "loss": 0.4915,
2180
+ "step": 310
2181
+ },
2182
+ {
2183
+ "epoch": 2.3230046948356806,
2184
+ "grad_norm": 0.98046875,
2185
+ "learning_rate": 3.843283582089552e-05,
2186
+ "loss": 0.3064,
2187
+ "step": 311
2188
+ },
2189
+ {
2190
+ "epoch": 2.330516431924883,
2191
+ "grad_norm": 0.953125,
2192
+ "learning_rate": 3.8395522388059704e-05,
2193
+ "loss": 0.2775,
2194
+ "step": 312
2195
+ },
2196
+ {
2197
+ "epoch": 2.3380281690140845,
2198
+ "grad_norm": 1.2421875,
2199
+ "learning_rate": 3.8358208955223886e-05,
2200
+ "loss": 0.4981,
2201
+ "step": 313
2202
+ },
2203
+ {
2204
+ "epoch": 2.345539906103286,
2205
+ "grad_norm": 0.9375,
2206
+ "learning_rate": 3.832089552238806e-05,
2207
+ "loss": 0.3448,
2208
+ "step": 314
2209
+ },
2210
+ {
2211
+ "epoch": 2.3530516431924884,
2212
+ "grad_norm": 0.953125,
2213
+ "learning_rate": 3.828358208955224e-05,
2214
+ "loss": 0.3782,
2215
+ "step": 315
2216
+ },
2217
+ {
2218
+ "epoch": 2.36056338028169,
2219
+ "grad_norm": 1.4453125,
2220
+ "learning_rate": 3.824626865671642e-05,
2221
+ "loss": 0.6991,
2222
+ "step": 316
2223
+ },
2224
+ {
2225
+ "epoch": 2.368075117370892,
2226
+ "grad_norm": 1.0234375,
2227
+ "learning_rate": 3.8208955223880596e-05,
2228
+ "loss": 0.3665,
2229
+ "step": 317
2230
+ },
2231
+ {
2232
+ "epoch": 2.375586854460094,
2233
+ "grad_norm": 1.21875,
2234
+ "learning_rate": 3.817164179104478e-05,
2235
+ "loss": 0.4168,
2236
+ "step": 318
2237
+ },
2238
+ {
2239
+ "epoch": 2.3830985915492957,
2240
+ "grad_norm": 1.1796875,
2241
+ "learning_rate": 3.813432835820896e-05,
2242
+ "loss": 0.4422,
2243
+ "step": 319
2244
+ },
2245
+ {
2246
+ "epoch": 2.3906103286384974,
2247
+ "grad_norm": 1.0234375,
2248
+ "learning_rate": 3.809701492537314e-05,
2249
+ "loss": 0.3613,
2250
+ "step": 320
2251
+ },
2252
+ {
2253
+ "epoch": 2.3981220657276996,
2254
+ "grad_norm": 1.0703125,
2255
+ "learning_rate": 3.805970149253731e-05,
2256
+ "loss": 0.4135,
2257
+ "step": 321
2258
+ },
2259
+ {
2260
+ "epoch": 2.4056338028169013,
2261
+ "grad_norm": 0.91015625,
2262
+ "learning_rate": 3.8022388059701495e-05,
2263
+ "loss": 0.3028,
2264
+ "step": 322
2265
+ },
2266
+ {
2267
+ "epoch": 2.4131455399061035,
2268
+ "grad_norm": 1.0625,
2269
+ "learning_rate": 3.798507462686567e-05,
2270
+ "loss": 0.3747,
2271
+ "step": 323
2272
+ },
2273
+ {
2274
+ "epoch": 2.420657276995305,
2275
+ "grad_norm": 0.8671875,
2276
+ "learning_rate": 3.7947761194029854e-05,
2277
+ "loss": 0.2747,
2278
+ "step": 324
2279
+ },
2280
+ {
2281
+ "epoch": 2.428169014084507,
2282
+ "grad_norm": 0.8984375,
2283
+ "learning_rate": 3.791044776119403e-05,
2284
+ "loss": 0.3254,
2285
+ "step": 325
2286
+ },
2287
+ {
2288
+ "epoch": 2.435680751173709,
2289
+ "grad_norm": 0.984375,
2290
+ "learning_rate": 3.787313432835821e-05,
2291
+ "loss": 0.4222,
2292
+ "step": 326
2293
+ },
2294
+ {
2295
+ "epoch": 2.443192488262911,
2296
+ "grad_norm": 0.96484375,
2297
+ "learning_rate": 3.783582089552239e-05,
2298
+ "loss": 0.3116,
2299
+ "step": 327
2300
+ },
2301
+ {
2302
+ "epoch": 2.4507042253521125,
2303
+ "grad_norm": 4.25,
2304
+ "learning_rate": 3.7798507462686563e-05,
2305
+ "loss": 0.4996,
2306
+ "step": 328
2307
+ },
2308
+ {
2309
+ "epoch": 2.4582159624413147,
2310
+ "grad_norm": 1.15625,
2311
+ "learning_rate": 3.776119402985075e-05,
2312
+ "loss": 0.4413,
2313
+ "step": 329
2314
+ },
2315
+ {
2316
+ "epoch": 2.4657276995305164,
2317
+ "grad_norm": 1.1484375,
2318
+ "learning_rate": 3.772388059701493e-05,
2319
+ "loss": 0.4048,
2320
+ "step": 330
2321
+ },
2322
+ {
2323
+ "epoch": 2.473239436619718,
2324
+ "grad_norm": 0.96484375,
2325
+ "learning_rate": 3.7686567164179104e-05,
2326
+ "loss": 0.4141,
2327
+ "step": 331
2328
+ },
2329
+ {
2330
+ "epoch": 2.4807511737089203,
2331
+ "grad_norm": 1.1640625,
2332
+ "learning_rate": 3.764925373134329e-05,
2333
+ "loss": 0.358,
2334
+ "step": 332
2335
+ },
2336
+ {
2337
+ "epoch": 2.488262910798122,
2338
+ "grad_norm": 1.0390625,
2339
+ "learning_rate": 3.761194029850746e-05,
2340
+ "loss": 0.358,
2341
+ "step": 333
2342
+ },
2343
+ {
2344
+ "epoch": 2.495774647887324,
2345
+ "grad_norm": 1.3046875,
2346
+ "learning_rate": 3.757462686567164e-05,
2347
+ "loss": 0.5101,
2348
+ "step": 334
2349
+ },
2350
+ {
2351
+ "epoch": 2.503286384976526,
2352
+ "grad_norm": 1.125,
2353
+ "learning_rate": 3.753731343283583e-05,
2354
+ "loss": 0.4161,
2355
+ "step": 335
2356
+ },
2357
+ {
2358
+ "epoch": 2.5107981220657276,
2359
+ "grad_norm": 1.15625,
2360
+ "learning_rate": 3.7500000000000003e-05,
2361
+ "loss": 0.4621,
2362
+ "step": 336
2363
+ },
2364
+ {
2365
+ "epoch": 2.5183098591549298,
2366
+ "grad_norm": 0.99609375,
2367
+ "learning_rate": 3.746268656716418e-05,
2368
+ "loss": 0.4423,
2369
+ "step": 337
2370
+ },
2371
+ {
2372
+ "epoch": 2.5258215962441315,
2373
+ "grad_norm": 0.921875,
2374
+ "learning_rate": 3.742537313432836e-05,
2375
+ "loss": 0.3696,
2376
+ "step": 338
2377
+ },
2378
+ {
2379
+ "epoch": 2.533333333333333,
2380
+ "grad_norm": 0.97265625,
2381
+ "learning_rate": 3.738805970149254e-05,
2382
+ "loss": 0.3723,
2383
+ "step": 339
2384
+ },
2385
+ {
2386
+ "epoch": 2.5408450704225354,
2387
+ "grad_norm": 1.1875,
2388
+ "learning_rate": 3.735074626865671e-05,
2389
+ "loss": 0.478,
2390
+ "step": 340
2391
+ },
2392
+ {
2393
+ "epoch": 2.548356807511737,
2394
+ "grad_norm": 1.0625,
2395
+ "learning_rate": 3.73134328358209e-05,
2396
+ "loss": 0.4382,
2397
+ "step": 341
2398
+ },
2399
+ {
2400
+ "epoch": 2.555868544600939,
2401
+ "grad_norm": 1.03125,
2402
+ "learning_rate": 3.727611940298508e-05,
2403
+ "loss": 0.3668,
2404
+ "step": 342
2405
+ },
2406
+ {
2407
+ "epoch": 2.563380281690141,
2408
+ "grad_norm": 1.1796875,
2409
+ "learning_rate": 3.7238805970149254e-05,
2410
+ "loss": 0.4139,
2411
+ "step": 343
2412
+ },
2413
+ {
2414
+ "epoch": 2.5708920187793427,
2415
+ "grad_norm": 1.0625,
2416
+ "learning_rate": 3.720149253731344e-05,
2417
+ "loss": 0.4257,
2418
+ "step": 344
2419
+ },
2420
+ {
2421
+ "epoch": 2.5784037558685444,
2422
+ "grad_norm": 1.0703125,
2423
+ "learning_rate": 3.716417910447761e-05,
2424
+ "loss": 0.49,
2425
+ "step": 345
2426
+ },
2427
+ {
2428
+ "epoch": 2.5859154929577466,
2429
+ "grad_norm": 1.1015625,
2430
+ "learning_rate": 3.7126865671641795e-05,
2431
+ "loss": 0.4485,
2432
+ "step": 346
2433
+ },
2434
+ {
2435
+ "epoch": 2.5934272300469483,
2436
+ "grad_norm": 1.1875,
2437
+ "learning_rate": 3.708955223880598e-05,
2438
+ "loss": 0.451,
2439
+ "step": 347
2440
+ },
2441
+ {
2442
+ "epoch": 2.60093896713615,
2443
+ "grad_norm": 1.078125,
2444
+ "learning_rate": 3.7052238805970153e-05,
2445
+ "loss": 0.4173,
2446
+ "step": 348
2447
+ },
2448
+ {
2449
+ "epoch": 2.608450704225352,
2450
+ "grad_norm": 1.140625,
2451
+ "learning_rate": 3.701492537313433e-05,
2452
+ "loss": 0.3067,
2453
+ "step": 349
2454
+ },
2455
+ {
2456
+ "epoch": 2.615962441314554,
2457
+ "grad_norm": 1.171875,
2458
+ "learning_rate": 3.6977611940298505e-05,
2459
+ "loss": 0.3344,
2460
+ "step": 350
2461
+ },
2462
+ {
2463
+ "epoch": 2.6234741784037556,
2464
+ "grad_norm": 1.0,
2465
+ "learning_rate": 3.694029850746269e-05,
2466
+ "loss": 0.3942,
2467
+ "step": 351
2468
+ },
2469
+ {
2470
+ "epoch": 2.630985915492958,
2471
+ "grad_norm": 1.359375,
2472
+ "learning_rate": 3.690298507462687e-05,
2473
+ "loss": 0.5222,
2474
+ "step": 352
2475
+ },
2476
+ {
2477
+ "epoch": 2.6384976525821595,
2478
+ "grad_norm": 1.03125,
2479
+ "learning_rate": 3.6865671641791046e-05,
2480
+ "loss": 0.2982,
2481
+ "step": 353
2482
+ },
2483
+ {
2484
+ "epoch": 2.6460093896713612,
2485
+ "grad_norm": 1.828125,
2486
+ "learning_rate": 3.682835820895523e-05,
2487
+ "loss": 0.4765,
2488
+ "step": 354
2489
+ },
2490
+ {
2491
+ "epoch": 2.6535211267605634,
2492
+ "grad_norm": 0.9921875,
2493
+ "learning_rate": 3.6791044776119404e-05,
2494
+ "loss": 0.3604,
2495
+ "step": 355
2496
+ },
2497
+ {
2498
+ "epoch": 2.661032863849765,
2499
+ "grad_norm": 1.2109375,
2500
+ "learning_rate": 3.675373134328358e-05,
2501
+ "loss": 0.4295,
2502
+ "step": 356
2503
+ },
2504
+ {
2505
+ "epoch": 2.6685446009389673,
2506
+ "grad_norm": 1.09375,
2507
+ "learning_rate": 3.671641791044776e-05,
2508
+ "loss": 0.3636,
2509
+ "step": 357
2510
+ },
2511
+ {
2512
+ "epoch": 2.676056338028169,
2513
+ "grad_norm": 1.1640625,
2514
+ "learning_rate": 3.6679104477611945e-05,
2515
+ "loss": 0.4205,
2516
+ "step": 358
2517
+ },
2518
+ {
2519
+ "epoch": 2.683568075117371,
2520
+ "grad_norm": 1.2578125,
2521
+ "learning_rate": 3.664179104477612e-05,
2522
+ "loss": 0.6766,
2523
+ "step": 359
2524
+ },
2525
+ {
2526
+ "epoch": 2.691079812206573,
2527
+ "grad_norm": 0.94921875,
2528
+ "learning_rate": 3.66044776119403e-05,
2529
+ "loss": 0.3813,
2530
+ "step": 360
2531
+ },
2532
+ {
2533
+ "epoch": 2.6985915492957746,
2534
+ "grad_norm": 1.09375,
2535
+ "learning_rate": 3.656716417910448e-05,
2536
+ "loss": 0.4775,
2537
+ "step": 361
2538
+ },
2539
+ {
2540
+ "epoch": 2.7061032863849768,
2541
+ "grad_norm": 1.3359375,
2542
+ "learning_rate": 3.6529850746268655e-05,
2543
+ "loss": 0.4854,
2544
+ "step": 362
2545
+ },
2546
+ {
2547
+ "epoch": 2.7136150234741785,
2548
+ "grad_norm": 1.5625,
2549
+ "learning_rate": 3.649253731343284e-05,
2550
+ "loss": 0.5122,
2551
+ "step": 363
2552
+ },
2553
+ {
2554
+ "epoch": 2.72112676056338,
2555
+ "grad_norm": 1.0234375,
2556
+ "learning_rate": 3.645522388059702e-05,
2557
+ "loss": 0.389,
2558
+ "step": 364
2559
+ },
2560
+ {
2561
+ "epoch": 2.7286384976525824,
2562
+ "grad_norm": 1.0546875,
2563
+ "learning_rate": 3.6417910447761196e-05,
2564
+ "loss": 0.4442,
2565
+ "step": 365
2566
+ },
2567
+ {
2568
+ "epoch": 2.736150234741784,
2569
+ "grad_norm": 1.0390625,
2570
+ "learning_rate": 3.638059701492538e-05,
2571
+ "loss": 0.3946,
2572
+ "step": 366
2573
+ },
2574
+ {
2575
+ "epoch": 2.743661971830986,
2576
+ "grad_norm": 1.2578125,
2577
+ "learning_rate": 3.6343283582089554e-05,
2578
+ "loss": 0.3582,
2579
+ "step": 367
2580
+ },
2581
+ {
2582
+ "epoch": 2.751173708920188,
2583
+ "grad_norm": 1.15625,
2584
+ "learning_rate": 3.630597014925373e-05,
2585
+ "loss": 0.4962,
2586
+ "step": 368
2587
+ },
2588
+ {
2589
+ "epoch": 2.7586854460093897,
2590
+ "grad_norm": 1.140625,
2591
+ "learning_rate": 3.626865671641791e-05,
2592
+ "loss": 0.4519,
2593
+ "step": 369
2594
+ },
2595
+ {
2596
+ "epoch": 2.7661971830985914,
2597
+ "grad_norm": 1.046875,
2598
+ "learning_rate": 3.6231343283582095e-05,
2599
+ "loss": 0.4805,
2600
+ "step": 370
2601
+ },
2602
+ {
2603
+ "epoch": 2.7737089201877936,
2604
+ "grad_norm": 1.203125,
2605
+ "learning_rate": 3.619402985074627e-05,
2606
+ "loss": 0.3858,
2607
+ "step": 371
2608
+ },
2609
+ {
2610
+ "epoch": 2.7812206572769953,
2611
+ "grad_norm": 1.25,
2612
+ "learning_rate": 3.6156716417910446e-05,
2613
+ "loss": 0.3975,
2614
+ "step": 372
2615
+ },
2616
+ {
2617
+ "epoch": 2.788732394366197,
2618
+ "grad_norm": 1.2109375,
2619
+ "learning_rate": 3.611940298507463e-05,
2620
+ "loss": 0.5112,
2621
+ "step": 373
2622
+ },
2623
+ {
2624
+ "epoch": 2.796244131455399,
2625
+ "grad_norm": 1.0625,
2626
+ "learning_rate": 3.6082089552238805e-05,
2627
+ "loss": 0.4478,
2628
+ "step": 374
2629
+ },
2630
+ {
2631
+ "epoch": 2.803755868544601,
2632
+ "grad_norm": 1.2265625,
2633
+ "learning_rate": 3.604477611940299e-05,
2634
+ "loss": 0.3357,
2635
+ "step": 375
2636
+ },
2637
+ {
2638
+ "epoch": 2.8112676056338026,
2639
+ "grad_norm": 1.0859375,
2640
+ "learning_rate": 3.600746268656717e-05,
2641
+ "loss": 0.3859,
2642
+ "step": 376
2643
+ },
2644
+ {
2645
+ "epoch": 2.818779342723005,
2646
+ "grad_norm": 0.921875,
2647
+ "learning_rate": 3.5970149253731346e-05,
2648
+ "loss": 0.3302,
2649
+ "step": 377
2650
+ },
2651
+ {
2652
+ "epoch": 2.8262910798122065,
2653
+ "grad_norm": 1.0703125,
2654
+ "learning_rate": 3.593283582089552e-05,
2655
+ "loss": 0.4415,
2656
+ "step": 378
2657
+ },
2658
+ {
2659
+ "epoch": 2.8338028169014082,
2660
+ "grad_norm": 1.0703125,
2661
+ "learning_rate": 3.5895522388059704e-05,
2662
+ "loss": 0.322,
2663
+ "step": 379
2664
+ },
2665
+ {
2666
+ "epoch": 2.8413145539906104,
2667
+ "grad_norm": 1.09375,
2668
+ "learning_rate": 3.585820895522388e-05,
2669
+ "loss": 0.3689,
2670
+ "step": 380
2671
+ },
2672
+ {
2673
+ "epoch": 2.848826291079812,
2674
+ "grad_norm": 1.1484375,
2675
+ "learning_rate": 3.582089552238806e-05,
2676
+ "loss": 0.4491,
2677
+ "step": 381
2678
+ },
2679
+ {
2680
+ "epoch": 2.856338028169014,
2681
+ "grad_norm": 1.1171875,
2682
+ "learning_rate": 3.5783582089552245e-05,
2683
+ "loss": 0.5455,
2684
+ "step": 382
2685
+ },
2686
+ {
2687
+ "epoch": 2.863849765258216,
2688
+ "grad_norm": 1.625,
2689
+ "learning_rate": 3.574626865671642e-05,
2690
+ "loss": 0.4963,
2691
+ "step": 383
2692
+ },
2693
+ {
2694
+ "epoch": 2.8713615023474177,
2695
+ "grad_norm": 1.140625,
2696
+ "learning_rate": 3.5708955223880596e-05,
2697
+ "loss": 0.5053,
2698
+ "step": 384
2699
+ },
2700
+ {
2701
+ "epoch": 2.87887323943662,
2702
+ "grad_norm": 1.0,
2703
+ "learning_rate": 3.567164179104478e-05,
2704
+ "loss": 0.4466,
2705
+ "step": 385
2706
+ },
2707
+ {
2708
+ "epoch": 2.8863849765258216,
2709
+ "grad_norm": 0.9375,
2710
+ "learning_rate": 3.5634328358208955e-05,
2711
+ "loss": 0.3599,
2712
+ "step": 386
2713
+ },
2714
+ {
2715
+ "epoch": 2.8938967136150233,
2716
+ "grad_norm": 1.390625,
2717
+ "learning_rate": 3.559701492537314e-05,
2718
+ "loss": 0.4515,
2719
+ "step": 387
2720
+ },
2721
+ {
2722
+ "epoch": 2.9014084507042255,
2723
+ "grad_norm": 1.0390625,
2724
+ "learning_rate": 3.555970149253732e-05,
2725
+ "loss": 0.4526,
2726
+ "step": 388
2727
+ },
2728
+ {
2729
+ "epoch": 2.908920187793427,
2730
+ "grad_norm": 1.2890625,
2731
+ "learning_rate": 3.5522388059701495e-05,
2732
+ "loss": 0.4239,
2733
+ "step": 389
2734
+ },
2735
+ {
2736
+ "epoch": 2.9164319248826294,
2737
+ "grad_norm": 1.046875,
2738
+ "learning_rate": 3.548507462686567e-05,
2739
+ "loss": 0.3883,
2740
+ "step": 390
2741
+ },
2742
+ {
2743
+ "epoch": 2.923943661971831,
2744
+ "grad_norm": 1.2578125,
2745
+ "learning_rate": 3.5447761194029854e-05,
2746
+ "loss": 0.486,
2747
+ "step": 391
2748
+ },
2749
+ {
2750
+ "epoch": 2.931455399061033,
2751
+ "grad_norm": 1.109375,
2752
+ "learning_rate": 3.541044776119403e-05,
2753
+ "loss": 0.502,
2754
+ "step": 392
2755
+ },
2756
+ {
2757
+ "epoch": 2.938967136150235,
2758
+ "grad_norm": 1.0703125,
2759
+ "learning_rate": 3.537313432835821e-05,
2760
+ "loss": 0.5833,
2761
+ "step": 393
2762
+ },
2763
+ {
2764
+ "epoch": 2.9464788732394367,
2765
+ "grad_norm": 1.1015625,
2766
+ "learning_rate": 3.5335820895522395e-05,
2767
+ "loss": 0.3583,
2768
+ "step": 394
2769
+ },
2770
+ {
2771
+ "epoch": 2.9539906103286384,
2772
+ "grad_norm": 1.078125,
2773
+ "learning_rate": 3.529850746268657e-05,
2774
+ "loss": 0.3327,
2775
+ "step": 395
2776
+ },
2777
+ {
2778
+ "epoch": 2.9615023474178406,
2779
+ "grad_norm": 1.2265625,
2780
+ "learning_rate": 3.5261194029850746e-05,
2781
+ "loss": 0.4078,
2782
+ "step": 396
2783
+ },
2784
+ {
2785
+ "epoch": 2.9690140845070423,
2786
+ "grad_norm": 0.97265625,
2787
+ "learning_rate": 3.522388059701492e-05,
2788
+ "loss": 0.3805,
2789
+ "step": 397
2790
+ },
2791
+ {
2792
+ "epoch": 2.976525821596244,
2793
+ "grad_norm": 0.953125,
2794
+ "learning_rate": 3.5186567164179105e-05,
2795
+ "loss": 0.3543,
2796
+ "step": 398
2797
+ },
2798
+ {
2799
+ "epoch": 2.984037558685446,
2800
+ "grad_norm": 1.03125,
2801
+ "learning_rate": 3.514925373134329e-05,
2802
+ "loss": 0.3794,
2803
+ "step": 399
2804
+ },
2805
+ {
2806
+ "epoch": 2.991549295774648,
2807
+ "grad_norm": 1.1015625,
2808
+ "learning_rate": 3.511194029850746e-05,
2809
+ "loss": 0.4541,
2810
+ "step": 400
2811
+ },
2812
+ {
2813
+ "epoch": 2.9990610328638496,
2814
+ "grad_norm": 1.0625,
2815
+ "learning_rate": 3.5074626865671645e-05,
2816
+ "loss": 0.4267,
2817
+ "step": 401
2818
+ },
2819
+ {
2820
+ "epoch": 3.0,
2821
+ "grad_norm": 3.6875,
2822
+ "learning_rate": 3.503731343283582e-05,
2823
+ "loss": 0.3003,
2824
+ "step": 402
2825
+ }
2826
+ ],
2827
+ "logging_steps": 1,
2828
+ "max_steps": 1340,
2829
+ "num_input_tokens_seen": 0,
2830
+ "num_train_epochs": 10,
2831
+ "save_steps": 500,
2832
+ "stateful_callbacks": {
2833
+ "TrainerControl": {
2834
+ "args": {
2835
+ "should_epoch_stop": false,
2836
+ "should_evaluate": false,
2837
+ "should_log": false,
2838
+ "should_save": true,
2839
+ "should_training_stop": false
2840
+ },
2841
+ "attributes": {}
2842
+ }
2843
+ },
2844
+ "total_flos": 2.056801189059625e+17,
2845
+ "train_batch_size": 4,
2846
+ "trial_name": null,
2847
+ "trial_params": null
2848
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfa2899e4c4bb6a7526e75ce4dbc8f7e426edaac5450013d167dfd3f5e446c26
3
+ size 5969
vocab.json ADDED
The diff for this file is too large to render. See raw diff