Hugging Face's logo

nkkbr
/
ViCA2-thinkng

Video-Text-to-Text
Transformers
Safetensors
sam2
English
vica_qwen
text-generation
multimodal
vision-language
video understanding
visuospatial cognition
spatial reasoning
vlm
llava
qwen
siglip
hiera
dual-encoder
Model card Files
xet
Community
nkkbr commited on
Commit
ba93489
·
1 Parent(s): 5681962

Upload

Browse files
Files changed (15) hide show
  1. added_tokens.json +6 -0
  2. config.json +204 -0
  3. generation_config.json +6 -0
  4. merges.txt +0 -0
  5. model-00001-of-00004.safetensors +3 -0
  6. model-00002-of-00004.safetensors +3 -0
  7. model-00003-of-00004.safetensors +3 -0
  8. model-00004-of-00004.safetensors +3 -0
  9. model.safetensors.index.json +0 -0
  10. special_tokens_map.json +20 -0
  11. tokenizer.json +0 -0
  12. tokenizer_config.json +53 -0
  13. trainer_state.json +2375 -0
  14. training_args.bin +3 -0
  15. vocab.json +0 -0
added_tokens.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "<image>": 151646,
3
+ "<|endoftext|>": 151643,
4
+ "<|im_end|>": 151645,
5
+ "<|im_start|>": 151644
6
+ }
config.json ADDED
@@ -0,0 +1,204 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/home/user1/old_server/250407_vica2_sam2.1_B+_playwithtrainer/work_dirs/0424_final",
3
+ "add_faster_video": false,
4
+ "add_time_instruction": true,
5
+ "architectures": [
6
+ "VicaQwenForCausalLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 151643,
10
+ "eos_token_id": 151645,
11
+ "faster_token_stride": 10,
12
+ "force_sample": true,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 3584,
15
+ "image_aspect_ratio": "anyres_max_9",
16
+ "image_crop_resolution": null,
17
+ "image_grid_pinpoints": [
18
+ [
19
+ 384,
20
+ 384
21
+ ],
22
+ [
23
+ 384,
24
+ 768
25
+ ],
26
+ [
27
+ 384,
28
+ 1152
29
+ ],
30
+ [
31
+ 384,
32
+ 1536
33
+ ],
34
+ [
35
+ 384,
36
+ 1920
37
+ ],
38
+ [
39
+ 384,
40
+ 2304
41
+ ],
42
+ [
43
+ 768,
44
+ 384
45
+ ],
46
+ [
47
+ 768,
48
+ 768
49
+ ],
50
+ [
51
+ 768,
52
+ 1152
53
+ ],
54
+ [
55
+ 768,
56
+ 1536
57
+ ],
58
+ [
59
+ 768,
60
+ 1920
61
+ ],
62
+ [
63
+ 768,
64
+ 2304
65
+ ],
66
+ [
67
+ 1152,
68
+ 384
69
+ ],
70
+ [
71
+ 1152,
72
+ 768
73
+ ],
74
+ [
75
+ 1152,
76
+ 1152
77
+ ],
78
+ [
79
+ 1152,
80
+ 1536
81
+ ],
82
+ [
83
+ 1152,
84
+ 1920
85
+ ],
86
+ [
87
+ 1152,
88
+ 2304
89
+ ],
90
+ [
91
+ 1536,
92
+ 384
93
+ ],
94
+ [
95
+ 1536,
96
+ 768
97
+ ],
98
+ [
99
+ 1536,
100
+ 1152
101
+ ],
102
+ [
103
+ 1536,
104
+ 1536
105
+ ],
106
+ [
107
+ 1536,
108
+ 1920
109
+ ],
110
+ [
111
+ 1536,
112
+ 2304
113
+ ],
114
+ [
115
+ 1920,
116
+ 384
117
+ ],
118
+ [
119
+ 1920,
120
+ 768
121
+ ],
122
+ [
123
+ 1920,
124
+ 1152
125
+ ],
126
+ [
127
+ 1920,
128
+ 1536
129
+ ],
130
+ [
131
+ 1920,
132
+ 1920
133
+ ],
134
+ [
135
+ 1920,
136
+ 2304
137
+ ],
138
+ [
139
+ 2304,
140
+ 384
141
+ ],
142
+ [
143
+ 2304,
144
+ 768
145
+ ],
146
+ [
147
+ 2304,
148
+ 1152
149
+ ],
150
+ [
151
+ 2304,
152
+ 1536
153
+ ],
154
+ [
155
+ 2304,
156
+ 1920
157
+ ],
158
+ [
159
+ 2304,
160
+ 2304
161
+ ]
162
+ ],
163
+ "image_split_resolution": null,
164
+ "image_token_index": 151646,
165
+ "initializer_range": 0.02,
166
+ "intermediate_size": 18944,
167
+ "max_position_embeddings": 32768,
168
+ "max_window_layers": 28,
169
+ "mm_hidden_size": 1152,
170
+ "mm_newline_position": "grid",
171
+ "mm_patch_merge_type": "spatial_unpad",
172
+ "mm_projector_lr": null,
173
+ "mm_projector_type": "mlp2x_gelu",
174
+ "mm_resampler_type": null,
175
+ "mm_spatial_pool_mode": "bilinear",
176
+ "mm_spatial_pool_stride": 2,
177
+ "mm_tunable_parts": "mm_vision_tower,mm_mlp_adapter,mm_language_model,sam_projector",
178
+ "mm_use_im_patch_token": false,
179
+ "mm_use_im_start_end": false,
180
+ "mm_vision_select_feature": "patch",
181
+ "mm_vision_select_layer": -2,
182
+ "mm_vision_tower": "google/siglip-so400m-patch14-384",
183
+ "mm_vision_tower_lr": 2e-06,
184
+ "model_type": "vica_qwen",
185
+ "num_attention_heads": 28,
186
+ "num_hidden_layers": 28,
187
+ "num_key_value_heads": 4,
188
+ "pos_skipping_range": 4096,
189
+ "rms_norm_eps": 1e-06,
190
+ "rope_scaling": null,
191
+ "rope_theta": 1000000.0,
192
+ "sliding_window": 131072,
193
+ "tie_word_embeddings": false,
194
+ "tokenizer_model_max_length": 32768,
195
+ "tokenizer_padding_side": "right",
196
+ "torch_dtype": "bfloat16",
197
+ "transformers_version": "4.40.0.dev0",
198
+ "use_cache": true,
199
+ "use_mm_proj": true,
200
+ "use_pos_skipping": false,
201
+ "use_sliding_window": false,
202
+ "vision_tower_pretrained": null,
203
+ "vocab_size": 151647
204
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 151643,
4
+ "eos_token_id": 151645,
5
+ "transformers_version": "4.40.0.dev0"
6
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model-00001-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9dc34828713cd26566ec60f7f22d34d952f5df36af720f7add4038ab1513b84d
3
+ size 4874678976
model-00002-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d85b633470d5d9ddbf8989bca741ab339ffac68d946e6b3ea1db94f976900b2
3
+ size 4932751008
model-00003-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d5fb1b78ff91dd3523799c4fe01fbc105aa482a95f59c78d06b44b42e28fdae
3
+ size 4994571904
model-00004-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f76ed68e722ba020361f8653255bc78d63a763283aca13a81a55a04df8ef694
3
+ size 1422340008
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
special_tokens_map.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>"
5
+ ],
6
+ "eos_token": {
7
+ "content": "<|im_end|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false
12
+ },
13
+ "pad_token": {
14
+ "content": "<|endoftext|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false
19
+ }
20
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "151643": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": false,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "151644": {
13
+ "content": "<|im_start|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "151645": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "151646": {
29
+ "content": "<image>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ }
36
+ },
37
+ "additional_special_tokens": [
38
+ "<|im_start|>",
39
+ "<|im_end|>"
40
+ ],
41
+ "bos_token": null,
42
+ "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
43
+ "clean_up_tokenization_spaces": false,
44
+ "eos_token": "<|im_end|>",
45
+ "errors": "replace",
46
+ "model_max_length": 32768,
47
+ "pad_token": "<|endoftext|>",
48
+ "padding_side": "right",
49
+ "processor_class": "LlavaProcessor",
50
+ "split_special_tokens": false,
51
+ "tokenizer_class": "Qwen2Tokenizer",
52
+ "unk_token": null
53
+ }
trainer_state.json ADDED
@@ -0,0 +1,2375 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 335,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "grad_norm": 9.992014557607169,
14
+ "learning_rate": 9.090909090909091e-07,
15
+ "loss": 2.7651,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.01,
20
+ "grad_norm": 9.855357081400161,
21
+ "learning_rate": 1.8181818181818183e-06,
22
+ "loss": 2.7106,
23
+ "step": 2
24
+ },
25
+ {
26
+ "epoch": 0.01,
27
+ "grad_norm": 10.452032696456367,
28
+ "learning_rate": 2.7272727272727272e-06,
29
+ "loss": 2.7538,
30
+ "step": 3
31
+ },
32
+ {
33
+ "epoch": 0.01,
34
+ "grad_norm": 9.376280328426684,
35
+ "learning_rate": 3.6363636363636366e-06,
36
+ "loss": 2.5369,
37
+ "step": 4
38
+ },
39
+ {
40
+ "epoch": 0.01,
41
+ "grad_norm": 8.188703375753345,
42
+ "learning_rate": 4.5454545454545455e-06,
43
+ "loss": 2.4723,
44
+ "step": 5
45
+ },
46
+ {
47
+ "epoch": 0.02,
48
+ "grad_norm": 7.173561512899845,
49
+ "learning_rate": 5.4545454545454545e-06,
50
+ "loss": 2.4736,
51
+ "step": 6
52
+ },
53
+ {
54
+ "epoch": 0.02,
55
+ "grad_norm": 7.1450790437206235,
56
+ "learning_rate": 6.363636363636364e-06,
57
+ "loss": 2.4403,
58
+ "step": 7
59
+ },
60
+ {
61
+ "epoch": 0.02,
62
+ "grad_norm": 4.247742053090214,
63
+ "learning_rate": 7.272727272727273e-06,
64
+ "loss": 2.0348,
65
+ "step": 8
66
+ },
67
+ {
68
+ "epoch": 0.03,
69
+ "grad_norm": 4.481406635010285,
70
+ "learning_rate": 8.181818181818183e-06,
71
+ "loss": 1.9127,
72
+ "step": 9
73
+ },
74
+ {
75
+ "epoch": 0.03,
76
+ "grad_norm": 5.521133417095248,
77
+ "learning_rate": 9.090909090909091e-06,
78
+ "loss": 1.891,
79
+ "step": 10
80
+ },
81
+ {
82
+ "epoch": 0.03,
83
+ "grad_norm": 8.731593362969328,
84
+ "learning_rate": 1e-05,
85
+ "loss": 1.8907,
86
+ "step": 11
87
+ },
88
+ {
89
+ "epoch": 0.04,
90
+ "grad_norm": 6.005591131903454,
91
+ "learning_rate": 9.99976495753613e-06,
92
+ "loss": 1.7979,
93
+ "step": 12
94
+ },
95
+ {
96
+ "epoch": 0.04,
97
+ "grad_norm": 3.0731579551147816,
98
+ "learning_rate": 9.999059852242508e-06,
99
+ "loss": 1.6642,
100
+ "step": 13
101
+ },
102
+ {
103
+ "epoch": 0.04,
104
+ "grad_norm": 2.7779072164471383,
105
+ "learning_rate": 9.997884750411004e-06,
106
+ "loss": 1.4797,
107
+ "step": 14
108
+ },
109
+ {
110
+ "epoch": 0.04,
111
+ "grad_norm": 3.3917050377427156,
112
+ "learning_rate": 9.996239762521152e-06,
113
+ "loss": 1.6384,
114
+ "step": 15
115
+ },
116
+ {
117
+ "epoch": 0.05,
118
+ "grad_norm": 3.348652641302203,
119
+ "learning_rate": 9.994125043229753e-06,
120
+ "loss": 1.4289,
121
+ "step": 16
122
+ },
123
+ {
124
+ "epoch": 0.05,
125
+ "grad_norm": 2.7580016074034144,
126
+ "learning_rate": 9.991540791356342e-06,
127
+ "loss": 1.3537,
128
+ "step": 17
129
+ },
130
+ {
131
+ "epoch": 0.05,
132
+ "grad_norm": 2.600060931738886,
133
+ "learning_rate": 9.98848724986449e-06,
134
+ "loss": 1.4414,
135
+ "step": 18
136
+ },
137
+ {
138
+ "epoch": 0.06,
139
+ "grad_norm": 2.402583908854572,
140
+ "learning_rate": 9.98496470583896e-06,
141
+ "loss": 1.4358,
142
+ "step": 19
143
+ },
144
+ {
145
+ "epoch": 0.06,
146
+ "grad_norm": 2.3681534411005365,
147
+ "learning_rate": 9.980973490458728e-06,
148
+ "loss": 1.3664,
149
+ "step": 20
150
+ },
151
+ {
152
+ "epoch": 0.06,
153
+ "grad_norm": 2.417118446828482,
154
+ "learning_rate": 9.976513978965829e-06,
155
+ "loss": 1.2636,
156
+ "step": 21
157
+ },
158
+ {
159
+ "epoch": 0.07,
160
+ "grad_norm": 2.406034881091615,
161
+ "learning_rate": 9.971586590630094e-06,
162
+ "loss": 1.2646,
163
+ "step": 22
164
+ },
165
+ {
166
+ "epoch": 0.07,
167
+ "grad_norm": 2.2282767138519564,
168
+ "learning_rate": 9.966191788709716e-06,
169
+ "loss": 1.2749,
170
+ "step": 23
171
+ },
172
+ {
173
+ "epoch": 0.07,
174
+ "grad_norm": 2.0495586711063853,
175
+ "learning_rate": 9.960330080407712e-06,
176
+ "loss": 1.2495,
177
+ "step": 24
178
+ },
179
+ {
180
+ "epoch": 0.07,
181
+ "grad_norm": 2.232570025850529,
182
+ "learning_rate": 9.954002016824226e-06,
183
+ "loss": 1.255,
184
+ "step": 25
185
+ },
186
+ {
187
+ "epoch": 0.08,
188
+ "grad_norm": 1.9746334176340221,
189
+ "learning_rate": 9.947208192904722e-06,
190
+ "loss": 1.2702,
191
+ "step": 26
192
+ },
193
+ {
194
+ "epoch": 0.08,
195
+ "grad_norm": 1.90630424922607,
196
+ "learning_rate": 9.939949247384046e-06,
197
+ "loss": 1.1942,
198
+ "step": 27
199
+ },
200
+ {
201
+ "epoch": 0.08,
202
+ "grad_norm": 2.0737836420918563,
203
+ "learning_rate": 9.93222586272637e-06,
204
+ "loss": 1.259,
205
+ "step": 28
206
+ },
207
+ {
208
+ "epoch": 0.09,
209
+ "grad_norm": 2.0980819753190514,
210
+ "learning_rate": 9.924038765061042e-06,
211
+ "loss": 1.2254,
212
+ "step": 29
213
+ },
214
+ {
215
+ "epoch": 0.09,
216
+ "grad_norm": 2.07955602263235,
217
+ "learning_rate": 9.915388724114301e-06,
218
+ "loss": 1.1271,
219
+ "step": 30
220
+ },
221
+ {
222
+ "epoch": 0.09,
223
+ "grad_norm": 2.1659071183953023,
224
+ "learning_rate": 9.906276553136924e-06,
225
+ "loss": 1.2434,
226
+ "step": 31
227
+ },
228
+ {
229
+ "epoch": 0.1,
230
+ "grad_norm": 2.1275763181049014,
231
+ "learning_rate": 9.896703108827758e-06,
232
+ "loss": 1.1295,
233
+ "step": 32
234
+ },
235
+ {
236
+ "epoch": 0.1,
237
+ "grad_norm": 2.124992777307241,
238
+ "learning_rate": 9.886669291253178e-06,
239
+ "loss": 1.0909,
240
+ "step": 33
241
+ },
242
+ {
243
+ "epoch": 0.1,
244
+ "grad_norm": 2.1457535781359405,
245
+ "learning_rate": 9.876176043762467e-06,
246
+ "loss": 1.0579,
247
+ "step": 34
248
+ },
249
+ {
250
+ "epoch": 0.1,
251
+ "grad_norm": 2.299797197398565,
252
+ "learning_rate": 9.86522435289912e-06,
253
+ "loss": 1.1403,
254
+ "step": 35
255
+ },
256
+ {
257
+ "epoch": 0.11,
258
+ "grad_norm": 1.9841854125040834,
259
+ "learning_rate": 9.853815248308101e-06,
260
+ "loss": 1.1669,
261
+ "step": 36
262
+ },
263
+ {
264
+ "epoch": 0.11,
265
+ "grad_norm": 1.9449153939112522,
266
+ "learning_rate": 9.841949802639031e-06,
267
+ "loss": 1.177,
268
+ "step": 37
269
+ },
270
+ {
271
+ "epoch": 0.11,
272
+ "grad_norm": 1.823305444585483,
273
+ "learning_rate": 9.829629131445342e-06,
274
+ "loss": 1.0274,
275
+ "step": 38
276
+ },
277
+ {
278
+ "epoch": 0.12,
279
+ "grad_norm": 1.9967356357088837,
280
+ "learning_rate": 9.816854393079402e-06,
281
+ "loss": 1.052,
282
+ "step": 39
283
+ },
284
+ {
285
+ "epoch": 0.12,
286
+ "grad_norm": 2.0742167672185525,
287
+ "learning_rate": 9.803626788583603e-06,
288
+ "loss": 1.0244,
289
+ "step": 40
290
+ },
291
+ {
292
+ "epoch": 0.12,
293
+ "grad_norm": 2.030292578786332,
294
+ "learning_rate": 9.789947561577445e-06,
295
+ "loss": 1.0812,
296
+ "step": 41
297
+ },
298
+ {
299
+ "epoch": 0.13,
300
+ "grad_norm": 2.1285455940700455,
301
+ "learning_rate": 9.775817998140615e-06,
302
+ "loss": 1.0372,
303
+ "step": 42
304
+ },
305
+ {
306
+ "epoch": 0.13,
307
+ "grad_norm": 2.104132559156802,
308
+ "learning_rate": 9.761239426692077e-06,
309
+ "loss": 1.0363,
310
+ "step": 43
311
+ },
312
+ {
313
+ "epoch": 0.13,
314
+ "grad_norm": 2.1801107672452096,
315
+ "learning_rate": 9.74621321786517e-06,
316
+ "loss": 1.1396,
317
+ "step": 44
318
+ },
319
+ {
320
+ "epoch": 0.13,
321
+ "grad_norm": 1.930455012718363,
322
+ "learning_rate": 9.730740784378755e-06,
323
+ "loss": 1.0207,
324
+ "step": 45
325
+ },
326
+ {
327
+ "epoch": 0.14,
328
+ "grad_norm": 1.9114532903731203,
329
+ "learning_rate": 9.71482358090438e-06,
330
+ "loss": 1.0088,
331
+ "step": 46
332
+ },
333
+ {
334
+ "epoch": 0.14,
335
+ "grad_norm": 1.9603910862722937,
336
+ "learning_rate": 9.698463103929542e-06,
337
+ "loss": 1.0367,
338
+ "step": 47
339
+ },
340
+ {
341
+ "epoch": 0.14,
342
+ "grad_norm": 2.0049474141842483,
343
+ "learning_rate": 9.681660891616967e-06,
344
+ "loss": 1.04,
345
+ "step": 48
346
+ },
347
+ {
348
+ "epoch": 0.15,
349
+ "grad_norm": 1.9843717409843054,
350
+ "learning_rate": 9.664418523660004e-06,
351
+ "loss": 1.0541,
352
+ "step": 49
353
+ },
354
+ {
355
+ "epoch": 0.15,
356
+ "grad_norm": 1.8735616571233806,
357
+ "learning_rate": 9.646737621134112e-06,
358
+ "loss": 1.0682,
359
+ "step": 50
360
+ },
361
+ {
362
+ "epoch": 0.15,
363
+ "grad_norm": 2.0255894944970154,
364
+ "learning_rate": 9.628619846344453e-06,
365
+ "loss": 1.072,
366
+ "step": 51
367
+ },
368
+ {
369
+ "epoch": 0.16,
370
+ "grad_norm": 1.8854561919254644,
371
+ "learning_rate": 9.610066902669593e-06,
372
+ "loss": 1.0295,
373
+ "step": 52
374
+ },
375
+ {
376
+ "epoch": 0.16,
377
+ "grad_norm": 1.834056540014944,
378
+ "learning_rate": 9.591080534401371e-06,
379
+ "loss": 1.0185,
380
+ "step": 53
381
+ },
382
+ {
383
+ "epoch": 0.16,
384
+ "grad_norm": 1.9190598160066954,
385
+ "learning_rate": 9.571662526580898e-06,
386
+ "loss": 1.0398,
387
+ "step": 54
388
+ },
389
+ {
390
+ "epoch": 0.16,
391
+ "grad_norm": 1.927342754807542,
392
+ "learning_rate": 9.551814704830734e-06,
393
+ "loss": 1.0799,
394
+ "step": 55
395
+ },
396
+ {
397
+ "epoch": 0.17,
398
+ "grad_norm": 1.9976883134890686,
399
+ "learning_rate": 9.531538935183252e-06,
400
+ "loss": 0.9839,
401
+ "step": 56
402
+ },
403
+ {
404
+ "epoch": 0.17,
405
+ "grad_norm": 2.010561519470896,
406
+ "learning_rate": 9.51083712390519e-06,
407
+ "loss": 1.0452,
408
+ "step": 57
409
+ },
410
+ {
411
+ "epoch": 0.17,
412
+ "grad_norm": 1.8025815822535263,
413
+ "learning_rate": 9.48971121731844e-06,
414
+ "loss": 0.9052,
415
+ "step": 58
416
+ },
417
+ {
418
+ "epoch": 0.18,
419
+ "grad_norm": 2.1672411135729948,
420
+ "learning_rate": 9.468163201617063e-06,
421
+ "loss": 0.9764,
422
+ "step": 59
423
+ },
424
+ {
425
+ "epoch": 0.18,
426
+ "grad_norm": 2.1274344440002197,
427
+ "learning_rate": 9.446195102680531e-06,
428
+ "loss": 1.0161,
429
+ "step": 60
430
+ },
431
+ {
432
+ "epoch": 0.18,
433
+ "grad_norm": 1.9154643005374183,
434
+ "learning_rate": 9.423808985883289e-06,
435
+ "loss": 0.9014,
436
+ "step": 61
437
+ },
438
+ {
439
+ "epoch": 0.19,
440
+ "grad_norm": 1.9253885360090393,
441
+ "learning_rate": 9.401006955900555e-06,
442
+ "loss": 0.9921,
443
+ "step": 62
444
+ },
445
+ {
446
+ "epoch": 0.19,
447
+ "grad_norm": 1.8131597896526979,
448
+ "learning_rate": 9.377791156510456e-06,
449
+ "loss": 0.9996,
450
+ "step": 63
451
+ },
452
+ {
453
+ "epoch": 0.19,
454
+ "grad_norm": 2.047402757936845,
455
+ "learning_rate": 9.35416377039246e-06,
456
+ "loss": 1.0432,
457
+ "step": 64
458
+ },
459
+ {
460
+ "epoch": 0.19,
461
+ "grad_norm": 2.42467076534372,
462
+ "learning_rate": 9.330127018922195e-06,
463
+ "loss": 1.0092,
464
+ "step": 65
465
+ },
466
+ {
467
+ "epoch": 0.2,
468
+ "grad_norm": 2.0406186242464015,
469
+ "learning_rate": 9.305683161962569e-06,
470
+ "loss": 1.0197,
471
+ "step": 66
472
+ },
473
+ {
474
+ "epoch": 0.2,
475
+ "grad_norm": 1.996085329280642,
476
+ "learning_rate": 9.280834497651334e-06,
477
+ "loss": 0.999,
478
+ "step": 67
479
+ },
480
+ {
481
+ "epoch": 0.2,
482
+ "grad_norm": 2.050405093618206,
483
+ "learning_rate": 9.255583362184998e-06,
484
+ "loss": 1.0275,
485
+ "step": 68
486
+ },
487
+ {
488
+ "epoch": 0.21,
489
+ "grad_norm": 1.9391779248159844,
490
+ "learning_rate": 9.229932129599206e-06,
491
+ "loss": 1.0069,
492
+ "step": 69
493
+ },
494
+ {
495
+ "epoch": 0.21,
496
+ "grad_norm": 1.8688340050178358,
497
+ "learning_rate": 9.203883211545517e-06,
498
+ "loss": 1.024,
499
+ "step": 70
500
+ },
501
+ {
502
+ "epoch": 0.21,
503
+ "grad_norm": 1.9107718382648906,
504
+ "learning_rate": 9.177439057064684e-06,
505
+ "loss": 0.9796,
506
+ "step": 71
507
+ },
508
+ {
509
+ "epoch": 0.21,
510
+ "grad_norm": 1.8884098537907934,
511
+ "learning_rate": 9.150602152356394e-06,
512
+ "loss": 1.0094,
513
+ "step": 72
514
+ },
515
+ {
516
+ "epoch": 0.22,
517
+ "grad_norm": 2.0666939259341337,
518
+ "learning_rate": 9.123375020545534e-06,
519
+ "loss": 1.1245,
520
+ "step": 73
521
+ },
522
+ {
523
+ "epoch": 0.22,
524
+ "grad_norm": 1.9312576824640348,
525
+ "learning_rate": 9.09576022144496e-06,
526
+ "loss": 0.9542,
527
+ "step": 74
528
+ },
529
+ {
530
+ "epoch": 0.22,
531
+ "grad_norm": 2.150486624339303,
532
+ "learning_rate": 9.067760351314838e-06,
533
+ "loss": 0.9796,
534
+ "step": 75
535
+ },
536
+ {
537
+ "epoch": 0.23,
538
+ "grad_norm": 2.1900483954756464,
539
+ "learning_rate": 9.039378042618556e-06,
540
+ "loss": 1.1277,
541
+ "step": 76
542
+ },
543
+ {
544
+ "epoch": 0.23,
545
+ "grad_norm": 1.8759460923333156,
546
+ "learning_rate": 9.01061596377522e-06,
547
+ "loss": 0.9205,
548
+ "step": 77
549
+ },
550
+ {
551
+ "epoch": 0.23,
552
+ "grad_norm": 2.0684517325582923,
553
+ "learning_rate": 8.981476818908778e-06,
554
+ "loss": 1.0655,
555
+ "step": 78
556
+ },
557
+ {
558
+ "epoch": 0.24,
559
+ "grad_norm": 1.7015404285579259,
560
+ "learning_rate": 8.951963347593797e-06,
561
+ "loss": 0.9322,
562
+ "step": 79
563
+ },
564
+ {
565
+ "epoch": 0.24,
566
+ "grad_norm": 2.0128565561264113,
567
+ "learning_rate": 8.92207832459788e-06,
568
+ "loss": 1.0096,
569
+ "step": 80
570
+ },
571
+ {
572
+ "epoch": 0.24,
573
+ "grad_norm": 2.2955683670099836,
574
+ "learning_rate": 8.891824559620801e-06,
575
+ "loss": 0.93,
576
+ "step": 81
577
+ },
578
+ {
579
+ "epoch": 0.24,
580
+ "grad_norm": 1.9819221656020602,
581
+ "learning_rate": 8.861204897030346e-06,
582
+ "loss": 0.8111,
583
+ "step": 82
584
+ },
585
+ {
586
+ "epoch": 0.25,
587
+ "grad_norm": 1.8840968076917093,
588
+ "learning_rate": 8.83022221559489e-06,
589
+ "loss": 0.8334,
590
+ "step": 83
591
+ },
592
+ {
593
+ "epoch": 0.25,
594
+ "grad_norm": 1.8148160639213984,
595
+ "learning_rate": 8.798879428212748e-06,
596
+ "loss": 0.9797,
597
+ "step": 84
598
+ },
599
+ {
600
+ "epoch": 0.25,
601
+ "grad_norm": 1.7219644312125897,
602
+ "learning_rate": 8.767179481638303e-06,
603
+ "loss": 0.9265,
604
+ "step": 85
605
+ },
606
+ {
607
+ "epoch": 0.26,
608
+ "grad_norm": 1.8153657619801007,
609
+ "learning_rate": 8.735125356204982e-06,
610
+ "loss": 0.9308,
611
+ "step": 86
612
+ },
613
+ {
614
+ "epoch": 0.26,
615
+ "grad_norm": 2.1864487165433717,
616
+ "learning_rate": 8.702720065545024e-06,
617
+ "loss": 0.8714,
618
+ "step": 87
619
+ },
620
+ {
621
+ "epoch": 0.26,
622
+ "grad_norm": 1.9319176084767922,
623
+ "learning_rate": 8.669966656306176e-06,
624
+ "loss": 0.8996,
625
+ "step": 88
626
+ },
627
+ {
628
+ "epoch": 0.27,
629
+ "grad_norm": 1.8941836323323051,
630
+ "learning_rate": 8.636868207865244e-06,
631
+ "loss": 0.9559,
632
+ "step": 89
633
+ },
634
+ {
635
+ "epoch": 0.27,
636
+ "grad_norm": 2.0942714525772264,
637
+ "learning_rate": 8.603427832038574e-06,
638
+ "loss": 0.9642,
639
+ "step": 90
640
+ },
641
+ {
642
+ "epoch": 0.27,
643
+ "grad_norm": 1.956607567158922,
644
+ "learning_rate": 8.569648672789496e-06,
645
+ "loss": 0.9556,
646
+ "step": 91
647
+ },
648
+ {
649
+ "epoch": 0.27,
650
+ "grad_norm": 2.0359194563727008,
651
+ "learning_rate": 8.535533905932739e-06,
652
+ "loss": 0.9433,
653
+ "step": 92
654
+ },
655
+ {
656
+ "epoch": 0.28,
657
+ "grad_norm": 1.8831570100390131,
658
+ "learning_rate": 8.501086738835843e-06,
659
+ "loss": 0.9653,
660
+ "step": 93
661
+ },
662
+ {
663
+ "epoch": 0.28,
664
+ "grad_norm": 1.970446143582905,
665
+ "learning_rate": 8.466310410117622e-06,
666
+ "loss": 0.9637,
667
+ "step": 94
668
+ },
669
+ {
670
+ "epoch": 0.28,
671
+ "grad_norm": 2.0437287034957903,
672
+ "learning_rate": 8.43120818934367e-06,
673
+ "loss": 0.9639,
674
+ "step": 95
675
+ },
676
+ {
677
+ "epoch": 0.29,
678
+ "grad_norm": 1.9551401347597726,
679
+ "learning_rate": 8.395783376718967e-06,
680
+ "loss": 0.908,
681
+ "step": 96
682
+ },
683
+ {
684
+ "epoch": 0.29,
685
+ "grad_norm": 1.9195968647766901,
686
+ "learning_rate": 8.360039302777614e-06,
687
+ "loss": 0.9048,
688
+ "step": 97
689
+ },
690
+ {
691
+ "epoch": 0.29,
692
+ "grad_norm": 1.853992686163623,
693
+ "learning_rate": 8.323979328069689e-06,
694
+ "loss": 0.9351,
695
+ "step": 98
696
+ },
697
+ {
698
+ "epoch": 0.3,
699
+ "grad_norm": 1.8308983331681774,
700
+ "learning_rate": 8.28760684284532e-06,
701
+ "loss": 0.9443,
702
+ "step": 99
703
+ },
704
+ {
705
+ "epoch": 0.3,
706
+ "grad_norm": 2.1769204299308758,
707
+ "learning_rate": 8.25092526673592e-06,
708
+ "loss": 1.0337,
709
+ "step": 100
710
+ },
711
+ {
712
+ "epoch": 0.3,
713
+ "grad_norm": 2.0101187852614335,
714
+ "learning_rate": 8.213938048432697e-06,
715
+ "loss": 0.9319,
716
+ "step": 101
717
+ },
718
+ {
719
+ "epoch": 0.3,
720
+ "grad_norm": 2.1098348893292616,
721
+ "learning_rate": 8.176648665362426e-06,
722
+ "loss": 0.8138,
723
+ "step": 102
724
+ },
725
+ {
726
+ "epoch": 0.31,
727
+ "grad_norm": 2.018298542101901,
728
+ "learning_rate": 8.139060623360494e-06,
729
+ "loss": 0.9322,
730
+ "step": 103
731
+ },
732
+ {
733
+ "epoch": 0.31,
734
+ "grad_norm": 1.97268436525157,
735
+ "learning_rate": 8.101177456341301e-06,
736
+ "loss": 0.908,
737
+ "step": 104
738
+ },
739
+ {
740
+ "epoch": 0.31,
741
+ "grad_norm": 1.8905564445487661,
742
+ "learning_rate": 8.063002725966014e-06,
743
+ "loss": 0.9059,
744
+ "step": 105
745
+ },
746
+ {
747
+ "epoch": 0.32,
748
+ "grad_norm": 1.8040796552589426,
749
+ "learning_rate": 8.024540021307709e-06,
750
+ "loss": 0.8834,
751
+ "step": 106
752
+ },
753
+ {
754
+ "epoch": 0.32,
755
+ "grad_norm": 1.8099754444058382,
756
+ "learning_rate": 7.985792958513932e-06,
757
+ "loss": 0.949,
758
+ "step": 107
759
+ },
760
+ {
761
+ "epoch": 0.32,
762
+ "grad_norm": 1.859283445011557,
763
+ "learning_rate": 7.946765180466725e-06,
764
+ "loss": 0.8598,
765
+ "step": 108
766
+ },
767
+ {
768
+ "epoch": 0.33,
769
+ "grad_norm": 2.017173689625784,
770
+ "learning_rate": 7.907460356440133e-06,
771
+ "loss": 0.98,
772
+ "step": 109
773
+ },
774
+ {
775
+ "epoch": 0.33,
776
+ "grad_norm": 1.9067977524786353,
777
+ "learning_rate": 7.86788218175523e-06,
778
+ "loss": 0.9113,
779
+ "step": 110
780
+ },
781
+ {
782
+ "epoch": 0.33,
783
+ "grad_norm": 1.8356374617177302,
784
+ "learning_rate": 7.828034377432694e-06,
785
+ "loss": 0.9379,
786
+ "step": 111
787
+ },
788
+ {
789
+ "epoch": 0.33,
790
+ "grad_norm": 1.9671234496145011,
791
+ "learning_rate": 7.787920689842965e-06,
792
+ "loss": 0.971,
793
+ "step": 112
794
+ },
795
+ {
796
+ "epoch": 0.34,
797
+ "grad_norm": 1.9057301922594099,
798
+ "learning_rate": 7.747544890354031e-06,
799
+ "loss": 0.9351,
800
+ "step": 113
801
+ },
802
+ {
803
+ "epoch": 0.34,
804
+ "grad_norm": 1.8176638279870596,
805
+ "learning_rate": 7.706910774976849e-06,
806
+ "loss": 0.8129,
807
+ "step": 114
808
+ },
809
+ {
810
+ "epoch": 0.34,
811
+ "grad_norm": 1.8480732550035386,
812
+ "learning_rate": 7.666022164008458e-06,
813
+ "loss": 0.9127,
814
+ "step": 115
815
+ },
816
+ {
817
+ "epoch": 0.35,
818
+ "grad_norm": 1.8342675263668593,
819
+ "learning_rate": 7.624882901672801e-06,
820
+ "loss": 0.8875,
821
+ "step": 116
822
+ },
823
+ {
824
+ "epoch": 0.35,
825
+ "grad_norm": 1.8479911871602939,
826
+ "learning_rate": 7.5834968557593155e-06,
827
+ "loss": 0.9814,
828
+ "step": 117
829
+ },
830
+ {
831
+ "epoch": 0.35,
832
+ "grad_norm": 1.953930833517739,
833
+ "learning_rate": 7.541867917259278e-06,
834
+ "loss": 1.0459,
835
+ "step": 118
836
+ },
837
+ {
838
+ "epoch": 0.36,
839
+ "grad_norm": 1.816248946402888,
840
+ "learning_rate": 7.500000000000001e-06,
841
+ "loss": 0.8606,
842
+ "step": 119
843
+ },
844
+ {
845
+ "epoch": 0.36,
846
+ "grad_norm": 1.829758704776544,
847
+ "learning_rate": 7.457897040276853e-06,
848
+ "loss": 0.9917,
849
+ "step": 120
850
+ },
851
+ {
852
+ "epoch": 0.36,
853
+ "grad_norm": 1.7798304171360442,
854
+ "learning_rate": 7.415562996483193e-06,
855
+ "loss": 0.9226,
856
+ "step": 121
857
+ },
858
+ {
859
+ "epoch": 0.36,
860
+ "grad_norm": 1.949746360544544,
861
+ "learning_rate": 7.373001848738203e-06,
862
+ "loss": 1.0089,
863
+ "step": 122
864
+ },
865
+ {
866
+ "epoch": 0.37,
867
+ "grad_norm": 1.7709880415336614,
868
+ "learning_rate": 7.330217598512696e-06,
869
+ "loss": 0.8108,
870
+ "step": 123
871
+ },
872
+ {
873
+ "epoch": 0.37,
874
+ "grad_norm": 1.8079747568178568,
875
+ "learning_rate": 7.2872142682529045e-06,
876
+ "loss": 0.8246,
877
+ "step": 124
878
+ },
879
+ {
880
+ "epoch": 0.37,
881
+ "grad_norm": 1.8171645027732521,
882
+ "learning_rate": 7.243995901002312e-06,
883
+ "loss": 0.9032,
884
+ "step": 125
885
+ },
886
+ {
887
+ "epoch": 0.38,
888
+ "grad_norm": 1.875345818100136,
889
+ "learning_rate": 7.200566560021525e-06,
890
+ "loss": 0.8692,
891
+ "step": 126
892
+ },
893
+ {
894
+ "epoch": 0.38,
895
+ "grad_norm": 1.9933228492567425,
896
+ "learning_rate": 7.156930328406268e-06,
897
+ "loss": 0.93,
898
+ "step": 127
899
+ },
900
+ {
901
+ "epoch": 0.38,
902
+ "grad_norm": 1.9411283328658857,
903
+ "learning_rate": 7.113091308703498e-06,
904
+ "loss": 0.9071,
905
+ "step": 128
906
+ },
907
+ {
908
+ "epoch": 0.39,
909
+ "grad_norm": 2.090341113079029,
910
+ "learning_rate": 7.069053622525697e-06,
911
+ "loss": 1.043,
912
+ "step": 129
913
+ },
914
+ {
915
+ "epoch": 0.39,
916
+ "grad_norm": 1.91058658362823,
917
+ "learning_rate": 7.0248214101633685e-06,
918
+ "loss": 0.8674,
919
+ "step": 130
920
+ },
921
+ {
922
+ "epoch": 0.39,
923
+ "grad_norm": 1.9696219193781328,
924
+ "learning_rate": 6.980398830195785e-06,
925
+ "loss": 1.0577,
926
+ "step": 131
927
+ },
928
+ {
929
+ "epoch": 0.39,
930
+ "grad_norm": 1.9555717096721157,
931
+ "learning_rate": 6.9357900591000034e-06,
932
+ "loss": 0.9531,
933
+ "step": 132
934
+ },
935
+ {
936
+ "epoch": 0.4,
937
+ "grad_norm": 1.8951036355458846,
938
+ "learning_rate": 6.890999290858213e-06,
939
+ "loss": 1.006,
940
+ "step": 133
941
+ },
942
+ {
943
+ "epoch": 0.4,
944
+ "grad_norm": 1.8568126429433485,
945
+ "learning_rate": 6.8460307365634225e-06,
946
+ "loss": 0.8623,
947
+ "step": 134
948
+ },
949
+ {
950
+ "epoch": 0.4,
951
+ "grad_norm": 1.8704210955739529,
952
+ "learning_rate": 6.800888624023552e-06,
953
+ "loss": 0.9493,
954
+ "step": 135
955
+ },
956
+ {
957
+ "epoch": 0.41,
958
+ "grad_norm": 1.832260849480108,
959
+ "learning_rate": 6.755577197363945e-06,
960
+ "loss": 0.8205,
961
+ "step": 136
962
+ },
963
+ {
964
+ "epoch": 0.41,
965
+ "grad_norm": 2.031520077283429,
966
+ "learning_rate": 6.710100716628345e-06,
967
+ "loss": 0.8792,
968
+ "step": 137
969
+ },
970
+ {
971
+ "epoch": 0.41,
972
+ "grad_norm": 1.7410235600999984,
973
+ "learning_rate": 6.6644634573783825e-06,
974
+ "loss": 0.8951,
975
+ "step": 138
976
+ },
977
+ {
978
+ "epoch": 0.41,
979
+ "grad_norm": 1.843737618356284,
980
+ "learning_rate": 6.618669710291607e-06,
981
+ "loss": 0.8085,
982
+ "step": 139
983
+ },
984
+ {
985
+ "epoch": 0.42,
986
+ "grad_norm": 1.765762644440963,
987
+ "learning_rate": 6.572723780758069e-06,
988
+ "loss": 0.8957,
989
+ "step": 140
990
+ },
991
+ {
992
+ "epoch": 0.42,
993
+ "grad_norm": 1.9511281601416317,
994
+ "learning_rate": 6.526629988475567e-06,
995
+ "loss": 0.9,
996
+ "step": 141
997
+ },
998
+ {
999
+ "epoch": 0.42,
1000
+ "grad_norm": 1.7773818546444415,
1001
+ "learning_rate": 6.4803926670435e-06,
1002
+ "loss": 0.9234,
1003
+ "step": 142
1004
+ },
1005
+ {
1006
+ "epoch": 0.43,
1007
+ "grad_norm": 1.8692189103556698,
1008
+ "learning_rate": 6.434016163555452e-06,
1009
+ "loss": 0.8899,
1010
+ "step": 143
1011
+ },
1012
+ {
1013
+ "epoch": 0.43,
1014
+ "grad_norm": 1.8255302240308842,
1015
+ "learning_rate": 6.387504838190479e-06,
1016
+ "loss": 0.8086,
1017
+ "step": 144
1018
+ },
1019
+ {
1020
+ "epoch": 0.43,
1021
+ "grad_norm": 1.8578841539686264,
1022
+ "learning_rate": 6.340863063803187e-06,
1023
+ "loss": 0.8248,
1024
+ "step": 145
1025
+ },
1026
+ {
1027
+ "epoch": 0.44,
1028
+ "grad_norm": 1.7994136623563066,
1029
+ "learning_rate": 6.294095225512604e-06,
1030
+ "loss": 0.933,
1031
+ "step": 146
1032
+ },
1033
+ {
1034
+ "epoch": 0.44,
1035
+ "grad_norm": 2.0121355236133978,
1036
+ "learning_rate": 6.247205720289907e-06,
1037
+ "loss": 1.0521,
1038
+ "step": 147
1039
+ },
1040
+ {
1041
+ "epoch": 0.44,
1042
+ "grad_norm": 1.869460952399974,
1043
+ "learning_rate": 6.2001989565450305e-06,
1044
+ "loss": 0.8515,
1045
+ "step": 148
1046
+ },
1047
+ {
1048
+ "epoch": 0.44,
1049
+ "grad_norm": 1.8246450967462686,
1050
+ "learning_rate": 6.153079353712201e-06,
1051
+ "loss": 0.852,
1052
+ "step": 149
1053
+ },
1054
+ {
1055
+ "epoch": 0.45,
1056
+ "grad_norm": 1.9648734937960934,
1057
+ "learning_rate": 6.105851341834439e-06,
1058
+ "loss": 0.8434,
1059
+ "step": 150
1060
+ },
1061
+ {
1062
+ "epoch": 0.45,
1063
+ "grad_norm": 1.7658190156459108,
1064
+ "learning_rate": 6.058519361147055e-06,
1065
+ "loss": 0.8406,
1066
+ "step": 151
1067
+ },
1068
+ {
1069
+ "epoch": 0.45,
1070
+ "grad_norm": 1.8896576006369,
1071
+ "learning_rate": 6.011087861660191e-06,
1072
+ "loss": 0.8235,
1073
+ "step": 152
1074
+ },
1075
+ {
1076
+ "epoch": 0.46,
1077
+ "grad_norm": 1.701511265987839,
1078
+ "learning_rate": 5.9635613027404495e-06,
1079
+ "loss": 0.83,
1080
+ "step": 153
1081
+ },
1082
+ {
1083
+ "epoch": 0.46,
1084
+ "grad_norm": 1.8458019088871649,
1085
+ "learning_rate": 5.915944152691634e-06,
1086
+ "loss": 0.8061,
1087
+ "step": 154
1088
+ },
1089
+ {
1090
+ "epoch": 0.46,
1091
+ "grad_norm": 1.8758600805480732,
1092
+ "learning_rate": 5.8682408883346535e-06,
1093
+ "loss": 0.8738,
1094
+ "step": 155
1095
+ },
1096
+ {
1097
+ "epoch": 0.47,
1098
+ "grad_norm": 1.7770461619321762,
1099
+ "learning_rate": 5.820455994586621e-06,
1100
+ "loss": 0.8144,
1101
+ "step": 156
1102
+ },
1103
+ {
1104
+ "epoch": 0.47,
1105
+ "grad_norm": 2.06521139231319,
1106
+ "learning_rate": 5.772593964039203e-06,
1107
+ "loss": 1.0683,
1108
+ "step": 157
1109
+ },
1110
+ {
1111
+ "epoch": 0.47,
1112
+ "grad_norm": 1.8949940538942156,
1113
+ "learning_rate": 5.724659296536234e-06,
1114
+ "loss": 0.8378,
1115
+ "step": 158
1116
+ },
1117
+ {
1118
+ "epoch": 0.47,
1119
+ "grad_norm": 1.8329328421055082,
1120
+ "learning_rate": 5.6766564987506564e-06,
1121
+ "loss": 0.8059,
1122
+ "step": 159
1123
+ },
1124
+ {
1125
+ "epoch": 0.48,
1126
+ "grad_norm": 1.9356837835306697,
1127
+ "learning_rate": 5.628590083760815e-06,
1128
+ "loss": 0.8759,
1129
+ "step": 160
1130
+ },
1131
+ {
1132
+ "epoch": 0.48,
1133
+ "grad_norm": 1.914824151289401,
1134
+ "learning_rate": 5.5804645706261515e-06,
1135
+ "loss": 0.9575,
1136
+ "step": 161
1137
+ },
1138
+ {
1139
+ "epoch": 0.48,
1140
+ "grad_norm": 1.9610745203544777,
1141
+ "learning_rate": 5.532284483962341e-06,
1142
+ "loss": 0.8548,
1143
+ "step": 162
1144
+ },
1145
+ {
1146
+ "epoch": 0.49,
1147
+ "grad_norm": 1.9847850375923817,
1148
+ "learning_rate": 5.484054353515896e-06,
1149
+ "loss": 0.9005,
1150
+ "step": 163
1151
+ },
1152
+ {
1153
+ "epoch": 0.49,
1154
+ "grad_norm": 1.900418899811939,
1155
+ "learning_rate": 5.435778713738292e-06,
1156
+ "loss": 0.9209,
1157
+ "step": 164
1158
+ },
1159
+ {
1160
+ "epoch": 0.49,
1161
+ "grad_norm": 1.7892930994134306,
1162
+ "learning_rate": 5.387462103359655e-06,
1163
+ "loss": 0.8555,
1164
+ "step": 165
1165
+ },
1166
+ {
1167
+ "epoch": 0.5,
1168
+ "grad_norm": 1.774160649412005,
1169
+ "learning_rate": 5.339109064962047e-06,
1170
+ "loss": 0.7846,
1171
+ "step": 166
1172
+ },
1173
+ {
1174
+ "epoch": 0.5,
1175
+ "grad_norm": 1.9563920124901706,
1176
+ "learning_rate": 5.290724144552379e-06,
1177
+ "loss": 0.7538,
1178
+ "step": 167
1179
+ },
1180
+ {
1181
+ "epoch": 0.5,
1182
+ "grad_norm": 1.9120830031418432,
1183
+ "learning_rate": 5.242311891135016e-06,
1184
+ "loss": 0.9287,
1185
+ "step": 168
1186
+ },
1187
+ {
1188
+ "epoch": 0.5,
1189
+ "grad_norm": 1.7820515084191753,
1190
+ "learning_rate": 5.193876856284085e-06,
1191
+ "loss": 0.7843,
1192
+ "step": 169
1193
+ },
1194
+ {
1195
+ "epoch": 0.51,
1196
+ "grad_norm": 1.8838516311972966,
1197
+ "learning_rate": 5.145423593715558e-06,
1198
+ "loss": 0.8848,
1199
+ "step": 170
1200
+ },
1201
+ {
1202
+ "epoch": 0.51,
1203
+ "grad_norm": 1.8000718579254187,
1204
+ "learning_rate": 5.096956658859122e-06,
1205
+ "loss": 0.8422,
1206
+ "step": 171
1207
+ },
1208
+ {
1209
+ "epoch": 0.51,
1210
+ "grad_norm": 1.9252361586236277,
1211
+ "learning_rate": 5.048480608429893e-06,
1212
+ "loss": 0.788,
1213
+ "step": 172
1214
+ },
1215
+ {
1216
+ "epoch": 0.52,
1217
+ "grad_norm": 1.7838973065656873,
1218
+ "learning_rate": 5e-06,
1219
+ "loss": 0.8419,
1220
+ "step": 173
1221
+ },
1222
+ {
1223
+ "epoch": 0.52,
1224
+ "grad_norm": 1.7406962046278718,
1225
+ "learning_rate": 4.951519391570108e-06,
1226
+ "loss": 0.8617,
1227
+ "step": 174
1228
+ },
1229
+ {
1230
+ "epoch": 0.52,
1231
+ "grad_norm": 1.833052653443519,
1232
+ "learning_rate": 4.903043341140879e-06,
1233
+ "loss": 0.7839,
1234
+ "step": 175
1235
+ },
1236
+ {
1237
+ "epoch": 0.53,
1238
+ "grad_norm": 1.9635309949293125,
1239
+ "learning_rate": 4.854576406284443e-06,
1240
+ "loss": 0.745,
1241
+ "step": 176
1242
+ },
1243
+ {
1244
+ "epoch": 0.53,
1245
+ "grad_norm": 1.8878508304922175,
1246
+ "learning_rate": 4.806123143715916e-06,
1247
+ "loss": 0.898,
1248
+ "step": 177
1249
+ },
1250
+ {
1251
+ "epoch": 0.53,
1252
+ "grad_norm": 1.8490253561704422,
1253
+ "learning_rate": 4.7576881088649865e-06,
1254
+ "loss": 0.8764,
1255
+ "step": 178
1256
+ },
1257
+ {
1258
+ "epoch": 0.53,
1259
+ "grad_norm": 1.8492342801201405,
1260
+ "learning_rate": 4.7092758554476215e-06,
1261
+ "loss": 0.8734,
1262
+ "step": 179
1263
+ },
1264
+ {
1265
+ "epoch": 0.54,
1266
+ "grad_norm": 1.678333641735353,
1267
+ "learning_rate": 4.660890935037954e-06,
1268
+ "loss": 0.8476,
1269
+ "step": 180
1270
+ },
1271
+ {
1272
+ "epoch": 0.54,
1273
+ "grad_norm": 1.9197951126550667,
1274
+ "learning_rate": 4.6125378966403465e-06,
1275
+ "loss": 0.8502,
1276
+ "step": 181
1277
+ },
1278
+ {
1279
+ "epoch": 0.54,
1280
+ "grad_norm": 1.8065516269516981,
1281
+ "learning_rate": 4.564221286261709e-06,
1282
+ "loss": 0.8434,
1283
+ "step": 182
1284
+ },
1285
+ {
1286
+ "epoch": 0.55,
1287
+ "grad_norm": 2.090063964079408,
1288
+ "learning_rate": 4.515945646484105e-06,
1289
+ "loss": 0.9075,
1290
+ "step": 183
1291
+ },
1292
+ {
1293
+ "epoch": 0.55,
1294
+ "grad_norm": 1.714803378036167,
1295
+ "learning_rate": 4.467715516037659e-06,
1296
+ "loss": 0.8979,
1297
+ "step": 184
1298
+ },
1299
+ {
1300
+ "epoch": 0.55,
1301
+ "grad_norm": 1.8920297684515237,
1302
+ "learning_rate": 4.4195354293738484e-06,
1303
+ "loss": 0.917,
1304
+ "step": 185
1305
+ },
1306
+ {
1307
+ "epoch": 0.56,
1308
+ "grad_norm": 1.9646366988613642,
1309
+ "learning_rate": 4.371409916239188e-06,
1310
+ "loss": 0.8406,
1311
+ "step": 186
1312
+ },
1313
+ {
1314
+ "epoch": 0.56,
1315
+ "grad_norm": 2.0095779435113643,
1316
+ "learning_rate": 4.323343501249346e-06,
1317
+ "loss": 0.8105,
1318
+ "step": 187
1319
+ },
1320
+ {
1321
+ "epoch": 0.56,
1322
+ "grad_norm": 1.84960716754979,
1323
+ "learning_rate": 4.275340703463767e-06,
1324
+ "loss": 0.709,
1325
+ "step": 188
1326
+ },
1327
+ {
1328
+ "epoch": 0.56,
1329
+ "grad_norm": 1.8784422747688236,
1330
+ "learning_rate": 4.227406035960798e-06,
1331
+ "loss": 0.8721,
1332
+ "step": 189
1333
+ },
1334
+ {
1335
+ "epoch": 0.57,
1336
+ "grad_norm": 1.7892389168859104,
1337
+ "learning_rate": 4.17954400541338e-06,
1338
+ "loss": 0.8237,
1339
+ "step": 190
1340
+ },
1341
+ {
1342
+ "epoch": 0.57,
1343
+ "grad_norm": 1.834621493465168,
1344
+ "learning_rate": 4.131759111665349e-06,
1345
+ "loss": 0.858,
1346
+ "step": 191
1347
+ },
1348
+ {
1349
+ "epoch": 0.57,
1350
+ "grad_norm": 2.0244248532984312,
1351
+ "learning_rate": 4.084055847308367e-06,
1352
+ "loss": 0.84,
1353
+ "step": 192
1354
+ },
1355
+ {
1356
+ "epoch": 0.58,
1357
+ "grad_norm": 1.7956990250048366,
1358
+ "learning_rate": 4.036438697259551e-06,
1359
+ "loss": 0.8343,
1360
+ "step": 193
1361
+ },
1362
+ {
1363
+ "epoch": 0.58,
1364
+ "grad_norm": 2.008026972496449,
1365
+ "learning_rate": 3.988912138339812e-06,
1366
+ "loss": 0.858,
1367
+ "step": 194
1368
+ },
1369
+ {
1370
+ "epoch": 0.58,
1371
+ "grad_norm": 2.02397192889752,
1372
+ "learning_rate": 3.941480638852948e-06,
1373
+ "loss": 0.8685,
1374
+ "step": 195
1375
+ },
1376
+ {
1377
+ "epoch": 0.59,
1378
+ "grad_norm": 1.8202366956390423,
1379
+ "learning_rate": 3.894148658165562e-06,
1380
+ "loss": 0.8428,
1381
+ "step": 196
1382
+ },
1383
+ {
1384
+ "epoch": 0.59,
1385
+ "grad_norm": 1.8728629012942652,
1386
+ "learning_rate": 3.8469206462878e-06,
1387
+ "loss": 0.833,
1388
+ "step": 197
1389
+ },
1390
+ {
1391
+ "epoch": 0.59,
1392
+ "grad_norm": 1.920110342063804,
1393
+ "learning_rate": 3.7998010434549716e-06,
1394
+ "loss": 0.923,
1395
+ "step": 198
1396
+ },
1397
+ {
1398
+ "epoch": 0.59,
1399
+ "grad_norm": 1.8750275450908436,
1400
+ "learning_rate": 3.752794279710094e-06,
1401
+ "loss": 0.8113,
1402
+ "step": 199
1403
+ },
1404
+ {
1405
+ "epoch": 0.6,
1406
+ "grad_norm": 1.7555094657763348,
1407
+ "learning_rate": 3.705904774487396e-06,
1408
+ "loss": 0.8442,
1409
+ "step": 200
1410
+ },
1411
+ {
1412
+ "epoch": 0.6,
1413
+ "grad_norm": 1.9946734158465282,
1414
+ "learning_rate": 3.6591369361968127e-06,
1415
+ "loss": 0.8566,
1416
+ "step": 201
1417
+ },
1418
+ {
1419
+ "epoch": 0.6,
1420
+ "grad_norm": 1.792346256212117,
1421
+ "learning_rate": 3.6124951618095224e-06,
1422
+ "loss": 0.8306,
1423
+ "step": 202
1424
+ },
1425
+ {
1426
+ "epoch": 0.61,
1427
+ "grad_norm": 1.7737697004497117,
1428
+ "learning_rate": 3.5659838364445505e-06,
1429
+ "loss": 0.827,
1430
+ "step": 203
1431
+ },
1432
+ {
1433
+ "epoch": 0.61,
1434
+ "grad_norm": 1.9896055506585861,
1435
+ "learning_rate": 3.519607332956502e-06,
1436
+ "loss": 0.8967,
1437
+ "step": 204
1438
+ },
1439
+ {
1440
+ "epoch": 0.61,
1441
+ "grad_norm": 2.0112406931128954,
1442
+ "learning_rate": 3.473370011524435e-06,
1443
+ "loss": 0.9582,
1444
+ "step": 205
1445
+ },
1446
+ {
1447
+ "epoch": 0.61,
1448
+ "grad_norm": 1.6933223451923685,
1449
+ "learning_rate": 3.427276219241933e-06,
1450
+ "loss": 0.7972,
1451
+ "step": 206
1452
+ },
1453
+ {
1454
+ "epoch": 0.62,
1455
+ "grad_norm": 1.7960032297149964,
1456
+ "learning_rate": 3.3813302897083955e-06,
1457
+ "loss": 0.91,
1458
+ "step": 207
1459
+ },
1460
+ {
1461
+ "epoch": 0.62,
1462
+ "grad_norm": 1.8409756298969775,
1463
+ "learning_rate": 3.335536542621617e-06,
1464
+ "loss": 0.8442,
1465
+ "step": 208
1466
+ },
1467
+ {
1468
+ "epoch": 0.62,
1469
+ "grad_norm": 1.8943440767619342,
1470
+ "learning_rate": 3.289899283371657e-06,
1471
+ "loss": 0.8782,
1472
+ "step": 209
1473
+ },
1474
+ {
1475
+ "epoch": 0.63,
1476
+ "grad_norm": 1.983602448512353,
1477
+ "learning_rate": 3.244422802636057e-06,
1478
+ "loss": 0.9121,
1479
+ "step": 210
1480
+ },
1481
+ {
1482
+ "epoch": 0.63,
1483
+ "grad_norm": 1.919904148023849,
1484
+ "learning_rate": 3.1991113759764493e-06,
1485
+ "loss": 0.8452,
1486
+ "step": 211
1487
+ },
1488
+ {
1489
+ "epoch": 0.63,
1490
+ "grad_norm": 1.7436938594681772,
1491
+ "learning_rate": 3.1539692634365788e-06,
1492
+ "loss": 0.8367,
1493
+ "step": 212
1494
+ },
1495
+ {
1496
+ "epoch": 0.64,
1497
+ "grad_norm": 1.9023867686458018,
1498
+ "learning_rate": 3.1090007091417884e-06,
1499
+ "loss": 0.7576,
1500
+ "step": 213
1501
+ },
1502
+ {
1503
+ "epoch": 0.64,
1504
+ "grad_norm": 1.8340656721562485,
1505
+ "learning_rate": 3.0642099408999982e-06,
1506
+ "loss": 0.9066,
1507
+ "step": 214
1508
+ },
1509
+ {
1510
+ "epoch": 0.64,
1511
+ "grad_norm": 2.123899315024522,
1512
+ "learning_rate": 3.019601169804216e-06,
1513
+ "loss": 0.8908,
1514
+ "step": 215
1515
+ },
1516
+ {
1517
+ "epoch": 0.64,
1518
+ "grad_norm": 1.970762510888818,
1519
+ "learning_rate": 2.975178589836632e-06,
1520
+ "loss": 0.8324,
1521
+ "step": 216
1522
+ },
1523
+ {
1524
+ "epoch": 0.65,
1525
+ "grad_norm": 1.7131201352172498,
1526
+ "learning_rate": 2.9309463774743047e-06,
1527
+ "loss": 0.9371,
1528
+ "step": 217
1529
+ },
1530
+ {
1531
+ "epoch": 0.65,
1532
+ "grad_norm": 1.692464765973636,
1533
+ "learning_rate": 2.886908691296504e-06,
1534
+ "loss": 0.8384,
1535
+ "step": 218
1536
+ },
1537
+ {
1538
+ "epoch": 0.65,
1539
+ "grad_norm": 1.754853534401181,
1540
+ "learning_rate": 2.843069671593734e-06,
1541
+ "loss": 0.8639,
1542
+ "step": 219
1543
+ },
1544
+ {
1545
+ "epoch": 0.66,
1546
+ "grad_norm": 1.7746221308047176,
1547
+ "learning_rate": 2.7994334399784773e-06,
1548
+ "loss": 0.8123,
1549
+ "step": 220
1550
+ },
1551
+ {
1552
+ "epoch": 0.66,
1553
+ "grad_norm": 1.7797496232199694,
1554
+ "learning_rate": 2.7560040989976894e-06,
1555
+ "loss": 0.8011,
1556
+ "step": 221
1557
+ },
1558
+ {
1559
+ "epoch": 0.66,
1560
+ "grad_norm": 1.7965856236183813,
1561
+ "learning_rate": 2.7127857317470967e-06,
1562
+ "loss": 0.8614,
1563
+ "step": 222
1564
+ },
1565
+ {
1566
+ "epoch": 0.67,
1567
+ "grad_norm": 1.895456729064966,
1568
+ "learning_rate": 2.6697824014873076e-06,
1569
+ "loss": 0.808,
1570
+ "step": 223
1571
+ },
1572
+ {
1573
+ "epoch": 0.67,
1574
+ "grad_norm": 1.6731115830560004,
1575
+ "learning_rate": 2.626998151261798e-06,
1576
+ "loss": 0.7644,
1577
+ "step": 224
1578
+ },
1579
+ {
1580
+ "epoch": 0.67,
1581
+ "grad_norm": 1.722068279724202,
1582
+ "learning_rate": 2.5844370035168077e-06,
1583
+ "loss": 0.7876,
1584
+ "step": 225
1585
+ },
1586
+ {
1587
+ "epoch": 0.67,
1588
+ "grad_norm": 1.8560082384134242,
1589
+ "learning_rate": 2.5421029597231476e-06,
1590
+ "loss": 0.8448,
1591
+ "step": 226
1592
+ },
1593
+ {
1594
+ "epoch": 0.68,
1595
+ "grad_norm": 1.8503877478285864,
1596
+ "learning_rate": 2.5000000000000015e-06,
1597
+ "loss": 0.8241,
1598
+ "step": 227
1599
+ },
1600
+ {
1601
+ "epoch": 0.68,
1602
+ "grad_norm": 1.9115184771008713,
1603
+ "learning_rate": 2.458132082740724e-06,
1604
+ "loss": 0.8465,
1605
+ "step": 228
1606
+ },
1607
+ {
1608
+ "epoch": 0.68,
1609
+ "grad_norm": 1.9363363832594316,
1610
+ "learning_rate": 2.4165031442406857e-06,
1611
+ "loss": 0.8231,
1612
+ "step": 229
1613
+ },
1614
+ {
1615
+ "epoch": 0.69,
1616
+ "grad_norm": 1.9209239281193067,
1617
+ "learning_rate": 2.3751170983272e-06,
1618
+ "loss": 0.9008,
1619
+ "step": 230
1620
+ },
1621
+ {
1622
+ "epoch": 0.69,
1623
+ "grad_norm": 1.8848510564703413,
1624
+ "learning_rate": 2.333977835991545e-06,
1625
+ "loss": 0.789,
1626
+ "step": 231
1627
+ },
1628
+ {
1629
+ "epoch": 0.69,
1630
+ "grad_norm": 1.885613538323485,
1631
+ "learning_rate": 2.293089225023152e-06,
1632
+ "loss": 0.9544,
1633
+ "step": 232
1634
+ },
1635
+ {
1636
+ "epoch": 0.7,
1637
+ "grad_norm": 1.9378339418142523,
1638
+ "learning_rate": 2.2524551096459703e-06,
1639
+ "loss": 0.8512,
1640
+ "step": 233
1641
+ },
1642
+ {
1643
+ "epoch": 0.7,
1644
+ "grad_norm": 1.8499392383494713,
1645
+ "learning_rate": 2.2120793101570366e-06,
1646
+ "loss": 0.7881,
1647
+ "step": 234
1648
+ },
1649
+ {
1650
+ "epoch": 0.7,
1651
+ "grad_norm": 1.8123212923542507,
1652
+ "learning_rate": 2.171965622567308e-06,
1653
+ "loss": 0.8658,
1654
+ "step": 235
1655
+ },
1656
+ {
1657
+ "epoch": 0.7,
1658
+ "grad_norm": 1.8978780385206444,
1659
+ "learning_rate": 2.132117818244771e-06,
1660
+ "loss": 0.8045,
1661
+ "step": 236
1662
+ },
1663
+ {
1664
+ "epoch": 0.71,
1665
+ "grad_norm": 1.7796628138142092,
1666
+ "learning_rate": 2.0925396435598665e-06,
1667
+ "loss": 0.813,
1668
+ "step": 237
1669
+ },
1670
+ {
1671
+ "epoch": 0.71,
1672
+ "grad_norm": 1.8442874060590642,
1673
+ "learning_rate": 2.053234819533276e-06,
1674
+ "loss": 0.7706,
1675
+ "step": 238
1676
+ },
1677
+ {
1678
+ "epoch": 0.71,
1679
+ "grad_norm": 1.8289341236387466,
1680
+ "learning_rate": 2.0142070414860704e-06,
1681
+ "loss": 0.8927,
1682
+ "step": 239
1683
+ },
1684
+ {
1685
+ "epoch": 0.72,
1686
+ "grad_norm": 1.9766538998425796,
1687
+ "learning_rate": 1.9754599786922913e-06,
1688
+ "loss": 0.848,
1689
+ "step": 240
1690
+ },
1691
+ {
1692
+ "epoch": 0.72,
1693
+ "grad_norm": 1.7214028779109611,
1694
+ "learning_rate": 1.936997274033986e-06,
1695
+ "loss": 0.7225,
1696
+ "step": 241
1697
+ },
1698
+ {
1699
+ "epoch": 0.72,
1700
+ "grad_norm": 1.5814120081508478,
1701
+ "learning_rate": 1.8988225436587005e-06,
1702
+ "loss": 0.7288,
1703
+ "step": 242
1704
+ },
1705
+ {
1706
+ "epoch": 0.73,
1707
+ "grad_norm": 1.7940729269240667,
1708
+ "learning_rate": 1.8609393766395083e-06,
1709
+ "loss": 0.9349,
1710
+ "step": 243
1711
+ },
1712
+ {
1713
+ "epoch": 0.73,
1714
+ "grad_norm": 1.878940225083856,
1715
+ "learning_rate": 1.823351334637576e-06,
1716
+ "loss": 0.8962,
1717
+ "step": 244
1718
+ },
1719
+ {
1720
+ "epoch": 0.73,
1721
+ "grad_norm": 1.8155856668279255,
1722
+ "learning_rate": 1.7860619515673034e-06,
1723
+ "loss": 0.88,
1724
+ "step": 245
1725
+ },
1726
+ {
1727
+ "epoch": 0.73,
1728
+ "grad_norm": 1.9394793705542583,
1729
+ "learning_rate": 1.7490747332640833e-06,
1730
+ "loss": 0.897,
1731
+ "step": 246
1732
+ },
1733
+ {
1734
+ "epoch": 0.74,
1735
+ "grad_norm": 1.8855178833940114,
1736
+ "learning_rate": 1.7123931571546826e-06,
1737
+ "loss": 0.8955,
1738
+ "step": 247
1739
+ },
1740
+ {
1741
+ "epoch": 0.74,
1742
+ "grad_norm": 1.680634839138001,
1743
+ "learning_rate": 1.6760206719303107e-06,
1744
+ "loss": 0.781,
1745
+ "step": 248
1746
+ },
1747
+ {
1748
+ "epoch": 0.74,
1749
+ "grad_norm": 1.8281579952033755,
1750
+ "learning_rate": 1.639960697222388e-06,
1751
+ "loss": 0.9578,
1752
+ "step": 249
1753
+ },
1754
+ {
1755
+ "epoch": 0.75,
1756
+ "grad_norm": 1.8390758231951718,
1757
+ "learning_rate": 1.6042166232810346e-06,
1758
+ "loss": 0.8377,
1759
+ "step": 250
1760
+ },
1761
+ {
1762
+ "epoch": 0.75,
1763
+ "grad_norm": 1.637452500506161,
1764
+ "learning_rate": 1.5687918106563326e-06,
1765
+ "loss": 0.7837,
1766
+ "step": 251
1767
+ },
1768
+ {
1769
+ "epoch": 0.75,
1770
+ "grad_norm": 1.8615373812049087,
1771
+ "learning_rate": 1.5336895898823801e-06,
1772
+ "loss": 0.8947,
1773
+ "step": 252
1774
+ },
1775
+ {
1776
+ "epoch": 0.76,
1777
+ "grad_norm": 1.8795666079438391,
1778
+ "learning_rate": 1.4989132611641576e-06,
1779
+ "loss": 0.8214,
1780
+ "step": 253
1781
+ },
1782
+ {
1783
+ "epoch": 0.76,
1784
+ "grad_norm": 2.1220008217273385,
1785
+ "learning_rate": 1.4644660940672628e-06,
1786
+ "loss": 0.8473,
1787
+ "step": 254
1788
+ },
1789
+ {
1790
+ "epoch": 0.76,
1791
+ "grad_norm": 2.016744480195016,
1792
+ "learning_rate": 1.4303513272105057e-06,
1793
+ "loss": 0.7648,
1794
+ "step": 255
1795
+ },
1796
+ {
1797
+ "epoch": 0.76,
1798
+ "grad_norm": 1.8273845868646243,
1799
+ "learning_rate": 1.396572167961427e-06,
1800
+ "loss": 0.9056,
1801
+ "step": 256
1802
+ },
1803
+ {
1804
+ "epoch": 0.77,
1805
+ "grad_norm": 1.9242593040018354,
1806
+ "learning_rate": 1.3631317921347564e-06,
1807
+ "loss": 0.8546,
1808
+ "step": 257
1809
+ },
1810
+ {
1811
+ "epoch": 0.77,
1812
+ "grad_norm": 1.7602720727922225,
1813
+ "learning_rate": 1.330033343693824e-06,
1814
+ "loss": 0.8168,
1815
+ "step": 258
1816
+ },
1817
+ {
1818
+ "epoch": 0.77,
1819
+ "grad_norm": 1.7981469172314437,
1820
+ "learning_rate": 1.297279934454978e-06,
1821
+ "loss": 0.8196,
1822
+ "step": 259
1823
+ },
1824
+ {
1825
+ "epoch": 0.78,
1826
+ "grad_norm": 1.934034355132605,
1827
+ "learning_rate": 1.264874643795021e-06,
1828
+ "loss": 0.8641,
1829
+ "step": 260
1830
+ },
1831
+ {
1832
+ "epoch": 0.78,
1833
+ "grad_norm": 1.731914923464836,
1834
+ "learning_rate": 1.2328205183616964e-06,
1835
+ "loss": 0.8316,
1836
+ "step": 261
1837
+ },
1838
+ {
1839
+ "epoch": 0.78,
1840
+ "grad_norm": 1.8671301669332583,
1841
+ "learning_rate": 1.2011205717872538e-06,
1842
+ "loss": 0.8973,
1843
+ "step": 262
1844
+ },
1845
+ {
1846
+ "epoch": 0.79,
1847
+ "grad_norm": 1.8258264225083662,
1848
+ "learning_rate": 1.1697777844051105e-06,
1849
+ "loss": 0.827,
1850
+ "step": 263
1851
+ },
1852
+ {
1853
+ "epoch": 0.79,
1854
+ "grad_norm": 1.76891519967969,
1855
+ "learning_rate": 1.1387951029696543e-06,
1856
+ "loss": 0.8073,
1857
+ "step": 264
1858
+ },
1859
+ {
1860
+ "epoch": 0.79,
1861
+ "grad_norm": 1.8056333135913247,
1862
+ "learning_rate": 1.1081754403792e-06,
1863
+ "loss": 0.8111,
1864
+ "step": 265
1865
+ },
1866
+ {
1867
+ "epoch": 0.79,
1868
+ "grad_norm": 1.6693657914010887,
1869
+ "learning_rate": 1.0779216754021215e-06,
1870
+ "loss": 0.8481,
1871
+ "step": 266
1872
+ },
1873
+ {
1874
+ "epoch": 0.8,
1875
+ "grad_norm": 1.8914182275757614,
1876
+ "learning_rate": 1.0480366524062041e-06,
1877
+ "loss": 0.7854,
1878
+ "step": 267
1879
+ },
1880
+ {
1881
+ "epoch": 0.8,
1882
+ "grad_norm": 1.7952669162789925,
1883
+ "learning_rate": 1.0185231810912223e-06,
1884
+ "loss": 0.8331,
1885
+ "step": 268
1886
+ },
1887
+ {
1888
+ "epoch": 0.8,
1889
+ "grad_norm": 1.8041970383317338,
1890
+ "learning_rate": 9.893840362247809e-07,
1891
+ "loss": 0.808,
1892
+ "step": 269
1893
+ },
1894
+ {
1895
+ "epoch": 0.81,
1896
+ "grad_norm": 1.7349503739814947,
1897
+ "learning_rate": 9.606219573814447e-07,
1898
+ "loss": 0.782,
1899
+ "step": 270
1900
+ },
1901
+ {
1902
+ "epoch": 0.81,
1903
+ "grad_norm": 1.971273696916632,
1904
+ "learning_rate": 9.322396486851626e-07,
1905
+ "loss": 0.8777,
1906
+ "step": 271
1907
+ },
1908
+ {
1909
+ "epoch": 0.81,
1910
+ "grad_norm": 1.6916809211406005,
1911
+ "learning_rate": 9.042397785550405e-07,
1912
+ "loss": 0.816,
1913
+ "step": 272
1914
+ },
1915
+ {
1916
+ "epoch": 0.81,
1917
+ "grad_norm": 1.859123501478176,
1918
+ "learning_rate": 8.766249794544662e-07,
1919
+ "loss": 0.8536,
1920
+ "step": 273
1921
+ },
1922
+ {
1923
+ "epoch": 0.82,
1924
+ "grad_norm": 1.9296428042034799,
1925
+ "learning_rate": 8.49397847643606e-07,
1926
+ "loss": 0.9167,
1927
+ "step": 274
1928
+ },
1929
+ {
1930
+ "epoch": 0.82,
1931
+ "grad_norm": 1.9014753831076923,
1932
+ "learning_rate": 8.225609429353187e-07,
1933
+ "loss": 0.8984,
1934
+ "step": 275
1935
+ },
1936
+ {
1937
+ "epoch": 0.82,
1938
+ "grad_norm": 1.9966126008374474,
1939
+ "learning_rate": 7.961167884544852e-07,
1940
+ "loss": 0.8615,
1941
+ "step": 276
1942
+ },
1943
+ {
1944
+ "epoch": 0.83,
1945
+ "grad_norm": 1.7845802462997415,
1946
+ "learning_rate": 7.700678704007947e-07,
1947
+ "loss": 0.8524,
1948
+ "step": 277
1949
+ },
1950
+ {
1951
+ "epoch": 0.83,
1952
+ "grad_norm": 1.786614071106851,
1953
+ "learning_rate": 7.444166378150014e-07,
1954
+ "loss": 0.8085,
1955
+ "step": 278
1956
+ },
1957
+ {
1958
+ "epoch": 0.83,
1959
+ "grad_norm": 1.6646768752142675,
1960
+ "learning_rate": 7.191655023486682e-07,
1961
+ "loss": 0.867,
1962
+ "step": 279
1963
+ },
1964
+ {
1965
+ "epoch": 0.84,
1966
+ "grad_norm": 1.934324323024249,
1967
+ "learning_rate": 6.94316838037431e-07,
1968
+ "loss": 0.724,
1969
+ "step": 280
1970
+ },
1971
+ {
1972
+ "epoch": 0.84,
1973
+ "grad_norm": 1.7670716798370054,
1974
+ "learning_rate": 6.698729810778065e-07,
1975
+ "loss": 0.7882,
1976
+ "step": 281
1977
+ },
1978
+ {
1979
+ "epoch": 0.84,
1980
+ "grad_norm": 1.8041863674588816,
1981
+ "learning_rate": 6.458362296075399e-07,
1982
+ "loss": 0.7681,
1983
+ "step": 282
1984
+ },
1985
+ {
1986
+ "epoch": 0.84,
1987
+ "grad_norm": 1.7612589809223944,
1988
+ "learning_rate": 6.222088434895462e-07,
1989
+ "loss": 0.8392,
1990
+ "step": 283
1991
+ },
1992
+ {
1993
+ "epoch": 0.85,
1994
+ "grad_norm": 1.749669895373114,
1995
+ "learning_rate": 5.989930440994451e-07,
1996
+ "loss": 0.7651,
1997
+ "step": 284
1998
+ },
1999
+ {
2000
+ "epoch": 0.85,
2001
+ "grad_norm": 1.856400426555766,
2002
+ "learning_rate": 5.76191014116711e-07,
2003
+ "loss": 0.854,
2004
+ "step": 285
2005
+ },
2006
+ {
2007
+ "epoch": 0.85,
2008
+ "grad_norm": 1.786307733562441,
2009
+ "learning_rate": 5.538048973194699e-07,
2010
+ "loss": 0.8124,
2011
+ "step": 286
2012
+ },
2013
+ {
2014
+ "epoch": 0.86,
2015
+ "grad_norm": 1.7598458461279902,
2016
+ "learning_rate": 5.318367983829393e-07,
2017
+ "loss": 0.822,
2018
+ "step": 287
2019
+ },
2020
+ {
2021
+ "epoch": 0.86,
2022
+ "grad_norm": 1.6759526002862657,
2023
+ "learning_rate": 5.102887826815589e-07,
2024
+ "loss": 0.7365,
2025
+ "step": 288
2026
+ },
2027
+ {
2028
+ "epoch": 0.86,
2029
+ "grad_norm": 1.7936397119836314,
2030
+ "learning_rate": 4.891628760948114e-07,
2031
+ "loss": 0.7666,
2032
+ "step": 289
2033
+ },
2034
+ {
2035
+ "epoch": 0.87,
2036
+ "grad_norm": 1.8498525973183433,
2037
+ "learning_rate": 4.6846106481675035e-07,
2038
+ "loss": 0.8682,
2039
+ "step": 290
2040
+ },
2041
+ {
2042
+ "epoch": 0.87,
2043
+ "grad_norm": 1.7420235116184377,
2044
+ "learning_rate": 4.481852951692672e-07,
2045
+ "loss": 0.7485,
2046
+ "step": 291
2047
+ },
2048
+ {
2049
+ "epoch": 0.87,
2050
+ "grad_norm": 1.877611598521712,
2051
+ "learning_rate": 4.283374734191037e-07,
2052
+ "loss": 0.9166,
2053
+ "step": 292
2054
+ },
2055
+ {
2056
+ "epoch": 0.87,
2057
+ "grad_norm": 1.9492518664893324,
2058
+ "learning_rate": 4.089194655986306e-07,
2059
+ "loss": 0.7579,
2060
+ "step": 293
2061
+ },
2062
+ {
2063
+ "epoch": 0.88,
2064
+ "grad_norm": 1.9457494252701653,
2065
+ "learning_rate": 3.899330973304083e-07,
2066
+ "loss": 0.9035,
2067
+ "step": 294
2068
+ },
2069
+ {
2070
+ "epoch": 0.88,
2071
+ "grad_norm": 1.9791967456606612,
2072
+ "learning_rate": 3.7138015365554834e-07,
2073
+ "loss": 0.7953,
2074
+ "step": 295
2075
+ },
2076
+ {
2077
+ "epoch": 0.88,
2078
+ "grad_norm": 1.794711427524111,
2079
+ "learning_rate": 3.5326237886588734e-07,
2080
+ "loss": 0.847,
2081
+ "step": 296
2082
+ },
2083
+ {
2084
+ "epoch": 0.89,
2085
+ "grad_norm": 1.8794850272880976,
2086
+ "learning_rate": 3.355814763399973e-07,
2087
+ "loss": 0.8514,
2088
+ "step": 297
2089
+ },
2090
+ {
2091
+ "epoch": 0.89,
2092
+ "grad_norm": 1.790194856863824,
2093
+ "learning_rate": 3.183391083830345e-07,
2094
+ "loss": 0.8278,
2095
+ "step": 298
2096
+ },
2097
+ {
2098
+ "epoch": 0.89,
2099
+ "grad_norm": 1.8717181093829611,
2100
+ "learning_rate": 3.015368960704584e-07,
2101
+ "loss": 0.8429,
2102
+ "step": 299
2103
+ },
2104
+ {
2105
+ "epoch": 0.9,
2106
+ "grad_norm": 1.7463890355109786,
2107
+ "learning_rate": 2.8517641909562075e-07,
2108
+ "loss": 0.8278,
2109
+ "step": 300
2110
+ },
2111
+ {
2112
+ "epoch": 0.9,
2113
+ "grad_norm": 1.855252737220356,
2114
+ "learning_rate": 2.6925921562124867e-07,
2115
+ "loss": 0.6983,
2116
+ "step": 301
2117
+ },
2118
+ {
2119
+ "epoch": 0.9,
2120
+ "grad_norm": 1.8352359161688767,
2121
+ "learning_rate": 2.5378678213483057e-07,
2122
+ "loss": 0.8415,
2123
+ "step": 302
2124
+ },
2125
+ {
2126
+ "epoch": 0.9,
2127
+ "grad_norm": 1.8346745144152707,
2128
+ "learning_rate": 2.3876057330792344e-07,
2129
+ "loss": 0.8257,
2130
+ "step": 303
2131
+ },
2132
+ {
2133
+ "epoch": 0.91,
2134
+ "grad_norm": 1.8337279964237236,
2135
+ "learning_rate": 2.2418200185938488e-07,
2136
+ "loss": 0.8155,
2137
+ "step": 304
2138
+ },
2139
+ {
2140
+ "epoch": 0.91,
2141
+ "grad_norm": 1.9002602072773978,
2142
+ "learning_rate": 2.1005243842255552e-07,
2143
+ "loss": 0.9803,
2144
+ "step": 305
2145
+ },
2146
+ {
2147
+ "epoch": 0.91,
2148
+ "grad_norm": 1.7966402315107077,
2149
+ "learning_rate": 1.9637321141639743e-07,
2150
+ "loss": 0.7586,
2151
+ "step": 306
2152
+ },
2153
+ {
2154
+ "epoch": 0.92,
2155
+ "grad_norm": 1.8900826834954527,
2156
+ "learning_rate": 1.8314560692059836e-07,
2157
+ "loss": 0.8701,
2158
+ "step": 307
2159
+ },
2160
+ {
2161
+ "epoch": 0.92,
2162
+ "grad_norm": 1.986051932218701,
2163
+ "learning_rate": 1.7037086855465902e-07,
2164
+ "loss": 0.854,
2165
+ "step": 308
2166
+ },
2167
+ {
2168
+ "epoch": 0.92,
2169
+ "grad_norm": 1.819134244934257,
2170
+ "learning_rate": 1.5805019736097105e-07,
2171
+ "loss": 0.8146,
2172
+ "step": 309
2173
+ },
2174
+ {
2175
+ "epoch": 0.93,
2176
+ "grad_norm": 1.7730810336503418,
2177
+ "learning_rate": 1.4618475169190017e-07,
2178
+ "loss": 0.8358,
2179
+ "step": 310
2180
+ },
2181
+ {
2182
+ "epoch": 0.93,
2183
+ "grad_norm": 1.8749266769059667,
2184
+ "learning_rate": 1.3477564710088097e-07,
2185
+ "loss": 0.8806,
2186
+ "step": 311
2187
+ },
2188
+ {
2189
+ "epoch": 0.93,
2190
+ "grad_norm": 1.7637217906423568,
2191
+ "learning_rate": 1.2382395623753484e-07,
2192
+ "loss": 0.8348,
2193
+ "step": 312
2194
+ },
2195
+ {
2196
+ "epoch": 0.93,
2197
+ "grad_norm": 1.7705930846058242,
2198
+ "learning_rate": 1.1333070874682217e-07,
2199
+ "loss": 0.8438,
2200
+ "step": 313
2201
+ },
2202
+ {
2203
+ "epoch": 0.94,
2204
+ "grad_norm": 1.9290513445644262,
2205
+ "learning_rate": 1.0329689117224262e-07,
2206
+ "loss": 0.8837,
2207
+ "step": 314
2208
+ },
2209
+ {
2210
+ "epoch": 0.94,
2211
+ "grad_norm": 1.7184442508285342,
2212
+ "learning_rate": 9.372344686307655e-08,
2213
+ "loss": 0.7542,
2214
+ "step": 315
2215
+ },
2216
+ {
2217
+ "epoch": 0.94,
2218
+ "grad_norm": 1.7508169719519233,
2219
+ "learning_rate": 8.461127588570039e-08,
2220
+ "loss": 0.8068,
2221
+ "step": 316
2222
+ },
2223
+ {
2224
+ "epoch": 0.95,
2225
+ "grad_norm": 2.072932043609583,
2226
+ "learning_rate": 7.59612349389599e-08,
2227
+ "loss": 0.8847,
2228
+ "step": 317
2229
+ },
2230
+ {
2231
+ "epoch": 0.95,
2232
+ "grad_norm": 1.8418737822218936,
2233
+ "learning_rate": 6.777413727363069e-08,
2234
+ "loss": 0.7607,
2235
+ "step": 318
2236
+ },
2237
+ {
2238
+ "epoch": 0.95,
2239
+ "grad_norm": 1.8728345287191652,
2240
+ "learning_rate": 6.005075261595495e-08,
2241
+ "loss": 0.8279,
2242
+ "step": 319
2243
+ },
2244
+ {
2245
+ "epoch": 0.96,
2246
+ "grad_norm": 1.9751421080539375,
2247
+ "learning_rate": 5.279180709527765e-08,
2248
+ "loss": 0.7666,
2249
+ "step": 320
2250
+ },
2251
+ {
2252
+ "epoch": 0.96,
2253
+ "grad_norm": 1.742439764943798,
2254
+ "learning_rate": 4.599798317577342e-08,
2255
+ "loss": 0.8479,
2256
+ "step": 321
2257
+ },
2258
+ {
2259
+ "epoch": 0.96,
2260
+ "grad_norm": 1.8692495061239316,
2261
+ "learning_rate": 3.9669919592288385e-08,
2262
+ "loss": 0.8109,
2263
+ "step": 322
2264
+ },
2265
+ {
2266
+ "epoch": 0.96,
2267
+ "grad_norm": 1.670765464035315,
2268
+ "learning_rate": 3.3808211290284886e-08,
2269
+ "loss": 0.7585,
2270
+ "step": 323
2271
+ },
2272
+ {
2273
+ "epoch": 0.97,
2274
+ "grad_norm": 2.062997844606162,
2275
+ "learning_rate": 2.8413409369907887e-08,
2276
+ "loss": 0.8709,
2277
+ "step": 324
2278
+ },
2279
+ {
2280
+ "epoch": 0.97,
2281
+ "grad_norm": 1.9091961716040289,
2282
+ "learning_rate": 2.3486021034170857e-08,
2283
+ "loss": 0.818,
2284
+ "step": 325
2285
+ },
2286
+ {
2287
+ "epoch": 0.97,
2288
+ "grad_norm": 1.8898711790354914,
2289
+ "learning_rate": 1.9026509541272276e-08,
2290
+ "loss": 0.8346,
2291
+ "step": 326
2292
+ },
2293
+ {
2294
+ "epoch": 0.98,
2295
+ "grad_norm": 1.6576693409221155,
2296
+ "learning_rate": 1.5035294161039882e-08,
2297
+ "loss": 0.7996,
2298
+ "step": 327
2299
+ },
2300
+ {
2301
+ "epoch": 0.98,
2302
+ "grad_norm": 1.7571989556959642,
2303
+ "learning_rate": 1.1512750135511674e-08,
2304
+ "loss": 0.7742,
2305
+ "step": 328
2306
+ },
2307
+ {
2308
+ "epoch": 0.98,
2309
+ "grad_norm": 1.8431201117480818,
2310
+ "learning_rate": 8.459208643659122e-09,
2311
+ "loss": 0.8858,
2312
+ "step": 329
2313
+ },
2314
+ {
2315
+ "epoch": 0.99,
2316
+ "grad_norm": 1.8220210754565818,
2317
+ "learning_rate": 5.874956770248186e-09,
2318
+ "loss": 0.8481,
2319
+ "step": 330
2320
+ },
2321
+ {
2322
+ "epoch": 0.99,
2323
+ "grad_norm": 1.8136218315284558,
2324
+ "learning_rate": 3.760237478849793e-09,
2325
+ "loss": 0.6899,
2326
+ "step": 331
2327
+ },
2328
+ {
2329
+ "epoch": 0.99,
2330
+ "grad_norm": 1.9765791609593397,
2331
+ "learning_rate": 2.1152495889970035e-09,
2332
+ "loss": 0.8208,
2333
+ "step": 332
2334
+ },
2335
+ {
2336
+ "epoch": 0.99,
2337
+ "grad_norm": 1.8275429866969195,
2338
+ "learning_rate": 9.401477574932927e-10,
2339
+ "loss": 0.8589,
2340
+ "step": 333
2341
+ },
2342
+ {
2343
+ "epoch": 1.0,
2344
+ "grad_norm": 1.8063515591696528,
2345
+ "learning_rate": 2.3504246386918394e-10,
2346
+ "loss": 0.781,
2347
+ "step": 334
2348
+ },
2349
+ {
2350
+ "epoch": 1.0,
2351
+ "grad_norm": 1.9072496888398838,
2352
+ "learning_rate": 0.0,
2353
+ "loss": 0.8156,
2354
+ "step": 335
2355
+ },
2356
+ {
2357
+ "epoch": 1.0,
2358
+ "step": 335,
2359
+ "total_flos": 79804470312960.0,
2360
+ "train_loss": 0.9642956034461064,
2361
+ "train_runtime": 7158.0735,
2362
+ "train_samples_per_second": 0.374,
2363
+ "train_steps_per_second": 0.047
2364
+ }
2365
+ ],
2366
+ "logging_steps": 1.0,
2367
+ "max_steps": 335,
2368
+ "num_input_tokens_seen": 0,
2369
+ "num_train_epochs": 1,
2370
+ "save_steps": 2056,
2371
+ "total_flos": 79804470312960.0,
2372
+ "train_batch_size": 1,
2373
+ "trial_name": null,
2374
+ "trial_params": null
2375
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eca46937d105715c6d1755800fc6e41b4316ee4d04c5561779d17e5e9666495c
3
+ size 7864
vocab.json ADDED
The diff for this file is too large to render. See raw diff