Upload folder using huggingface_hub
Browse files- checkpoint-420/config.json +66 -0
- checkpoint-420/generation_config.json +249 -0
- checkpoint-420/model.safetensors +3 -0
- checkpoint-420/optimizer.pt +3 -0
- checkpoint-420/preprocessor_config.json +14 -0
- checkpoint-420/rng_state.pth +3 -0
- checkpoint-420/scheduler.pt +3 -0
- checkpoint-420/trainer_state.json +1281 -0
- checkpoint-420/training_args.bin +3 -0
- checkpoint-840/config.json +66 -0
- checkpoint-840/generation_config.json +249 -0
- checkpoint-840/model.safetensors +3 -0
- checkpoint-840/optimizer.pt +3 -0
- checkpoint-840/preprocessor_config.json +14 -0
- checkpoint-840/rng_state.pth +3 -0
- checkpoint-840/scheduler.pt +3 -0
- checkpoint-840/trainer_state.json +1281 -0
- checkpoint-840/training_args.bin +3 -0
checkpoint-420/config.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-medium",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"apply_spec_augment": false,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"WhisperForConditionalGeneration"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"begin_suppress_tokens": [
|
| 11 |
+
220,
|
| 12 |
+
50257
|
| 13 |
+
],
|
| 14 |
+
"bos_token_id": 50257,
|
| 15 |
+
"classifier_proj_size": 256,
|
| 16 |
+
"d_model": 1024,
|
| 17 |
+
"decoder_attention_heads": 16,
|
| 18 |
+
"decoder_ffn_dim": 4096,
|
| 19 |
+
"decoder_layerdrop": 0.0,
|
| 20 |
+
"decoder_layers": 24,
|
| 21 |
+
"decoder_start_token_id": 50258,
|
| 22 |
+
"dropout": 0.0,
|
| 23 |
+
"encoder_attention_heads": 16,
|
| 24 |
+
"encoder_ffn_dim": 4096,
|
| 25 |
+
"encoder_layerdrop": 0.0,
|
| 26 |
+
"encoder_layers": 24,
|
| 27 |
+
"eos_token_id": 50257,
|
| 28 |
+
"forced_decoder_ids": [
|
| 29 |
+
[
|
| 30 |
+
1,
|
| 31 |
+
50259
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
2,
|
| 35 |
+
50359
|
| 36 |
+
],
|
| 37 |
+
[
|
| 38 |
+
3,
|
| 39 |
+
50363
|
| 40 |
+
]
|
| 41 |
+
],
|
| 42 |
+
"init_std": 0.02,
|
| 43 |
+
"is_encoder_decoder": true,
|
| 44 |
+
"language": "English",
|
| 45 |
+
"mask_feature_length": 10,
|
| 46 |
+
"mask_feature_min_masks": 0,
|
| 47 |
+
"mask_feature_prob": 0.0,
|
| 48 |
+
"mask_time_length": 10,
|
| 49 |
+
"mask_time_min_masks": 2,
|
| 50 |
+
"mask_time_prob": 0.05,
|
| 51 |
+
"max_length": 448,
|
| 52 |
+
"max_source_positions": 1500,
|
| 53 |
+
"max_target_positions": 448,
|
| 54 |
+
"median_filter_width": 7,
|
| 55 |
+
"model_type": "whisper",
|
| 56 |
+
"num_hidden_layers": 24,
|
| 57 |
+
"num_mel_bins": 80,
|
| 58 |
+
"pad_token_id": 50257,
|
| 59 |
+
"scale_embedding": false,
|
| 60 |
+
"suppress_tokens": [],
|
| 61 |
+
"torch_dtype": "float32",
|
| 62 |
+
"transformers_version": "4.37.2",
|
| 63 |
+
"use_cache": true,
|
| 64 |
+
"use_weighted_layer_sum": false,
|
| 65 |
+
"vocab_size": 51865
|
| 66 |
+
}
|
checkpoint-420/generation_config.json
ADDED
|
@@ -0,0 +1,249 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alignment_heads": [
|
| 3 |
+
[
|
| 4 |
+
13,
|
| 5 |
+
15
|
| 6 |
+
],
|
| 7 |
+
[
|
| 8 |
+
15,
|
| 9 |
+
4
|
| 10 |
+
],
|
| 11 |
+
[
|
| 12 |
+
15,
|
| 13 |
+
15
|
| 14 |
+
],
|
| 15 |
+
[
|
| 16 |
+
16,
|
| 17 |
+
1
|
| 18 |
+
],
|
| 19 |
+
[
|
| 20 |
+
20,
|
| 21 |
+
0
|
| 22 |
+
],
|
| 23 |
+
[
|
| 24 |
+
23,
|
| 25 |
+
4
|
| 26 |
+
]
|
| 27 |
+
],
|
| 28 |
+
"begin_suppress_tokens": [
|
| 29 |
+
220,
|
| 30 |
+
50257
|
| 31 |
+
],
|
| 32 |
+
"bos_token_id": 50257,
|
| 33 |
+
"decoder_start_token_id": 50258,
|
| 34 |
+
"eos_token_id": 50257,
|
| 35 |
+
"forced_decoder_ids": [
|
| 36 |
+
[
|
| 37 |
+
1,
|
| 38 |
+
null
|
| 39 |
+
],
|
| 40 |
+
[
|
| 41 |
+
2,
|
| 42 |
+
50359
|
| 43 |
+
]
|
| 44 |
+
],
|
| 45 |
+
"is_multilingual": true,
|
| 46 |
+
"lang_to_id": {
|
| 47 |
+
"<|af|>": 50327,
|
| 48 |
+
"<|am|>": 50334,
|
| 49 |
+
"<|ar|>": 50272,
|
| 50 |
+
"<|as|>": 50350,
|
| 51 |
+
"<|az|>": 50304,
|
| 52 |
+
"<|ba|>": 50355,
|
| 53 |
+
"<|be|>": 50330,
|
| 54 |
+
"<|bg|>": 50292,
|
| 55 |
+
"<|bn|>": 50302,
|
| 56 |
+
"<|bo|>": 50347,
|
| 57 |
+
"<|br|>": 50309,
|
| 58 |
+
"<|bs|>": 50315,
|
| 59 |
+
"<|ca|>": 50270,
|
| 60 |
+
"<|cs|>": 50283,
|
| 61 |
+
"<|cy|>": 50297,
|
| 62 |
+
"<|da|>": 50285,
|
| 63 |
+
"<|de|>": 50261,
|
| 64 |
+
"<|el|>": 50281,
|
| 65 |
+
"<|en|>": 50259,
|
| 66 |
+
"<|es|>": 50262,
|
| 67 |
+
"<|et|>": 50307,
|
| 68 |
+
"<|eu|>": 50310,
|
| 69 |
+
"<|fa|>": 50300,
|
| 70 |
+
"<|fi|>": 50277,
|
| 71 |
+
"<|fo|>": 50338,
|
| 72 |
+
"<|fr|>": 50265,
|
| 73 |
+
"<|gl|>": 50319,
|
| 74 |
+
"<|gu|>": 50333,
|
| 75 |
+
"<|haw|>": 50352,
|
| 76 |
+
"<|ha|>": 50354,
|
| 77 |
+
"<|he|>": 50279,
|
| 78 |
+
"<|hi|>": 50276,
|
| 79 |
+
"<|hr|>": 50291,
|
| 80 |
+
"<|ht|>": 50339,
|
| 81 |
+
"<|hu|>": 50286,
|
| 82 |
+
"<|hy|>": 50312,
|
| 83 |
+
"<|id|>": 50275,
|
| 84 |
+
"<|is|>": 50311,
|
| 85 |
+
"<|it|>": 50274,
|
| 86 |
+
"<|ja|>": 50266,
|
| 87 |
+
"<|jw|>": 50356,
|
| 88 |
+
"<|ka|>": 50329,
|
| 89 |
+
"<|kk|>": 50316,
|
| 90 |
+
"<|km|>": 50323,
|
| 91 |
+
"<|kn|>": 50306,
|
| 92 |
+
"<|ko|>": 50264,
|
| 93 |
+
"<|la|>": 50294,
|
| 94 |
+
"<|lb|>": 50345,
|
| 95 |
+
"<|ln|>": 50353,
|
| 96 |
+
"<|lo|>": 50336,
|
| 97 |
+
"<|lt|>": 50293,
|
| 98 |
+
"<|lv|>": 50301,
|
| 99 |
+
"<|mg|>": 50349,
|
| 100 |
+
"<|mi|>": 50295,
|
| 101 |
+
"<|mk|>": 50308,
|
| 102 |
+
"<|ml|>": 50296,
|
| 103 |
+
"<|mn|>": 50314,
|
| 104 |
+
"<|mr|>": 50320,
|
| 105 |
+
"<|ms|>": 50282,
|
| 106 |
+
"<|mt|>": 50343,
|
| 107 |
+
"<|my|>": 50346,
|
| 108 |
+
"<|ne|>": 50313,
|
| 109 |
+
"<|nl|>": 50271,
|
| 110 |
+
"<|nn|>": 50342,
|
| 111 |
+
"<|no|>": 50288,
|
| 112 |
+
"<|oc|>": 50328,
|
| 113 |
+
"<|pa|>": 50321,
|
| 114 |
+
"<|pl|>": 50269,
|
| 115 |
+
"<|ps|>": 50340,
|
| 116 |
+
"<|pt|>": 50267,
|
| 117 |
+
"<|ro|>": 50284,
|
| 118 |
+
"<|ru|>": 50263,
|
| 119 |
+
"<|sa|>": 50344,
|
| 120 |
+
"<|sd|>": 50332,
|
| 121 |
+
"<|si|>": 50322,
|
| 122 |
+
"<|sk|>": 50298,
|
| 123 |
+
"<|sl|>": 50305,
|
| 124 |
+
"<|sn|>": 50324,
|
| 125 |
+
"<|so|>": 50326,
|
| 126 |
+
"<|sq|>": 50317,
|
| 127 |
+
"<|sr|>": 50303,
|
| 128 |
+
"<|su|>": 50357,
|
| 129 |
+
"<|sv|>": 50273,
|
| 130 |
+
"<|sw|>": 50318,
|
| 131 |
+
"<|ta|>": 50287,
|
| 132 |
+
"<|te|>": 50299,
|
| 133 |
+
"<|tg|>": 50331,
|
| 134 |
+
"<|th|>": 50289,
|
| 135 |
+
"<|tk|>": 50341,
|
| 136 |
+
"<|tl|>": 50348,
|
| 137 |
+
"<|tr|>": 50268,
|
| 138 |
+
"<|tt|>": 50351,
|
| 139 |
+
"<|uk|>": 50280,
|
| 140 |
+
"<|ur|>": 50290,
|
| 141 |
+
"<|uz|>": 50337,
|
| 142 |
+
"<|vi|>": 50278,
|
| 143 |
+
"<|yi|>": 50335,
|
| 144 |
+
"<|yo|>": 50325,
|
| 145 |
+
"<|zh|>": 50260
|
| 146 |
+
},
|
| 147 |
+
"max_initial_timestamp_index": 50,
|
| 148 |
+
"max_length": 448,
|
| 149 |
+
"no_timestamps_token_id": 50363,
|
| 150 |
+
"pad_token_id": 50257,
|
| 151 |
+
"prev_sot_token_id": 50361,
|
| 152 |
+
"return_timestamps": false,
|
| 153 |
+
"suppress_tokens": [
|
| 154 |
+
1,
|
| 155 |
+
2,
|
| 156 |
+
7,
|
| 157 |
+
8,
|
| 158 |
+
9,
|
| 159 |
+
10,
|
| 160 |
+
14,
|
| 161 |
+
25,
|
| 162 |
+
26,
|
| 163 |
+
27,
|
| 164 |
+
28,
|
| 165 |
+
29,
|
| 166 |
+
31,
|
| 167 |
+
58,
|
| 168 |
+
59,
|
| 169 |
+
60,
|
| 170 |
+
61,
|
| 171 |
+
62,
|
| 172 |
+
63,
|
| 173 |
+
90,
|
| 174 |
+
91,
|
| 175 |
+
92,
|
| 176 |
+
93,
|
| 177 |
+
359,
|
| 178 |
+
503,
|
| 179 |
+
522,
|
| 180 |
+
542,
|
| 181 |
+
873,
|
| 182 |
+
893,
|
| 183 |
+
902,
|
| 184 |
+
918,
|
| 185 |
+
922,
|
| 186 |
+
931,
|
| 187 |
+
1350,
|
| 188 |
+
1853,
|
| 189 |
+
1982,
|
| 190 |
+
2460,
|
| 191 |
+
2627,
|
| 192 |
+
3246,
|
| 193 |
+
3253,
|
| 194 |
+
3268,
|
| 195 |
+
3536,
|
| 196 |
+
3846,
|
| 197 |
+
3961,
|
| 198 |
+
4183,
|
| 199 |
+
4667,
|
| 200 |
+
6585,
|
| 201 |
+
6647,
|
| 202 |
+
7273,
|
| 203 |
+
9061,
|
| 204 |
+
9383,
|
| 205 |
+
10428,
|
| 206 |
+
10929,
|
| 207 |
+
11938,
|
| 208 |
+
12033,
|
| 209 |
+
12331,
|
| 210 |
+
12562,
|
| 211 |
+
13793,
|
| 212 |
+
14157,
|
| 213 |
+
14635,
|
| 214 |
+
15265,
|
| 215 |
+
15618,
|
| 216 |
+
16553,
|
| 217 |
+
16604,
|
| 218 |
+
18362,
|
| 219 |
+
18956,
|
| 220 |
+
20075,
|
| 221 |
+
21675,
|
| 222 |
+
22520,
|
| 223 |
+
26130,
|
| 224 |
+
26161,
|
| 225 |
+
26435,
|
| 226 |
+
28279,
|
| 227 |
+
29464,
|
| 228 |
+
31650,
|
| 229 |
+
32302,
|
| 230 |
+
32470,
|
| 231 |
+
36865,
|
| 232 |
+
42863,
|
| 233 |
+
47425,
|
| 234 |
+
49870,
|
| 235 |
+
50254,
|
| 236 |
+
50258,
|
| 237 |
+
50358,
|
| 238 |
+
50359,
|
| 239 |
+
50360,
|
| 240 |
+
50361,
|
| 241 |
+
50362
|
| 242 |
+
],
|
| 243 |
+
"task": "transcribe",
|
| 244 |
+
"task_to_id": {
|
| 245 |
+
"transcribe": 50359,
|
| 246 |
+
"translate": 50358
|
| 247 |
+
},
|
| 248 |
+
"transformers_version": "4.37.2"
|
| 249 |
+
}
|
checkpoint-420/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:de63da89aced924a99e682e3342e7e259eccd9ecd5097b2a2e6900286f5258d1
|
| 3 |
+
size 3055544304
|
checkpoint-420/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:63a31617799fbf67c1c385194f90876c21c9ed0e412a45f254b791428f05d902
|
| 3 |
+
size 1531565088
|
checkpoint-420/preprocessor_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length": 30,
|
| 3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
| 4 |
+
"feature_size": 80,
|
| 5 |
+
"hop_length": 160,
|
| 6 |
+
"n_fft": 400,
|
| 7 |
+
"n_samples": 480000,
|
| 8 |
+
"nb_max_frames": 3000,
|
| 9 |
+
"padding_side": "right",
|
| 10 |
+
"padding_value": 0.0,
|
| 11 |
+
"processor_class": "WhisperProcessor",
|
| 12 |
+
"return_attention_mask": false,
|
| 13 |
+
"sampling_rate": 16000
|
| 14 |
+
}
|
checkpoint-420/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:07383d577af8e5374413ecc7766c43ad8dd3c65ff54c8a1ded0f83354021cdad
|
| 3 |
+
size 14244
|
checkpoint-420/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:829fe6630cb129cf4e080b77ee62b862042f9bfb8b551cae45d5f9db5fcf4937
|
| 3 |
+
size 1064
|
checkpoint-420/trainer_state.json
ADDED
|
@@ -0,0 +1,1281 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 420,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 0.0,
|
| 14 |
+
"loss": 5.3442,
|
| 15 |
+
"step": 2
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 0.0,
|
| 20 |
+
"loss": 5.2525,
|
| 21 |
+
"step": 4
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.01,
|
| 25 |
+
"learning_rate": 2e-08,
|
| 26 |
+
"loss": 5.0768,
|
| 27 |
+
"step": 6
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.02,
|
| 31 |
+
"learning_rate": 6e-08,
|
| 32 |
+
"loss": 5.1381,
|
| 33 |
+
"step": 8
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.02,
|
| 37 |
+
"learning_rate": 8e-08,
|
| 38 |
+
"loss": 5.3956,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.03,
|
| 43 |
+
"learning_rate": 1.2e-07,
|
| 44 |
+
"loss": 5.2933,
|
| 45 |
+
"step": 12
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.03,
|
| 49 |
+
"learning_rate": 1.6e-07,
|
| 50 |
+
"loss": 5.2991,
|
| 51 |
+
"step": 14
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.04,
|
| 55 |
+
"learning_rate": 2e-07,
|
| 56 |
+
"loss": 5.5463,
|
| 57 |
+
"step": 16
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.04,
|
| 61 |
+
"learning_rate": 2.4e-07,
|
| 62 |
+
"loss": 5.3438,
|
| 63 |
+
"step": 18
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.05,
|
| 67 |
+
"learning_rate": 2.6e-07,
|
| 68 |
+
"loss": 5.0126,
|
| 69 |
+
"step": 20
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.05,
|
| 73 |
+
"learning_rate": 3e-07,
|
| 74 |
+
"loss": 4.8808,
|
| 75 |
+
"step": 22
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.06,
|
| 79 |
+
"learning_rate": 3.4000000000000003e-07,
|
| 80 |
+
"loss": 4.5046,
|
| 81 |
+
"step": 24
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.06,
|
| 85 |
+
"learning_rate": 3.7999999999999996e-07,
|
| 86 |
+
"loss": 4.3779,
|
| 87 |
+
"step": 26
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.07,
|
| 91 |
+
"learning_rate": 4.1999999999999995e-07,
|
| 92 |
+
"loss": 4.2311,
|
| 93 |
+
"step": 28
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.07,
|
| 97 |
+
"learning_rate": 4.6e-07,
|
| 98 |
+
"loss": 4.0877,
|
| 99 |
+
"step": 30
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.08,
|
| 103 |
+
"learning_rate": 5e-07,
|
| 104 |
+
"loss": 4.0972,
|
| 105 |
+
"step": 32
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.08,
|
| 109 |
+
"learning_rate": 5.4e-07,
|
| 110 |
+
"loss": 4.1511,
|
| 111 |
+
"step": 34
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.09,
|
| 115 |
+
"learning_rate": 5.8e-07,
|
| 116 |
+
"loss": 3.9641,
|
| 117 |
+
"step": 36
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.09,
|
| 121 |
+
"learning_rate": 6.2e-07,
|
| 122 |
+
"loss": 3.8411,
|
| 123 |
+
"step": 38
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.1,
|
| 127 |
+
"learning_rate": 6.6e-07,
|
| 128 |
+
"loss": 3.5198,
|
| 129 |
+
"step": 40
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.1,
|
| 133 |
+
"learning_rate": 7e-07,
|
| 134 |
+
"loss": 3.5286,
|
| 135 |
+
"step": 42
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.1,
|
| 139 |
+
"learning_rate": 7.4e-07,
|
| 140 |
+
"loss": 3.8059,
|
| 141 |
+
"step": 44
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.11,
|
| 145 |
+
"learning_rate": 7.799999999999999e-07,
|
| 146 |
+
"loss": 3.374,
|
| 147 |
+
"step": 46
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.11,
|
| 151 |
+
"learning_rate": 8.199999999999999e-07,
|
| 152 |
+
"loss": 3.3654,
|
| 153 |
+
"step": 48
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.12,
|
| 157 |
+
"learning_rate": 8.599999999999999e-07,
|
| 158 |
+
"loss": 3.3283,
|
| 159 |
+
"step": 50
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.12,
|
| 163 |
+
"learning_rate": 9e-07,
|
| 164 |
+
"loss": 3.3164,
|
| 165 |
+
"step": 52
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.13,
|
| 169 |
+
"learning_rate": 9.399999999999999e-07,
|
| 170 |
+
"loss": 3.13,
|
| 171 |
+
"step": 54
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.13,
|
| 175 |
+
"learning_rate": 9.8e-07,
|
| 176 |
+
"loss": 2.9915,
|
| 177 |
+
"step": 56
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.14,
|
| 181 |
+
"learning_rate": 9.987341772151898e-07,
|
| 182 |
+
"loss": 2.7748,
|
| 183 |
+
"step": 58
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.14,
|
| 187 |
+
"learning_rate": 9.962025316455696e-07,
|
| 188 |
+
"loss": 2.8378,
|
| 189 |
+
"step": 60
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.15,
|
| 193 |
+
"learning_rate": 9.936708860759494e-07,
|
| 194 |
+
"loss": 2.7933,
|
| 195 |
+
"step": 62
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.15,
|
| 199 |
+
"learning_rate": 9.91139240506329e-07,
|
| 200 |
+
"loss": 2.7108,
|
| 201 |
+
"step": 64
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.16,
|
| 205 |
+
"learning_rate": 9.88607594936709e-07,
|
| 206 |
+
"loss": 2.4029,
|
| 207 |
+
"step": 66
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.16,
|
| 211 |
+
"learning_rate": 9.860759493670887e-07,
|
| 212 |
+
"loss": 2.4199,
|
| 213 |
+
"step": 68
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.17,
|
| 217 |
+
"learning_rate": 9.835443037974683e-07,
|
| 218 |
+
"loss": 2.3015,
|
| 219 |
+
"step": 70
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.17,
|
| 223 |
+
"learning_rate": 9.81012658227848e-07,
|
| 224 |
+
"loss": 2.4849,
|
| 225 |
+
"step": 72
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.18,
|
| 229 |
+
"learning_rate": 9.784810126582278e-07,
|
| 230 |
+
"loss": 2.265,
|
| 231 |
+
"step": 74
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.18,
|
| 235 |
+
"learning_rate": 9.759493670886076e-07,
|
| 236 |
+
"loss": 2.2422,
|
| 237 |
+
"step": 76
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.19,
|
| 241 |
+
"learning_rate": 9.734177215189873e-07,
|
| 242 |
+
"loss": 2.188,
|
| 243 |
+
"step": 78
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.19,
|
| 247 |
+
"learning_rate": 9.70886075949367e-07,
|
| 248 |
+
"loss": 2.1488,
|
| 249 |
+
"step": 80
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.2,
|
| 253 |
+
"learning_rate": 9.683544303797469e-07,
|
| 254 |
+
"loss": 2.0795,
|
| 255 |
+
"step": 82
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.2,
|
| 259 |
+
"learning_rate": 9.658227848101267e-07,
|
| 260 |
+
"loss": 1.9767,
|
| 261 |
+
"step": 84
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.2,
|
| 265 |
+
"learning_rate": 9.632911392405062e-07,
|
| 266 |
+
"loss": 1.9811,
|
| 267 |
+
"step": 86
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.21,
|
| 271 |
+
"learning_rate": 9.60759493670886e-07,
|
| 272 |
+
"loss": 2.0459,
|
| 273 |
+
"step": 88
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.21,
|
| 277 |
+
"learning_rate": 9.582278481012658e-07,
|
| 278 |
+
"loss": 1.936,
|
| 279 |
+
"step": 90
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.22,
|
| 283 |
+
"learning_rate": 9.556962025316455e-07,
|
| 284 |
+
"loss": 1.9373,
|
| 285 |
+
"step": 92
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.22,
|
| 289 |
+
"learning_rate": 9.531645569620253e-07,
|
| 290 |
+
"loss": 2.0272,
|
| 291 |
+
"step": 94
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.23,
|
| 295 |
+
"learning_rate": 9.50632911392405e-07,
|
| 296 |
+
"loss": 1.8032,
|
| 297 |
+
"step": 96
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.23,
|
| 301 |
+
"learning_rate": 9.481012658227847e-07,
|
| 302 |
+
"loss": 1.8799,
|
| 303 |
+
"step": 98
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.24,
|
| 307 |
+
"learning_rate": 9.455696202531646e-07,
|
| 308 |
+
"loss": 1.9145,
|
| 309 |
+
"step": 100
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.24,
|
| 313 |
+
"learning_rate": 9.430379746835443e-07,
|
| 314 |
+
"loss": 1.8476,
|
| 315 |
+
"step": 102
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.25,
|
| 319 |
+
"learning_rate": 9.40506329113924e-07,
|
| 320 |
+
"loss": 1.7027,
|
| 321 |
+
"step": 104
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.25,
|
| 325 |
+
"learning_rate": 9.379746835443037e-07,
|
| 326 |
+
"loss": 1.6791,
|
| 327 |
+
"step": 106
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.26,
|
| 331 |
+
"learning_rate": 9.354430379746835e-07,
|
| 332 |
+
"loss": 1.6089,
|
| 333 |
+
"step": 108
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.26,
|
| 337 |
+
"learning_rate": 9.329113924050633e-07,
|
| 338 |
+
"loss": 1.6315,
|
| 339 |
+
"step": 110
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.27,
|
| 343 |
+
"learning_rate": 9.303797468354429e-07,
|
| 344 |
+
"loss": 1.6194,
|
| 345 |
+
"step": 112
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.27,
|
| 349 |
+
"learning_rate": 9.278481012658227e-07,
|
| 350 |
+
"loss": 1.5367,
|
| 351 |
+
"step": 114
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.28,
|
| 355 |
+
"learning_rate": 9.253164556962026e-07,
|
| 356 |
+
"loss": 1.4617,
|
| 357 |
+
"step": 116
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.28,
|
| 361 |
+
"learning_rate": 9.227848101265822e-07,
|
| 362 |
+
"loss": 1.6019,
|
| 363 |
+
"step": 118
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.29,
|
| 367 |
+
"learning_rate": 9.20253164556962e-07,
|
| 368 |
+
"loss": 1.4418,
|
| 369 |
+
"step": 120
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.29,
|
| 373 |
+
"learning_rate": 9.177215189873418e-07,
|
| 374 |
+
"loss": 1.4306,
|
| 375 |
+
"step": 122
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.3,
|
| 379 |
+
"learning_rate": 9.151898734177214e-07,
|
| 380 |
+
"loss": 1.3711,
|
| 381 |
+
"step": 124
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.3,
|
| 385 |
+
"learning_rate": 9.126582278481012e-07,
|
| 386 |
+
"loss": 1.4102,
|
| 387 |
+
"step": 126
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3,
|
| 391 |
+
"learning_rate": 9.10126582278481e-07,
|
| 392 |
+
"loss": 1.348,
|
| 393 |
+
"step": 128
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.31,
|
| 397 |
+
"learning_rate": 9.075949367088606e-07,
|
| 398 |
+
"loss": 1.374,
|
| 399 |
+
"step": 130
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.31,
|
| 403 |
+
"learning_rate": 9.050632911392405e-07,
|
| 404 |
+
"loss": 1.2833,
|
| 405 |
+
"step": 132
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.32,
|
| 409 |
+
"learning_rate": 9.025316455696203e-07,
|
| 410 |
+
"loss": 1.2588,
|
| 411 |
+
"step": 134
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.32,
|
| 415 |
+
"learning_rate": 9e-07,
|
| 416 |
+
"loss": 1.2172,
|
| 417 |
+
"step": 136
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.33,
|
| 421 |
+
"learning_rate": 8.974683544303797e-07,
|
| 422 |
+
"loss": 1.3046,
|
| 423 |
+
"step": 138
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.33,
|
| 427 |
+
"learning_rate": 8.949367088607594e-07,
|
| 428 |
+
"loss": 1.0875,
|
| 429 |
+
"step": 140
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.34,
|
| 433 |
+
"learning_rate": 8.924050632911392e-07,
|
| 434 |
+
"loss": 1.2346,
|
| 435 |
+
"step": 142
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.34,
|
| 439 |
+
"learning_rate": 8.898734177215189e-07,
|
| 440 |
+
"loss": 1.0975,
|
| 441 |
+
"step": 144
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.35,
|
| 445 |
+
"learning_rate": 8.873417721518986e-07,
|
| 446 |
+
"loss": 1.2068,
|
| 447 |
+
"step": 146
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.35,
|
| 451 |
+
"learning_rate": 8.848101265822785e-07,
|
| 452 |
+
"loss": 1.1378,
|
| 453 |
+
"step": 148
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.36,
|
| 457 |
+
"learning_rate": 8.822784810126583e-07,
|
| 458 |
+
"loss": 1.0359,
|
| 459 |
+
"step": 150
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.36,
|
| 463 |
+
"learning_rate": 8.797468354430379e-07,
|
| 464 |
+
"loss": 1.2252,
|
| 465 |
+
"step": 152
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.37,
|
| 469 |
+
"learning_rate": 8.772151898734177e-07,
|
| 470 |
+
"loss": 1.062,
|
| 471 |
+
"step": 154
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.37,
|
| 475 |
+
"learning_rate": 8.746835443037975e-07,
|
| 476 |
+
"loss": 0.9852,
|
| 477 |
+
"step": 156
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.38,
|
| 481 |
+
"learning_rate": 8.721518987341771e-07,
|
| 482 |
+
"loss": 1.0557,
|
| 483 |
+
"step": 158
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.38,
|
| 487 |
+
"learning_rate": 8.696202531645569e-07,
|
| 488 |
+
"loss": 0.9546,
|
| 489 |
+
"step": 160
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.39,
|
| 493 |
+
"learning_rate": 8.670886075949367e-07,
|
| 494 |
+
"loss": 0.9371,
|
| 495 |
+
"step": 162
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.39,
|
| 499 |
+
"learning_rate": 8.645569620253164e-07,
|
| 500 |
+
"loss": 0.9128,
|
| 501 |
+
"step": 164
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.4,
|
| 505 |
+
"learning_rate": 8.620253164556962e-07,
|
| 506 |
+
"loss": 0.9508,
|
| 507 |
+
"step": 166
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.4,
|
| 511 |
+
"learning_rate": 8.59493670886076e-07,
|
| 512 |
+
"loss": 1.0656,
|
| 513 |
+
"step": 168
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.4,
|
| 517 |
+
"learning_rate": 8.569620253164556e-07,
|
| 518 |
+
"loss": 0.8574,
|
| 519 |
+
"step": 170
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.41,
|
| 523 |
+
"learning_rate": 8.544303797468354e-07,
|
| 524 |
+
"loss": 0.7957,
|
| 525 |
+
"step": 172
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.41,
|
| 529 |
+
"learning_rate": 8.518987341772152e-07,
|
| 530 |
+
"loss": 0.8379,
|
| 531 |
+
"step": 174
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.42,
|
| 535 |
+
"learning_rate": 8.493670886075949e-07,
|
| 536 |
+
"loss": 0.7316,
|
| 537 |
+
"step": 176
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.42,
|
| 541 |
+
"learning_rate": 8.468354430379746e-07,
|
| 542 |
+
"loss": 0.8871,
|
| 543 |
+
"step": 178
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.43,
|
| 547 |
+
"learning_rate": 8.443037974683543e-07,
|
| 548 |
+
"loss": 0.7857,
|
| 549 |
+
"step": 180
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.43,
|
| 553 |
+
"learning_rate": 8.417721518987342e-07,
|
| 554 |
+
"loss": 0.6546,
|
| 555 |
+
"step": 182
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.44,
|
| 559 |
+
"learning_rate": 8.392405063291139e-07,
|
| 560 |
+
"loss": 0.7563,
|
| 561 |
+
"step": 184
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.44,
|
| 565 |
+
"learning_rate": 8.367088607594936e-07,
|
| 566 |
+
"loss": 0.6105,
|
| 567 |
+
"step": 186
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.45,
|
| 571 |
+
"learning_rate": 8.341772151898734e-07,
|
| 572 |
+
"loss": 0.6173,
|
| 573 |
+
"step": 188
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.45,
|
| 577 |
+
"learning_rate": 8.316455696202531e-07,
|
| 578 |
+
"loss": 0.536,
|
| 579 |
+
"step": 190
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.46,
|
| 583 |
+
"learning_rate": 8.291139240506328e-07,
|
| 584 |
+
"loss": 0.4445,
|
| 585 |
+
"step": 192
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.46,
|
| 589 |
+
"learning_rate": 8.265822784810126e-07,
|
| 590 |
+
"loss": 0.377,
|
| 591 |
+
"step": 194
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.47,
|
| 595 |
+
"learning_rate": 8.240506329113924e-07,
|
| 596 |
+
"loss": 0.4066,
|
| 597 |
+
"step": 196
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.47,
|
| 601 |
+
"learning_rate": 8.215189873417721e-07,
|
| 602 |
+
"loss": 0.3306,
|
| 603 |
+
"step": 198
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.48,
|
| 607 |
+
"learning_rate": 8.189873417721519e-07,
|
| 608 |
+
"loss": 0.3437,
|
| 609 |
+
"step": 200
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.48,
|
| 613 |
+
"learning_rate": 8.164556962025317e-07,
|
| 614 |
+
"loss": 0.3812,
|
| 615 |
+
"step": 202
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.49,
|
| 619 |
+
"learning_rate": 8.139240506329113e-07,
|
| 620 |
+
"loss": 0.3661,
|
| 621 |
+
"step": 204
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.49,
|
| 625 |
+
"learning_rate": 8.113924050632911e-07,
|
| 626 |
+
"loss": 0.3579,
|
| 627 |
+
"step": 206
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.5,
|
| 631 |
+
"learning_rate": 8.088607594936709e-07,
|
| 632 |
+
"loss": 0.3228,
|
| 633 |
+
"step": 208
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.5,
|
| 637 |
+
"learning_rate": 8.063291139240505e-07,
|
| 638 |
+
"loss": 0.3723,
|
| 639 |
+
"step": 210
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5,
|
| 643 |
+
"learning_rate": 8.037974683544303e-07,
|
| 644 |
+
"loss": 0.3276,
|
| 645 |
+
"step": 212
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.51,
|
| 649 |
+
"learning_rate": 8.012658227848102e-07,
|
| 650 |
+
"loss": 0.3386,
|
| 651 |
+
"step": 214
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.51,
|
| 655 |
+
"learning_rate": 7.987341772151898e-07,
|
| 656 |
+
"loss": 0.3415,
|
| 657 |
+
"step": 216
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.52,
|
| 661 |
+
"learning_rate": 7.962025316455696e-07,
|
| 662 |
+
"loss": 0.3055,
|
| 663 |
+
"step": 218
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 0.52,
|
| 667 |
+
"learning_rate": 7.936708860759494e-07,
|
| 668 |
+
"loss": 0.2819,
|
| 669 |
+
"step": 220
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.53,
|
| 673 |
+
"learning_rate": 7.911392405063291e-07,
|
| 674 |
+
"loss": 0.3376,
|
| 675 |
+
"step": 222
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.53,
|
| 679 |
+
"learning_rate": 7.886075949367088e-07,
|
| 680 |
+
"loss": 0.306,
|
| 681 |
+
"step": 224
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.54,
|
| 685 |
+
"learning_rate": 7.860759493670885e-07,
|
| 686 |
+
"loss": 0.2942,
|
| 687 |
+
"step": 226
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.54,
|
| 691 |
+
"learning_rate": 7.835443037974683e-07,
|
| 692 |
+
"loss": 0.3164,
|
| 693 |
+
"step": 228
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.55,
|
| 697 |
+
"learning_rate": 7.810126582278481e-07,
|
| 698 |
+
"loss": 0.2918,
|
| 699 |
+
"step": 230
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.55,
|
| 703 |
+
"learning_rate": 7.784810126582278e-07,
|
| 704 |
+
"loss": 0.3092,
|
| 705 |
+
"step": 232
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 0.56,
|
| 709 |
+
"learning_rate": 7.759493670886076e-07,
|
| 710 |
+
"loss": 0.4132,
|
| 711 |
+
"step": 234
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.56,
|
| 715 |
+
"learning_rate": 7.734177215189873e-07,
|
| 716 |
+
"loss": 0.3428,
|
| 717 |
+
"step": 236
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.57,
|
| 721 |
+
"learning_rate": 7.70886075949367e-07,
|
| 722 |
+
"loss": 0.2002,
|
| 723 |
+
"step": 238
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.57,
|
| 727 |
+
"learning_rate": 7.683544303797468e-07,
|
| 728 |
+
"loss": 0.333,
|
| 729 |
+
"step": 240
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.58,
|
| 733 |
+
"learning_rate": 7.658227848101266e-07,
|
| 734 |
+
"loss": 0.3346,
|
| 735 |
+
"step": 242
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.58,
|
| 739 |
+
"learning_rate": 7.632911392405062e-07,
|
| 740 |
+
"loss": 0.282,
|
| 741 |
+
"step": 244
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.59,
|
| 745 |
+
"learning_rate": 7.60759493670886e-07,
|
| 746 |
+
"loss": 0.1892,
|
| 747 |
+
"step": 246
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.59,
|
| 751 |
+
"learning_rate": 7.582278481012659e-07,
|
| 752 |
+
"loss": 0.2727,
|
| 753 |
+
"step": 248
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.6,
|
| 757 |
+
"learning_rate": 7.556962025316455e-07,
|
| 758 |
+
"loss": 0.3528,
|
| 759 |
+
"step": 250
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.6,
|
| 763 |
+
"learning_rate": 7.531645569620253e-07,
|
| 764 |
+
"loss": 0.3449,
|
| 765 |
+
"step": 252
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.6,
|
| 769 |
+
"learning_rate": 7.506329113924051e-07,
|
| 770 |
+
"loss": 0.2044,
|
| 771 |
+
"step": 254
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 0.61,
|
| 775 |
+
"learning_rate": 7.481012658227847e-07,
|
| 776 |
+
"loss": 0.2418,
|
| 777 |
+
"step": 256
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.61,
|
| 781 |
+
"learning_rate": 7.455696202531645e-07,
|
| 782 |
+
"loss": 0.2712,
|
| 783 |
+
"step": 258
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.62,
|
| 787 |
+
"learning_rate": 7.430379746835443e-07,
|
| 788 |
+
"loss": 0.2555,
|
| 789 |
+
"step": 260
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.62,
|
| 793 |
+
"learning_rate": 7.40506329113924e-07,
|
| 794 |
+
"loss": 0.231,
|
| 795 |
+
"step": 262
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.63,
|
| 799 |
+
"learning_rate": 7.379746835443038e-07,
|
| 800 |
+
"loss": 0.2705,
|
| 801 |
+
"step": 264
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.63,
|
| 805 |
+
"learning_rate": 7.354430379746836e-07,
|
| 806 |
+
"loss": 0.4189,
|
| 807 |
+
"step": 266
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.64,
|
| 811 |
+
"learning_rate": 7.329113924050633e-07,
|
| 812 |
+
"loss": 0.2864,
|
| 813 |
+
"step": 268
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.64,
|
| 817 |
+
"learning_rate": 7.30379746835443e-07,
|
| 818 |
+
"loss": 0.278,
|
| 819 |
+
"step": 270
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.65,
|
| 823 |
+
"learning_rate": 7.278481012658227e-07,
|
| 824 |
+
"loss": 0.308,
|
| 825 |
+
"step": 272
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.65,
|
| 829 |
+
"learning_rate": 7.253164556962025e-07,
|
| 830 |
+
"loss": 0.3659,
|
| 831 |
+
"step": 274
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.66,
|
| 835 |
+
"learning_rate": 7.227848101265822e-07,
|
| 836 |
+
"loss": 0.2763,
|
| 837 |
+
"step": 276
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.66,
|
| 841 |
+
"learning_rate": 7.202531645569619e-07,
|
| 842 |
+
"loss": 0.3028,
|
| 843 |
+
"step": 278
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.67,
|
| 847 |
+
"learning_rate": 7.177215189873418e-07,
|
| 848 |
+
"loss": 0.2959,
|
| 849 |
+
"step": 280
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.67,
|
| 853 |
+
"learning_rate": 7.151898734177216e-07,
|
| 854 |
+
"loss": 0.2135,
|
| 855 |
+
"step": 282
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 0.68,
|
| 859 |
+
"learning_rate": 7.126582278481012e-07,
|
| 860 |
+
"loss": 0.2805,
|
| 861 |
+
"step": 284
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 0.68,
|
| 865 |
+
"learning_rate": 7.10126582278481e-07,
|
| 866 |
+
"loss": 0.2925,
|
| 867 |
+
"step": 286
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.69,
|
| 871 |
+
"learning_rate": 7.075949367088608e-07,
|
| 872 |
+
"loss": 0.2914,
|
| 873 |
+
"step": 288
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.69,
|
| 877 |
+
"learning_rate": 7.050632911392404e-07,
|
| 878 |
+
"loss": 0.2366,
|
| 879 |
+
"step": 290
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.7,
|
| 883 |
+
"learning_rate": 7.025316455696202e-07,
|
| 884 |
+
"loss": 0.2246,
|
| 885 |
+
"step": 292
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.7,
|
| 889 |
+
"learning_rate": 7e-07,
|
| 890 |
+
"loss": 0.3242,
|
| 891 |
+
"step": 294
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.7,
|
| 895 |
+
"learning_rate": 6.974683544303797e-07,
|
| 896 |
+
"loss": 0.3683,
|
| 897 |
+
"step": 296
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 0.71,
|
| 901 |
+
"learning_rate": 6.949367088607595e-07,
|
| 902 |
+
"loss": 0.2697,
|
| 903 |
+
"step": 298
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.71,
|
| 907 |
+
"learning_rate": 6.924050632911393e-07,
|
| 908 |
+
"loss": 0.2314,
|
| 909 |
+
"step": 300
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.72,
|
| 913 |
+
"learning_rate": 6.898734177215189e-07,
|
| 914 |
+
"loss": 0.268,
|
| 915 |
+
"step": 302
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.72,
|
| 919 |
+
"learning_rate": 6.873417721518987e-07,
|
| 920 |
+
"loss": 0.3017,
|
| 921 |
+
"step": 304
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.73,
|
| 925 |
+
"learning_rate": 6.848101265822784e-07,
|
| 926 |
+
"loss": 0.2147,
|
| 927 |
+
"step": 306
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.73,
|
| 931 |
+
"learning_rate": 6.822784810126582e-07,
|
| 932 |
+
"loss": 0.2574,
|
| 933 |
+
"step": 308
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.74,
|
| 937 |
+
"learning_rate": 6.797468354430379e-07,
|
| 938 |
+
"loss": 0.2828,
|
| 939 |
+
"step": 310
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 0.74,
|
| 943 |
+
"learning_rate": 6.772151898734176e-07,
|
| 944 |
+
"loss": 0.2967,
|
| 945 |
+
"step": 312
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 0.75,
|
| 949 |
+
"learning_rate": 6.746835443037975e-07,
|
| 950 |
+
"loss": 0.2444,
|
| 951 |
+
"step": 314
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 0.75,
|
| 955 |
+
"learning_rate": 6.721518987341772e-07,
|
| 956 |
+
"loss": 0.2925,
|
| 957 |
+
"step": 316
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.76,
|
| 961 |
+
"learning_rate": 6.696202531645569e-07,
|
| 962 |
+
"loss": 0.243,
|
| 963 |
+
"step": 318
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 0.76,
|
| 967 |
+
"learning_rate": 6.670886075949367e-07,
|
| 968 |
+
"loss": 0.2559,
|
| 969 |
+
"step": 320
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.77,
|
| 973 |
+
"learning_rate": 6.645569620253164e-07,
|
| 974 |
+
"loss": 0.2143,
|
| 975 |
+
"step": 322
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.77,
|
| 979 |
+
"learning_rate": 6.620253164556961e-07,
|
| 980 |
+
"loss": 0.2637,
|
| 981 |
+
"step": 324
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 0.78,
|
| 985 |
+
"learning_rate": 6.594936708860759e-07,
|
| 986 |
+
"loss": 0.2977,
|
| 987 |
+
"step": 326
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 0.78,
|
| 991 |
+
"learning_rate": 6.569620253164557e-07,
|
| 992 |
+
"loss": 0.3284,
|
| 993 |
+
"step": 328
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 0.79,
|
| 997 |
+
"learning_rate": 6.544303797468354e-07,
|
| 998 |
+
"loss": 0.2531,
|
| 999 |
+
"step": 330
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.79,
|
| 1003 |
+
"learning_rate": 6.518987341772152e-07,
|
| 1004 |
+
"loss": 0.2793,
|
| 1005 |
+
"step": 332
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 0.8,
|
| 1009 |
+
"learning_rate": 6.49367088607595e-07,
|
| 1010 |
+
"loss": 0.3183,
|
| 1011 |
+
"step": 334
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.8,
|
| 1015 |
+
"learning_rate": 6.468354430379746e-07,
|
| 1016 |
+
"loss": 0.3329,
|
| 1017 |
+
"step": 336
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.8,
|
| 1021 |
+
"learning_rate": 6.443037974683544e-07,
|
| 1022 |
+
"loss": 0.2545,
|
| 1023 |
+
"step": 338
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 0.81,
|
| 1027 |
+
"learning_rate": 6.417721518987342e-07,
|
| 1028 |
+
"loss": 0.286,
|
| 1029 |
+
"step": 340
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.81,
|
| 1033 |
+
"learning_rate": 6.392405063291138e-07,
|
| 1034 |
+
"loss": 0.3286,
|
| 1035 |
+
"step": 342
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.82,
|
| 1039 |
+
"learning_rate": 6.367088607594936e-07,
|
| 1040 |
+
"loss": 0.2061,
|
| 1041 |
+
"step": 344
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.82,
|
| 1045 |
+
"learning_rate": 6.341772151898735e-07,
|
| 1046 |
+
"loss": 0.1816,
|
| 1047 |
+
"step": 346
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 0.83,
|
| 1051 |
+
"learning_rate": 6.316455696202532e-07,
|
| 1052 |
+
"loss": 0.262,
|
| 1053 |
+
"step": 348
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.83,
|
| 1057 |
+
"learning_rate": 6.291139240506329e-07,
|
| 1058 |
+
"loss": 0.3591,
|
| 1059 |
+
"step": 350
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.84,
|
| 1063 |
+
"learning_rate": 6.265822784810126e-07,
|
| 1064 |
+
"loss": 0.2565,
|
| 1065 |
+
"step": 352
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 0.84,
|
| 1069 |
+
"learning_rate": 6.240506329113924e-07,
|
| 1070 |
+
"loss": 0.2609,
|
| 1071 |
+
"step": 354
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.85,
|
| 1075 |
+
"learning_rate": 6.215189873417721e-07,
|
| 1076 |
+
"loss": 0.2728,
|
| 1077 |
+
"step": 356
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 0.85,
|
| 1081 |
+
"learning_rate": 6.189873417721518e-07,
|
| 1082 |
+
"loss": 0.3175,
|
| 1083 |
+
"step": 358
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.86,
|
| 1087 |
+
"learning_rate": 6.164556962025316e-07,
|
| 1088 |
+
"loss": 0.393,
|
| 1089 |
+
"step": 360
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.86,
|
| 1093 |
+
"learning_rate": 6.139240506329114e-07,
|
| 1094 |
+
"loss": 0.234,
|
| 1095 |
+
"step": 362
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.87,
|
| 1099 |
+
"learning_rate": 6.113924050632911e-07,
|
| 1100 |
+
"loss": 0.2264,
|
| 1101 |
+
"step": 364
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.87,
|
| 1105 |
+
"learning_rate": 6.088607594936709e-07,
|
| 1106 |
+
"loss": 0.3062,
|
| 1107 |
+
"step": 366
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 0.88,
|
| 1111 |
+
"learning_rate": 6.063291139240507e-07,
|
| 1112 |
+
"loss": 0.184,
|
| 1113 |
+
"step": 368
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 0.88,
|
| 1117 |
+
"learning_rate": 6.037974683544303e-07,
|
| 1118 |
+
"loss": 0.2396,
|
| 1119 |
+
"step": 370
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.89,
|
| 1123 |
+
"learning_rate": 6.012658227848101e-07,
|
| 1124 |
+
"loss": 0.2631,
|
| 1125 |
+
"step": 372
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.89,
|
| 1129 |
+
"learning_rate": 5.987341772151899e-07,
|
| 1130 |
+
"loss": 0.2776,
|
| 1131 |
+
"step": 374
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.9,
|
| 1135 |
+
"learning_rate": 5.962025316455695e-07,
|
| 1136 |
+
"loss": 0.2634,
|
| 1137 |
+
"step": 376
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 0.9,
|
| 1141 |
+
"learning_rate": 5.936708860759494e-07,
|
| 1142 |
+
"loss": 0.1984,
|
| 1143 |
+
"step": 378
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.9,
|
| 1147 |
+
"learning_rate": 5.911392405063292e-07,
|
| 1148 |
+
"loss": 0.2973,
|
| 1149 |
+
"step": 380
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 0.91,
|
| 1153 |
+
"learning_rate": 5.886075949367088e-07,
|
| 1154 |
+
"loss": 0.3261,
|
| 1155 |
+
"step": 382
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.91,
|
| 1159 |
+
"learning_rate": 5.860759493670886e-07,
|
| 1160 |
+
"loss": 0.1961,
|
| 1161 |
+
"step": 384
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.92,
|
| 1165 |
+
"learning_rate": 5.835443037974684e-07,
|
| 1166 |
+
"loss": 0.3002,
|
| 1167 |
+
"step": 386
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 0.92,
|
| 1171 |
+
"learning_rate": 5.81012658227848e-07,
|
| 1172 |
+
"loss": 0.2281,
|
| 1173 |
+
"step": 388
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 0.93,
|
| 1177 |
+
"learning_rate": 5.784810126582278e-07,
|
| 1178 |
+
"loss": 0.1912,
|
| 1179 |
+
"step": 390
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.93,
|
| 1183 |
+
"learning_rate": 5.759493670886075e-07,
|
| 1184 |
+
"loss": 0.2726,
|
| 1185 |
+
"step": 392
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.94,
|
| 1189 |
+
"learning_rate": 5.734177215189873e-07,
|
| 1190 |
+
"loss": 0.2855,
|
| 1191 |
+
"step": 394
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 0.94,
|
| 1195 |
+
"learning_rate": 5.708860759493671e-07,
|
| 1196 |
+
"loss": 0.2241,
|
| 1197 |
+
"step": 396
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.95,
|
| 1201 |
+
"learning_rate": 5.683544303797468e-07,
|
| 1202 |
+
"loss": 0.3273,
|
| 1203 |
+
"step": 398
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.95,
|
| 1207 |
+
"learning_rate": 5.658227848101266e-07,
|
| 1208 |
+
"loss": 0.209,
|
| 1209 |
+
"step": 400
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 0.96,
|
| 1213 |
+
"learning_rate": 5.632911392405063e-07,
|
| 1214 |
+
"loss": 0.2155,
|
| 1215 |
+
"step": 402
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 0.96,
|
| 1219 |
+
"learning_rate": 5.60759493670886e-07,
|
| 1220 |
+
"loss": 0.2218,
|
| 1221 |
+
"step": 404
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.97,
|
| 1225 |
+
"learning_rate": 5.582278481012658e-07,
|
| 1226 |
+
"loss": 0.3097,
|
| 1227 |
+
"step": 406
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.97,
|
| 1231 |
+
"learning_rate": 5.556962025316455e-07,
|
| 1232 |
+
"loss": 0.1995,
|
| 1233 |
+
"step": 408
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 0.98,
|
| 1237 |
+
"learning_rate": 5.531645569620252e-07,
|
| 1238 |
+
"loss": 0.2408,
|
| 1239 |
+
"step": 410
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 0.98,
|
| 1243 |
+
"learning_rate": 5.506329113924051e-07,
|
| 1244 |
+
"loss": 0.2443,
|
| 1245 |
+
"step": 412
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.99,
|
| 1249 |
+
"learning_rate": 5.481012658227849e-07,
|
| 1250 |
+
"loss": 0.2719,
|
| 1251 |
+
"step": 414
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 0.99,
|
| 1255 |
+
"learning_rate": 5.455696202531645e-07,
|
| 1256 |
+
"loss": 0.1995,
|
| 1257 |
+
"step": 416
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 1.0,
|
| 1261 |
+
"learning_rate": 5.430379746835443e-07,
|
| 1262 |
+
"loss": 0.2849,
|
| 1263 |
+
"step": 418
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.0,
|
| 1267 |
+
"learning_rate": 5.405063291139241e-07,
|
| 1268 |
+
"loss": 0.272,
|
| 1269 |
+
"step": 420
|
| 1270 |
+
}
|
| 1271 |
+
],
|
| 1272 |
+
"logging_steps": 2,
|
| 1273 |
+
"max_steps": 840,
|
| 1274 |
+
"num_input_tokens_seen": 0,
|
| 1275 |
+
"num_train_epochs": 2,
|
| 1276 |
+
"save_steps": 200,
|
| 1277 |
+
"total_flos": 1.371080632762368e+19,
|
| 1278 |
+
"train_batch_size": 32,
|
| 1279 |
+
"trial_name": null,
|
| 1280 |
+
"trial_params": null
|
| 1281 |
+
}
|
checkpoint-420/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ff97f6d90f494edb91f0da56db39ccf2ae46c05afb3fd7866aec405fb04d0c0
|
| 3 |
+
size 4984
|
checkpoint-840/config.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-medium",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"apply_spec_augment": false,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"WhisperForConditionalGeneration"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"begin_suppress_tokens": [
|
| 11 |
+
220,
|
| 12 |
+
50257
|
| 13 |
+
],
|
| 14 |
+
"bos_token_id": 50257,
|
| 15 |
+
"classifier_proj_size": 256,
|
| 16 |
+
"d_model": 1024,
|
| 17 |
+
"decoder_attention_heads": 16,
|
| 18 |
+
"decoder_ffn_dim": 4096,
|
| 19 |
+
"decoder_layerdrop": 0.0,
|
| 20 |
+
"decoder_layers": 24,
|
| 21 |
+
"decoder_start_token_id": 50258,
|
| 22 |
+
"dropout": 0.0,
|
| 23 |
+
"encoder_attention_heads": 16,
|
| 24 |
+
"encoder_ffn_dim": 4096,
|
| 25 |
+
"encoder_layerdrop": 0.0,
|
| 26 |
+
"encoder_layers": 24,
|
| 27 |
+
"eos_token_id": 50257,
|
| 28 |
+
"forced_decoder_ids": [
|
| 29 |
+
[
|
| 30 |
+
1,
|
| 31 |
+
50259
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
2,
|
| 35 |
+
50359
|
| 36 |
+
],
|
| 37 |
+
[
|
| 38 |
+
3,
|
| 39 |
+
50363
|
| 40 |
+
]
|
| 41 |
+
],
|
| 42 |
+
"init_std": 0.02,
|
| 43 |
+
"is_encoder_decoder": true,
|
| 44 |
+
"language": "English",
|
| 45 |
+
"mask_feature_length": 10,
|
| 46 |
+
"mask_feature_min_masks": 0,
|
| 47 |
+
"mask_feature_prob": 0.0,
|
| 48 |
+
"mask_time_length": 10,
|
| 49 |
+
"mask_time_min_masks": 2,
|
| 50 |
+
"mask_time_prob": 0.05,
|
| 51 |
+
"max_length": 448,
|
| 52 |
+
"max_source_positions": 1500,
|
| 53 |
+
"max_target_positions": 448,
|
| 54 |
+
"median_filter_width": 7,
|
| 55 |
+
"model_type": "whisper",
|
| 56 |
+
"num_hidden_layers": 24,
|
| 57 |
+
"num_mel_bins": 80,
|
| 58 |
+
"pad_token_id": 50257,
|
| 59 |
+
"scale_embedding": false,
|
| 60 |
+
"suppress_tokens": [],
|
| 61 |
+
"torch_dtype": "float32",
|
| 62 |
+
"transformers_version": "4.37.2",
|
| 63 |
+
"use_cache": true,
|
| 64 |
+
"use_weighted_layer_sum": false,
|
| 65 |
+
"vocab_size": 51865
|
| 66 |
+
}
|
checkpoint-840/generation_config.json
ADDED
|
@@ -0,0 +1,249 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alignment_heads": [
|
| 3 |
+
[
|
| 4 |
+
13,
|
| 5 |
+
15
|
| 6 |
+
],
|
| 7 |
+
[
|
| 8 |
+
15,
|
| 9 |
+
4
|
| 10 |
+
],
|
| 11 |
+
[
|
| 12 |
+
15,
|
| 13 |
+
15
|
| 14 |
+
],
|
| 15 |
+
[
|
| 16 |
+
16,
|
| 17 |
+
1
|
| 18 |
+
],
|
| 19 |
+
[
|
| 20 |
+
20,
|
| 21 |
+
0
|
| 22 |
+
],
|
| 23 |
+
[
|
| 24 |
+
23,
|
| 25 |
+
4
|
| 26 |
+
]
|
| 27 |
+
],
|
| 28 |
+
"begin_suppress_tokens": [
|
| 29 |
+
220,
|
| 30 |
+
50257
|
| 31 |
+
],
|
| 32 |
+
"bos_token_id": 50257,
|
| 33 |
+
"decoder_start_token_id": 50258,
|
| 34 |
+
"eos_token_id": 50257,
|
| 35 |
+
"forced_decoder_ids": [
|
| 36 |
+
[
|
| 37 |
+
1,
|
| 38 |
+
null
|
| 39 |
+
],
|
| 40 |
+
[
|
| 41 |
+
2,
|
| 42 |
+
50359
|
| 43 |
+
]
|
| 44 |
+
],
|
| 45 |
+
"is_multilingual": true,
|
| 46 |
+
"lang_to_id": {
|
| 47 |
+
"<|af|>": 50327,
|
| 48 |
+
"<|am|>": 50334,
|
| 49 |
+
"<|ar|>": 50272,
|
| 50 |
+
"<|as|>": 50350,
|
| 51 |
+
"<|az|>": 50304,
|
| 52 |
+
"<|ba|>": 50355,
|
| 53 |
+
"<|be|>": 50330,
|
| 54 |
+
"<|bg|>": 50292,
|
| 55 |
+
"<|bn|>": 50302,
|
| 56 |
+
"<|bo|>": 50347,
|
| 57 |
+
"<|br|>": 50309,
|
| 58 |
+
"<|bs|>": 50315,
|
| 59 |
+
"<|ca|>": 50270,
|
| 60 |
+
"<|cs|>": 50283,
|
| 61 |
+
"<|cy|>": 50297,
|
| 62 |
+
"<|da|>": 50285,
|
| 63 |
+
"<|de|>": 50261,
|
| 64 |
+
"<|el|>": 50281,
|
| 65 |
+
"<|en|>": 50259,
|
| 66 |
+
"<|es|>": 50262,
|
| 67 |
+
"<|et|>": 50307,
|
| 68 |
+
"<|eu|>": 50310,
|
| 69 |
+
"<|fa|>": 50300,
|
| 70 |
+
"<|fi|>": 50277,
|
| 71 |
+
"<|fo|>": 50338,
|
| 72 |
+
"<|fr|>": 50265,
|
| 73 |
+
"<|gl|>": 50319,
|
| 74 |
+
"<|gu|>": 50333,
|
| 75 |
+
"<|haw|>": 50352,
|
| 76 |
+
"<|ha|>": 50354,
|
| 77 |
+
"<|he|>": 50279,
|
| 78 |
+
"<|hi|>": 50276,
|
| 79 |
+
"<|hr|>": 50291,
|
| 80 |
+
"<|ht|>": 50339,
|
| 81 |
+
"<|hu|>": 50286,
|
| 82 |
+
"<|hy|>": 50312,
|
| 83 |
+
"<|id|>": 50275,
|
| 84 |
+
"<|is|>": 50311,
|
| 85 |
+
"<|it|>": 50274,
|
| 86 |
+
"<|ja|>": 50266,
|
| 87 |
+
"<|jw|>": 50356,
|
| 88 |
+
"<|ka|>": 50329,
|
| 89 |
+
"<|kk|>": 50316,
|
| 90 |
+
"<|km|>": 50323,
|
| 91 |
+
"<|kn|>": 50306,
|
| 92 |
+
"<|ko|>": 50264,
|
| 93 |
+
"<|la|>": 50294,
|
| 94 |
+
"<|lb|>": 50345,
|
| 95 |
+
"<|ln|>": 50353,
|
| 96 |
+
"<|lo|>": 50336,
|
| 97 |
+
"<|lt|>": 50293,
|
| 98 |
+
"<|lv|>": 50301,
|
| 99 |
+
"<|mg|>": 50349,
|
| 100 |
+
"<|mi|>": 50295,
|
| 101 |
+
"<|mk|>": 50308,
|
| 102 |
+
"<|ml|>": 50296,
|
| 103 |
+
"<|mn|>": 50314,
|
| 104 |
+
"<|mr|>": 50320,
|
| 105 |
+
"<|ms|>": 50282,
|
| 106 |
+
"<|mt|>": 50343,
|
| 107 |
+
"<|my|>": 50346,
|
| 108 |
+
"<|ne|>": 50313,
|
| 109 |
+
"<|nl|>": 50271,
|
| 110 |
+
"<|nn|>": 50342,
|
| 111 |
+
"<|no|>": 50288,
|
| 112 |
+
"<|oc|>": 50328,
|
| 113 |
+
"<|pa|>": 50321,
|
| 114 |
+
"<|pl|>": 50269,
|
| 115 |
+
"<|ps|>": 50340,
|
| 116 |
+
"<|pt|>": 50267,
|
| 117 |
+
"<|ro|>": 50284,
|
| 118 |
+
"<|ru|>": 50263,
|
| 119 |
+
"<|sa|>": 50344,
|
| 120 |
+
"<|sd|>": 50332,
|
| 121 |
+
"<|si|>": 50322,
|
| 122 |
+
"<|sk|>": 50298,
|
| 123 |
+
"<|sl|>": 50305,
|
| 124 |
+
"<|sn|>": 50324,
|
| 125 |
+
"<|so|>": 50326,
|
| 126 |
+
"<|sq|>": 50317,
|
| 127 |
+
"<|sr|>": 50303,
|
| 128 |
+
"<|su|>": 50357,
|
| 129 |
+
"<|sv|>": 50273,
|
| 130 |
+
"<|sw|>": 50318,
|
| 131 |
+
"<|ta|>": 50287,
|
| 132 |
+
"<|te|>": 50299,
|
| 133 |
+
"<|tg|>": 50331,
|
| 134 |
+
"<|th|>": 50289,
|
| 135 |
+
"<|tk|>": 50341,
|
| 136 |
+
"<|tl|>": 50348,
|
| 137 |
+
"<|tr|>": 50268,
|
| 138 |
+
"<|tt|>": 50351,
|
| 139 |
+
"<|uk|>": 50280,
|
| 140 |
+
"<|ur|>": 50290,
|
| 141 |
+
"<|uz|>": 50337,
|
| 142 |
+
"<|vi|>": 50278,
|
| 143 |
+
"<|yi|>": 50335,
|
| 144 |
+
"<|yo|>": 50325,
|
| 145 |
+
"<|zh|>": 50260
|
| 146 |
+
},
|
| 147 |
+
"max_initial_timestamp_index": 50,
|
| 148 |
+
"max_length": 448,
|
| 149 |
+
"no_timestamps_token_id": 50363,
|
| 150 |
+
"pad_token_id": 50257,
|
| 151 |
+
"prev_sot_token_id": 50361,
|
| 152 |
+
"return_timestamps": false,
|
| 153 |
+
"suppress_tokens": [
|
| 154 |
+
1,
|
| 155 |
+
2,
|
| 156 |
+
7,
|
| 157 |
+
8,
|
| 158 |
+
9,
|
| 159 |
+
10,
|
| 160 |
+
14,
|
| 161 |
+
25,
|
| 162 |
+
26,
|
| 163 |
+
27,
|
| 164 |
+
28,
|
| 165 |
+
29,
|
| 166 |
+
31,
|
| 167 |
+
58,
|
| 168 |
+
59,
|
| 169 |
+
60,
|
| 170 |
+
61,
|
| 171 |
+
62,
|
| 172 |
+
63,
|
| 173 |
+
90,
|
| 174 |
+
91,
|
| 175 |
+
92,
|
| 176 |
+
93,
|
| 177 |
+
359,
|
| 178 |
+
503,
|
| 179 |
+
522,
|
| 180 |
+
542,
|
| 181 |
+
873,
|
| 182 |
+
893,
|
| 183 |
+
902,
|
| 184 |
+
918,
|
| 185 |
+
922,
|
| 186 |
+
931,
|
| 187 |
+
1350,
|
| 188 |
+
1853,
|
| 189 |
+
1982,
|
| 190 |
+
2460,
|
| 191 |
+
2627,
|
| 192 |
+
3246,
|
| 193 |
+
3253,
|
| 194 |
+
3268,
|
| 195 |
+
3536,
|
| 196 |
+
3846,
|
| 197 |
+
3961,
|
| 198 |
+
4183,
|
| 199 |
+
4667,
|
| 200 |
+
6585,
|
| 201 |
+
6647,
|
| 202 |
+
7273,
|
| 203 |
+
9061,
|
| 204 |
+
9383,
|
| 205 |
+
10428,
|
| 206 |
+
10929,
|
| 207 |
+
11938,
|
| 208 |
+
12033,
|
| 209 |
+
12331,
|
| 210 |
+
12562,
|
| 211 |
+
13793,
|
| 212 |
+
14157,
|
| 213 |
+
14635,
|
| 214 |
+
15265,
|
| 215 |
+
15618,
|
| 216 |
+
16553,
|
| 217 |
+
16604,
|
| 218 |
+
18362,
|
| 219 |
+
18956,
|
| 220 |
+
20075,
|
| 221 |
+
21675,
|
| 222 |
+
22520,
|
| 223 |
+
26130,
|
| 224 |
+
26161,
|
| 225 |
+
26435,
|
| 226 |
+
28279,
|
| 227 |
+
29464,
|
| 228 |
+
31650,
|
| 229 |
+
32302,
|
| 230 |
+
32470,
|
| 231 |
+
36865,
|
| 232 |
+
42863,
|
| 233 |
+
47425,
|
| 234 |
+
49870,
|
| 235 |
+
50254,
|
| 236 |
+
50258,
|
| 237 |
+
50358,
|
| 238 |
+
50359,
|
| 239 |
+
50360,
|
| 240 |
+
50361,
|
| 241 |
+
50362
|
| 242 |
+
],
|
| 243 |
+
"task": "transcribe",
|
| 244 |
+
"task_to_id": {
|
| 245 |
+
"transcribe": 50359,
|
| 246 |
+
"translate": 50358
|
| 247 |
+
},
|
| 248 |
+
"transformers_version": "4.37.2"
|
| 249 |
+
}
|
checkpoint-840/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfdba08f003d04a2f4f5c72d79d9334c4c7ef886d55357ca0fe7fe7247fcf429
|
| 3 |
+
size 3055544304
|
checkpoint-840/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a7c876a631fdfee0a0f6a9ffe5cab8f82cd2b07a1315f25043bb89ab4f3476fc
|
| 3 |
+
size 1531565088
|
checkpoint-840/preprocessor_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length": 30,
|
| 3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
| 4 |
+
"feature_size": 80,
|
| 5 |
+
"hop_length": 160,
|
| 6 |
+
"n_fft": 400,
|
| 7 |
+
"n_samples": 480000,
|
| 8 |
+
"nb_max_frames": 3000,
|
| 9 |
+
"padding_side": "right",
|
| 10 |
+
"padding_value": 0.0,
|
| 11 |
+
"processor_class": "WhisperProcessor",
|
| 12 |
+
"return_attention_mask": false,
|
| 13 |
+
"sampling_rate": 16000
|
| 14 |
+
}
|
checkpoint-840/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d5a42e8569123809e30694bb525f0a9dcb868aabf97f3f156c573870dd7f086b
|
| 3 |
+
size 14244
|
checkpoint-840/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c5c4722b91ccea5b7b184ef5a791709a3f952084c8d754d2b137874ea75f273c
|
| 3 |
+
size 1064
|
checkpoint-840/trainer_state.json
ADDED
|
@@ -0,0 +1,1281 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 840,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 0.0,
|
| 14 |
+
"loss": 5.3457,
|
| 15 |
+
"step": 4
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 4e-08,
|
| 20 |
+
"loss": 5.2606,
|
| 21 |
+
"step": 8
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.01,
|
| 25 |
+
"learning_rate": 7e-08,
|
| 26 |
+
"loss": 5.0776,
|
| 27 |
+
"step": 12
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.02,
|
| 31 |
+
"learning_rate": 1e-07,
|
| 32 |
+
"loss": 5.149,
|
| 33 |
+
"step": 16
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.02,
|
| 37 |
+
"learning_rate": 1.4e-07,
|
| 38 |
+
"loss": 5.3724,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.03,
|
| 43 |
+
"learning_rate": 1.7000000000000001e-07,
|
| 44 |
+
"loss": 5.2246,
|
| 45 |
+
"step": 24
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.03,
|
| 49 |
+
"learning_rate": 2.0999999999999997e-07,
|
| 50 |
+
"loss": 5.032,
|
| 51 |
+
"step": 28
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.04,
|
| 55 |
+
"learning_rate": 2.5e-07,
|
| 56 |
+
"loss": 5.0639,
|
| 57 |
+
"step": 32
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.04,
|
| 61 |
+
"learning_rate": 2.9e-07,
|
| 62 |
+
"loss": 4.5604,
|
| 63 |
+
"step": 36
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.05,
|
| 67 |
+
"learning_rate": 3.3e-07,
|
| 68 |
+
"loss": 4.1731,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.05,
|
| 73 |
+
"learning_rate": 3.7e-07,
|
| 74 |
+
"loss": 4.1152,
|
| 75 |
+
"step": 44
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.06,
|
| 79 |
+
"learning_rate": 4.0999999999999994e-07,
|
| 80 |
+
"loss": 4.101,
|
| 81 |
+
"step": 48
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.06,
|
| 85 |
+
"learning_rate": 4.5e-07,
|
| 86 |
+
"loss": 3.8602,
|
| 87 |
+
"step": 52
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.07,
|
| 91 |
+
"learning_rate": 4.9e-07,
|
| 92 |
+
"loss": 3.7321,
|
| 93 |
+
"step": 56
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.07,
|
| 97 |
+
"learning_rate": 5.3e-07,
|
| 98 |
+
"loss": 3.6759,
|
| 99 |
+
"step": 60
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.08,
|
| 103 |
+
"learning_rate": 5.699999999999999e-07,
|
| 104 |
+
"loss": 3.6372,
|
| 105 |
+
"step": 64
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.08,
|
| 109 |
+
"learning_rate": 6.1e-07,
|
| 110 |
+
"loss": 3.5755,
|
| 111 |
+
"step": 68
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.09,
|
| 115 |
+
"learning_rate": 6.5e-07,
|
| 116 |
+
"loss": 3.415,
|
| 117 |
+
"step": 72
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.09,
|
| 121 |
+
"learning_rate": 6.9e-07,
|
| 122 |
+
"loss": 3.215,
|
| 123 |
+
"step": 76
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.1,
|
| 127 |
+
"learning_rate": 7.3e-07,
|
| 128 |
+
"loss": 2.8689,
|
| 129 |
+
"step": 80
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.1,
|
| 133 |
+
"learning_rate": 7.699999999999999e-07,
|
| 134 |
+
"loss": 2.843,
|
| 135 |
+
"step": 84
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.1,
|
| 139 |
+
"learning_rate": 8.1e-07,
|
| 140 |
+
"loss": 2.9887,
|
| 141 |
+
"step": 88
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.11,
|
| 145 |
+
"learning_rate": 8.499999999999999e-07,
|
| 146 |
+
"loss": 2.6316,
|
| 147 |
+
"step": 92
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.11,
|
| 151 |
+
"learning_rate": 8.9e-07,
|
| 152 |
+
"loss": 2.4875,
|
| 153 |
+
"step": 96
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.12,
|
| 157 |
+
"learning_rate": 9.3e-07,
|
| 158 |
+
"loss": 2.4555,
|
| 159 |
+
"step": 100
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.12,
|
| 163 |
+
"learning_rate": 9.7e-07,
|
| 164 |
+
"loss": 2.3976,
|
| 165 |
+
"step": 104
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.13,
|
| 169 |
+
"learning_rate": 9.99367088607595e-07,
|
| 170 |
+
"loss": 2.2534,
|
| 171 |
+
"step": 108
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.13,
|
| 175 |
+
"learning_rate": 9.968354430379747e-07,
|
| 176 |
+
"loss": 2.1527,
|
| 177 |
+
"step": 112
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.14,
|
| 181 |
+
"learning_rate": 9.943037974683544e-07,
|
| 182 |
+
"loss": 2.0296,
|
| 183 |
+
"step": 116
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.14,
|
| 187 |
+
"learning_rate": 9.91772151898734e-07,
|
| 188 |
+
"loss": 2.0501,
|
| 189 |
+
"step": 120
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.15,
|
| 193 |
+
"learning_rate": 9.89240506329114e-07,
|
| 194 |
+
"loss": 2.0385,
|
| 195 |
+
"step": 124
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.15,
|
| 199 |
+
"learning_rate": 9.867088607594938e-07,
|
| 200 |
+
"loss": 1.9757,
|
| 201 |
+
"step": 128
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.16,
|
| 205 |
+
"learning_rate": 9.841772151898733e-07,
|
| 206 |
+
"loss": 1.7414,
|
| 207 |
+
"step": 132
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.16,
|
| 211 |
+
"learning_rate": 9.81645569620253e-07,
|
| 212 |
+
"loss": 1.7658,
|
| 213 |
+
"step": 136
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.17,
|
| 217 |
+
"learning_rate": 9.791139240506329e-07,
|
| 218 |
+
"loss": 1.6414,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.17,
|
| 223 |
+
"learning_rate": 9.765822784810126e-07,
|
| 224 |
+
"loss": 1.7797,
|
| 225 |
+
"step": 144
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.18,
|
| 229 |
+
"learning_rate": 9.740506329113924e-07,
|
| 230 |
+
"loss": 1.5513,
|
| 231 |
+
"step": 148
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.18,
|
| 235 |
+
"learning_rate": 9.715189873417722e-07,
|
| 236 |
+
"loss": 1.5515,
|
| 237 |
+
"step": 152
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.19,
|
| 241 |
+
"learning_rate": 9.689873417721517e-07,
|
| 242 |
+
"loss": 1.471,
|
| 243 |
+
"step": 156
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.19,
|
| 247 |
+
"learning_rate": 9.664556962025317e-07,
|
| 248 |
+
"loss": 1.3805,
|
| 249 |
+
"step": 160
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.2,
|
| 253 |
+
"learning_rate": 9.639240506329115e-07,
|
| 254 |
+
"loss": 1.3442,
|
| 255 |
+
"step": 164
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.2,
|
| 259 |
+
"learning_rate": 9.61392405063291e-07,
|
| 260 |
+
"loss": 1.2072,
|
| 261 |
+
"step": 168
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.2,
|
| 265 |
+
"learning_rate": 9.588607594936708e-07,
|
| 266 |
+
"loss": 1.2704,
|
| 267 |
+
"step": 172
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.21,
|
| 271 |
+
"learning_rate": 9.563291139240506e-07,
|
| 272 |
+
"loss": 1.3914,
|
| 273 |
+
"step": 176
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.21,
|
| 277 |
+
"learning_rate": 9.537974683544304e-07,
|
| 278 |
+
"loss": 1.2708,
|
| 279 |
+
"step": 180
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.22,
|
| 283 |
+
"learning_rate": 9.5126582278481e-07,
|
| 284 |
+
"loss": 1.2163,
|
| 285 |
+
"step": 184
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.22,
|
| 289 |
+
"learning_rate": 9.487341772151898e-07,
|
| 290 |
+
"loss": 1.2604,
|
| 291 |
+
"step": 188
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.23,
|
| 295 |
+
"learning_rate": 9.462025316455697e-07,
|
| 296 |
+
"loss": 1.0958,
|
| 297 |
+
"step": 192
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.23,
|
| 301 |
+
"learning_rate": 9.436708860759493e-07,
|
| 302 |
+
"loss": 1.1216,
|
| 303 |
+
"step": 196
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.24,
|
| 307 |
+
"learning_rate": 9.411392405063291e-07,
|
| 308 |
+
"loss": 1.1185,
|
| 309 |
+
"step": 200
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.24,
|
| 313 |
+
"learning_rate": 9.386075949367089e-07,
|
| 314 |
+
"loss": 1.0496,
|
| 315 |
+
"step": 204
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.25,
|
| 319 |
+
"learning_rate": 9.360759493670885e-07,
|
| 320 |
+
"loss": 0.9438,
|
| 321 |
+
"step": 208
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.25,
|
| 325 |
+
"learning_rate": 9.335443037974683e-07,
|
| 326 |
+
"loss": 0.8912,
|
| 327 |
+
"step": 212
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.26,
|
| 331 |
+
"learning_rate": 9.31012658227848e-07,
|
| 332 |
+
"loss": 0.8424,
|
| 333 |
+
"step": 216
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.26,
|
| 337 |
+
"learning_rate": 9.284810126582277e-07,
|
| 338 |
+
"loss": 0.9812,
|
| 339 |
+
"step": 220
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.27,
|
| 343 |
+
"learning_rate": 9.259493670886076e-07,
|
| 344 |
+
"loss": 0.8999,
|
| 345 |
+
"step": 224
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.27,
|
| 349 |
+
"learning_rate": 9.234177215189873e-07,
|
| 350 |
+
"loss": 0.7623,
|
| 351 |
+
"step": 228
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.28,
|
| 355 |
+
"learning_rate": 9.208860759493671e-07,
|
| 356 |
+
"loss": 0.6592,
|
| 357 |
+
"step": 232
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.28,
|
| 361 |
+
"learning_rate": 9.183544303797468e-07,
|
| 362 |
+
"loss": 0.7013,
|
| 363 |
+
"step": 236
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.29,
|
| 367 |
+
"learning_rate": 9.158227848101265e-07,
|
| 368 |
+
"loss": 0.5614,
|
| 369 |
+
"step": 240
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.29,
|
| 373 |
+
"learning_rate": 9.132911392405063e-07,
|
| 374 |
+
"loss": 0.5246,
|
| 375 |
+
"step": 244
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.3,
|
| 379 |
+
"learning_rate": 9.10759493670886e-07,
|
| 380 |
+
"loss": 0.3732,
|
| 381 |
+
"step": 248
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.3,
|
| 385 |
+
"learning_rate": 9.082278481012657e-07,
|
| 386 |
+
"loss": 0.3963,
|
| 387 |
+
"step": 252
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3,
|
| 391 |
+
"learning_rate": 9.056962025316456e-07,
|
| 392 |
+
"loss": 0.3281,
|
| 393 |
+
"step": 256
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.31,
|
| 397 |
+
"learning_rate": 9.031645569620254e-07,
|
| 398 |
+
"loss": 0.3707,
|
| 399 |
+
"step": 260
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.31,
|
| 403 |
+
"learning_rate": 9.00632911392405e-07,
|
| 404 |
+
"loss": 0.3442,
|
| 405 |
+
"step": 264
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.32,
|
| 409 |
+
"learning_rate": 8.981012658227848e-07,
|
| 410 |
+
"loss": 0.286,
|
| 411 |
+
"step": 268
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.32,
|
| 415 |
+
"learning_rate": 8.955696202531646e-07,
|
| 416 |
+
"loss": 0.2778,
|
| 417 |
+
"step": 272
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.33,
|
| 421 |
+
"learning_rate": 8.930379746835442e-07,
|
| 422 |
+
"loss": 0.3224,
|
| 423 |
+
"step": 276
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.33,
|
| 427 |
+
"learning_rate": 8.90506329113924e-07,
|
| 428 |
+
"loss": 0.2294,
|
| 429 |
+
"step": 280
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.34,
|
| 433 |
+
"learning_rate": 8.879746835443038e-07,
|
| 434 |
+
"loss": 0.3471,
|
| 435 |
+
"step": 284
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.34,
|
| 439 |
+
"learning_rate": 8.854430379746834e-07,
|
| 440 |
+
"loss": 0.2537,
|
| 441 |
+
"step": 288
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.35,
|
| 445 |
+
"learning_rate": 8.829113924050633e-07,
|
| 446 |
+
"loss": 0.3189,
|
| 447 |
+
"step": 292
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.35,
|
| 451 |
+
"learning_rate": 8.803797468354431e-07,
|
| 452 |
+
"loss": 0.2704,
|
| 453 |
+
"step": 296
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.36,
|
| 457 |
+
"learning_rate": 8.778481012658227e-07,
|
| 458 |
+
"loss": 0.2685,
|
| 459 |
+
"step": 300
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.36,
|
| 463 |
+
"learning_rate": 8.753164556962025e-07,
|
| 464 |
+
"loss": 0.4016,
|
| 465 |
+
"step": 304
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.37,
|
| 469 |
+
"learning_rate": 8.727848101265822e-07,
|
| 470 |
+
"loss": 0.2614,
|
| 471 |
+
"step": 308
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.37,
|
| 475 |
+
"learning_rate": 8.70253164556962e-07,
|
| 476 |
+
"loss": 0.2208,
|
| 477 |
+
"step": 312
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.38,
|
| 481 |
+
"learning_rate": 8.677215189873417e-07,
|
| 482 |
+
"loss": 0.3452,
|
| 483 |
+
"step": 316
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.38,
|
| 487 |
+
"learning_rate": 8.651898734177214e-07,
|
| 488 |
+
"loss": 0.2481,
|
| 489 |
+
"step": 320
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.39,
|
| 493 |
+
"learning_rate": 8.626582278481013e-07,
|
| 494 |
+
"loss": 0.2698,
|
| 495 |
+
"step": 324
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.39,
|
| 499 |
+
"learning_rate": 8.60126582278481e-07,
|
| 500 |
+
"loss": 0.2386,
|
| 501 |
+
"step": 328
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.4,
|
| 505 |
+
"learning_rate": 8.575949367088607e-07,
|
| 506 |
+
"loss": 0.27,
|
| 507 |
+
"step": 332
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.4,
|
| 511 |
+
"learning_rate": 8.550632911392405e-07,
|
| 512 |
+
"loss": 0.4073,
|
| 513 |
+
"step": 336
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.4,
|
| 517 |
+
"learning_rate": 8.525316455696202e-07,
|
| 518 |
+
"loss": 0.263,
|
| 519 |
+
"step": 340
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.41,
|
| 523 |
+
"learning_rate": 8.499999999999999e-07,
|
| 524 |
+
"loss": 0.2443,
|
| 525 |
+
"step": 344
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.41,
|
| 529 |
+
"learning_rate": 8.474683544303797e-07,
|
| 530 |
+
"loss": 0.1876,
|
| 531 |
+
"step": 348
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.42,
|
| 535 |
+
"learning_rate": 8.449367088607595e-07,
|
| 536 |
+
"loss": 0.1893,
|
| 537 |
+
"step": 352
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.42,
|
| 541 |
+
"learning_rate": 8.424050632911392e-07,
|
| 542 |
+
"loss": 0.356,
|
| 543 |
+
"step": 356
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.43,
|
| 547 |
+
"learning_rate": 8.39873417721519e-07,
|
| 548 |
+
"loss": 0.3026,
|
| 549 |
+
"step": 360
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.43,
|
| 553 |
+
"learning_rate": 8.373417721518988e-07,
|
| 554 |
+
"loss": 0.2265,
|
| 555 |
+
"step": 364
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.44,
|
| 559 |
+
"learning_rate": 8.348101265822784e-07,
|
| 560 |
+
"loss": 0.3389,
|
| 561 |
+
"step": 368
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.44,
|
| 565 |
+
"learning_rate": 8.322784810126582e-07,
|
| 566 |
+
"loss": 0.274,
|
| 567 |
+
"step": 372
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.45,
|
| 571 |
+
"learning_rate": 8.297468354430379e-07,
|
| 572 |
+
"loss": 0.3028,
|
| 573 |
+
"step": 376
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.45,
|
| 577 |
+
"learning_rate": 8.272151898734176e-07,
|
| 578 |
+
"loss": 0.2582,
|
| 579 |
+
"step": 380
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.46,
|
| 583 |
+
"learning_rate": 8.246835443037974e-07,
|
| 584 |
+
"loss": 0.253,
|
| 585 |
+
"step": 384
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.46,
|
| 589 |
+
"learning_rate": 8.221518987341772e-07,
|
| 590 |
+
"loss": 0.2579,
|
| 591 |
+
"step": 388
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.47,
|
| 595 |
+
"learning_rate": 8.19620253164557e-07,
|
| 596 |
+
"loss": 0.3189,
|
| 597 |
+
"step": 392
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.47,
|
| 601 |
+
"learning_rate": 8.170886075949367e-07,
|
| 602 |
+
"loss": 0.2599,
|
| 603 |
+
"step": 396
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.48,
|
| 607 |
+
"learning_rate": 8.145569620253164e-07,
|
| 608 |
+
"loss": 0.2791,
|
| 609 |
+
"step": 400
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.48,
|
| 613 |
+
"learning_rate": 8.120253164556962e-07,
|
| 614 |
+
"loss": 0.3284,
|
| 615 |
+
"step": 404
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.49,
|
| 619 |
+
"learning_rate": 8.094936708860759e-07,
|
| 620 |
+
"loss": 0.3026,
|
| 621 |
+
"step": 408
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.49,
|
| 625 |
+
"learning_rate": 8.069620253164556e-07,
|
| 626 |
+
"loss": 0.3195,
|
| 627 |
+
"step": 412
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.5,
|
| 631 |
+
"learning_rate": 8.044303797468354e-07,
|
| 632 |
+
"loss": 0.2776,
|
| 633 |
+
"step": 416
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.5,
|
| 637 |
+
"learning_rate": 8.018987341772152e-07,
|
| 638 |
+
"loss": 0.3193,
|
| 639 |
+
"step": 420
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5,
|
| 643 |
+
"learning_rate": 7.993670886075949e-07,
|
| 644 |
+
"loss": 0.2963,
|
| 645 |
+
"step": 424
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.51,
|
| 649 |
+
"learning_rate": 7.968354430379747e-07,
|
| 650 |
+
"loss": 0.2865,
|
| 651 |
+
"step": 428
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.51,
|
| 655 |
+
"learning_rate": 7.943037974683544e-07,
|
| 656 |
+
"loss": 0.3051,
|
| 657 |
+
"step": 432
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.52,
|
| 661 |
+
"learning_rate": 7.917721518987341e-07,
|
| 662 |
+
"loss": 0.2691,
|
| 663 |
+
"step": 436
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 0.52,
|
| 667 |
+
"learning_rate": 7.892405063291139e-07,
|
| 668 |
+
"loss": 0.2542,
|
| 669 |
+
"step": 440
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.53,
|
| 673 |
+
"learning_rate": 7.867088607594937e-07,
|
| 674 |
+
"loss": 0.3001,
|
| 675 |
+
"step": 444
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.53,
|
| 679 |
+
"learning_rate": 7.841772151898733e-07,
|
| 680 |
+
"loss": 0.2793,
|
| 681 |
+
"step": 448
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.54,
|
| 685 |
+
"learning_rate": 7.816455696202531e-07,
|
| 686 |
+
"loss": 0.2546,
|
| 687 |
+
"step": 452
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.54,
|
| 691 |
+
"learning_rate": 7.79113924050633e-07,
|
| 692 |
+
"loss": 0.2801,
|
| 693 |
+
"step": 456
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.55,
|
| 697 |
+
"learning_rate": 7.765822784810126e-07,
|
| 698 |
+
"loss": 0.2635,
|
| 699 |
+
"step": 460
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.55,
|
| 703 |
+
"learning_rate": 7.740506329113924e-07,
|
| 704 |
+
"loss": 0.2678,
|
| 705 |
+
"step": 464
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 0.56,
|
| 709 |
+
"learning_rate": 7.715189873417721e-07,
|
| 710 |
+
"loss": 0.379,
|
| 711 |
+
"step": 468
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.56,
|
| 715 |
+
"learning_rate": 7.689873417721518e-07,
|
| 716 |
+
"loss": 0.3077,
|
| 717 |
+
"step": 472
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.57,
|
| 721 |
+
"learning_rate": 7.664556962025316e-07,
|
| 722 |
+
"loss": 0.1788,
|
| 723 |
+
"step": 476
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.57,
|
| 727 |
+
"learning_rate": 7.639240506329113e-07,
|
| 728 |
+
"loss": 0.3147,
|
| 729 |
+
"step": 480
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.58,
|
| 733 |
+
"learning_rate": 7.61392405063291e-07,
|
| 734 |
+
"loss": 0.3064,
|
| 735 |
+
"step": 484
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.58,
|
| 739 |
+
"learning_rate": 7.588607594936709e-07,
|
| 740 |
+
"loss": 0.2644,
|
| 741 |
+
"step": 488
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.59,
|
| 745 |
+
"learning_rate": 7.563291139240506e-07,
|
| 746 |
+
"loss": 0.1713,
|
| 747 |
+
"step": 492
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.59,
|
| 751 |
+
"learning_rate": 7.537974683544304e-07,
|
| 752 |
+
"loss": 0.2332,
|
| 753 |
+
"step": 496
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.6,
|
| 757 |
+
"learning_rate": 7.512658227848101e-07,
|
| 758 |
+
"loss": 0.3326,
|
| 759 |
+
"step": 500
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.6,
|
| 763 |
+
"learning_rate": 7.487341772151898e-07,
|
| 764 |
+
"loss": 0.3293,
|
| 765 |
+
"step": 504
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.6,
|
| 769 |
+
"learning_rate": 7.462025316455696e-07,
|
| 770 |
+
"loss": 0.1996,
|
| 771 |
+
"step": 508
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 0.61,
|
| 775 |
+
"learning_rate": 7.436708860759493e-07,
|
| 776 |
+
"loss": 0.2282,
|
| 777 |
+
"step": 512
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.61,
|
| 781 |
+
"learning_rate": 7.41139240506329e-07,
|
| 782 |
+
"loss": 0.2591,
|
| 783 |
+
"step": 516
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.62,
|
| 787 |
+
"learning_rate": 7.386075949367089e-07,
|
| 788 |
+
"loss": 0.2376,
|
| 789 |
+
"step": 520
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.62,
|
| 793 |
+
"learning_rate": 7.360759493670887e-07,
|
| 794 |
+
"loss": 0.2115,
|
| 795 |
+
"step": 524
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.63,
|
| 799 |
+
"learning_rate": 7.335443037974683e-07,
|
| 800 |
+
"loss": 0.2515,
|
| 801 |
+
"step": 528
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.63,
|
| 805 |
+
"learning_rate": 7.310126582278481e-07,
|
| 806 |
+
"loss": 0.3904,
|
| 807 |
+
"step": 532
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.64,
|
| 811 |
+
"learning_rate": 7.284810126582279e-07,
|
| 812 |
+
"loss": 0.2693,
|
| 813 |
+
"step": 536
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.64,
|
| 817 |
+
"learning_rate": 7.259493670886075e-07,
|
| 818 |
+
"loss": 0.2653,
|
| 819 |
+
"step": 540
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.65,
|
| 823 |
+
"learning_rate": 7.234177215189873e-07,
|
| 824 |
+
"loss": 0.2901,
|
| 825 |
+
"step": 544
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.65,
|
| 829 |
+
"learning_rate": 7.20886075949367e-07,
|
| 830 |
+
"loss": 0.3418,
|
| 831 |
+
"step": 548
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.66,
|
| 835 |
+
"learning_rate": 7.183544303797468e-07,
|
| 836 |
+
"loss": 0.2681,
|
| 837 |
+
"step": 552
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.66,
|
| 841 |
+
"learning_rate": 7.158227848101266e-07,
|
| 842 |
+
"loss": 0.2793,
|
| 843 |
+
"step": 556
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.67,
|
| 847 |
+
"learning_rate": 7.132911392405063e-07,
|
| 848 |
+
"loss": 0.2777,
|
| 849 |
+
"step": 560
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.67,
|
| 853 |
+
"learning_rate": 7.10759493670886e-07,
|
| 854 |
+
"loss": 0.1978,
|
| 855 |
+
"step": 564
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 0.68,
|
| 859 |
+
"learning_rate": 7.082278481012658e-07,
|
| 860 |
+
"loss": 0.2681,
|
| 861 |
+
"step": 568
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 0.68,
|
| 865 |
+
"learning_rate": 7.056962025316455e-07,
|
| 866 |
+
"loss": 0.2808,
|
| 867 |
+
"step": 572
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.69,
|
| 871 |
+
"learning_rate": 7.031645569620253e-07,
|
| 872 |
+
"loss": 0.2718,
|
| 873 |
+
"step": 576
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.69,
|
| 877 |
+
"learning_rate": 7.00632911392405e-07,
|
| 878 |
+
"loss": 0.2197,
|
| 879 |
+
"step": 580
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.7,
|
| 883 |
+
"learning_rate": 6.981012658227847e-07,
|
| 884 |
+
"loss": 0.2063,
|
| 885 |
+
"step": 584
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.7,
|
| 889 |
+
"learning_rate": 6.955696202531646e-07,
|
| 890 |
+
"loss": 0.2994,
|
| 891 |
+
"step": 588
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.7,
|
| 895 |
+
"learning_rate": 6.930379746835443e-07,
|
| 896 |
+
"loss": 0.3452,
|
| 897 |
+
"step": 592
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 0.71,
|
| 901 |
+
"learning_rate": 6.90506329113924e-07,
|
| 902 |
+
"loss": 0.2579,
|
| 903 |
+
"step": 596
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.71,
|
| 907 |
+
"learning_rate": 6.879746835443038e-07,
|
| 908 |
+
"loss": 0.2208,
|
| 909 |
+
"step": 600
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.72,
|
| 913 |
+
"learning_rate": 6.854430379746835e-07,
|
| 914 |
+
"loss": 0.2463,
|
| 915 |
+
"step": 604
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.72,
|
| 919 |
+
"learning_rate": 6.829113924050632e-07,
|
| 920 |
+
"loss": 0.2835,
|
| 921 |
+
"step": 608
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.73,
|
| 925 |
+
"learning_rate": 6.80379746835443e-07,
|
| 926 |
+
"loss": 0.2034,
|
| 927 |
+
"step": 612
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.73,
|
| 931 |
+
"learning_rate": 6.778481012658228e-07,
|
| 932 |
+
"loss": 0.2464,
|
| 933 |
+
"step": 616
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.74,
|
| 937 |
+
"learning_rate": 6.753164556962025e-07,
|
| 938 |
+
"loss": 0.2682,
|
| 939 |
+
"step": 620
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 0.74,
|
| 943 |
+
"learning_rate": 6.727848101265823e-07,
|
| 944 |
+
"loss": 0.285,
|
| 945 |
+
"step": 624
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 0.75,
|
| 949 |
+
"learning_rate": 6.702531645569621e-07,
|
| 950 |
+
"loss": 0.2276,
|
| 951 |
+
"step": 628
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 0.75,
|
| 955 |
+
"learning_rate": 6.677215189873417e-07,
|
| 956 |
+
"loss": 0.2754,
|
| 957 |
+
"step": 632
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.76,
|
| 961 |
+
"learning_rate": 6.651898734177215e-07,
|
| 962 |
+
"loss": 0.2306,
|
| 963 |
+
"step": 636
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 0.76,
|
| 967 |
+
"learning_rate": 6.626582278481012e-07,
|
| 968 |
+
"loss": 0.2431,
|
| 969 |
+
"step": 640
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.77,
|
| 973 |
+
"learning_rate": 6.601265822784809e-07,
|
| 974 |
+
"loss": 0.2079,
|
| 975 |
+
"step": 644
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.77,
|
| 979 |
+
"learning_rate": 6.575949367088607e-07,
|
| 980 |
+
"loss": 0.2426,
|
| 981 |
+
"step": 648
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 0.78,
|
| 985 |
+
"learning_rate": 6.550632911392405e-07,
|
| 986 |
+
"loss": 0.2854,
|
| 987 |
+
"step": 652
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 0.78,
|
| 991 |
+
"learning_rate": 6.525316455696203e-07,
|
| 992 |
+
"loss": 0.3083,
|
| 993 |
+
"step": 656
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 0.79,
|
| 997 |
+
"learning_rate": 6.5e-07,
|
| 998 |
+
"loss": 0.2305,
|
| 999 |
+
"step": 660
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.79,
|
| 1003 |
+
"learning_rate": 6.474683544303797e-07,
|
| 1004 |
+
"loss": 0.2648,
|
| 1005 |
+
"step": 664
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 0.8,
|
| 1009 |
+
"learning_rate": 6.449367088607595e-07,
|
| 1010 |
+
"loss": 0.2992,
|
| 1011 |
+
"step": 668
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.8,
|
| 1015 |
+
"learning_rate": 6.424050632911392e-07,
|
| 1016 |
+
"loss": 0.3059,
|
| 1017 |
+
"step": 672
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.8,
|
| 1021 |
+
"learning_rate": 6.398734177215189e-07,
|
| 1022 |
+
"loss": 0.2373,
|
| 1023 |
+
"step": 676
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 0.81,
|
| 1027 |
+
"learning_rate": 6.373417721518987e-07,
|
| 1028 |
+
"loss": 0.2808,
|
| 1029 |
+
"step": 680
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.81,
|
| 1033 |
+
"learning_rate": 6.348101265822785e-07,
|
| 1034 |
+
"loss": 0.3125,
|
| 1035 |
+
"step": 684
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.82,
|
| 1039 |
+
"learning_rate": 6.322784810126582e-07,
|
| 1040 |
+
"loss": 0.1994,
|
| 1041 |
+
"step": 688
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.82,
|
| 1045 |
+
"learning_rate": 6.29746835443038e-07,
|
| 1046 |
+
"loss": 0.1612,
|
| 1047 |
+
"step": 692
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 0.83,
|
| 1051 |
+
"learning_rate": 6.272151898734178e-07,
|
| 1052 |
+
"loss": 0.2562,
|
| 1053 |
+
"step": 696
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.83,
|
| 1057 |
+
"learning_rate": 6.246835443037974e-07,
|
| 1058 |
+
"loss": 0.3455,
|
| 1059 |
+
"step": 700
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.84,
|
| 1063 |
+
"learning_rate": 6.221518987341772e-07,
|
| 1064 |
+
"loss": 0.2547,
|
| 1065 |
+
"step": 704
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 0.84,
|
| 1069 |
+
"learning_rate": 6.196202531645569e-07,
|
| 1070 |
+
"loss": 0.2416,
|
| 1071 |
+
"step": 708
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.85,
|
| 1075 |
+
"learning_rate": 6.170886075949366e-07,
|
| 1076 |
+
"loss": 0.2431,
|
| 1077 |
+
"step": 712
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 0.85,
|
| 1081 |
+
"learning_rate": 6.145569620253165e-07,
|
| 1082 |
+
"loss": 0.3004,
|
| 1083 |
+
"step": 716
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.86,
|
| 1087 |
+
"learning_rate": 6.120253164556962e-07,
|
| 1088 |
+
"loss": 0.378,
|
| 1089 |
+
"step": 720
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.86,
|
| 1093 |
+
"learning_rate": 6.094936708860759e-07,
|
| 1094 |
+
"loss": 0.2185,
|
| 1095 |
+
"step": 724
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.87,
|
| 1099 |
+
"learning_rate": 6.069620253164557e-07,
|
| 1100 |
+
"loss": 0.2064,
|
| 1101 |
+
"step": 728
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.87,
|
| 1105 |
+
"learning_rate": 6.044303797468354e-07,
|
| 1106 |
+
"loss": 0.2867,
|
| 1107 |
+
"step": 732
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 0.88,
|
| 1111 |
+
"learning_rate": 6.018987341772151e-07,
|
| 1112 |
+
"loss": 0.178,
|
| 1113 |
+
"step": 736
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 0.88,
|
| 1117 |
+
"learning_rate": 5.993670886075949e-07,
|
| 1118 |
+
"loss": 0.2377,
|
| 1119 |
+
"step": 740
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.89,
|
| 1123 |
+
"learning_rate": 5.968354430379746e-07,
|
| 1124 |
+
"loss": 0.2475,
|
| 1125 |
+
"step": 744
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.89,
|
| 1129 |
+
"learning_rate": 5.943037974683544e-07,
|
| 1130 |
+
"loss": 0.2546,
|
| 1131 |
+
"step": 748
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.9,
|
| 1135 |
+
"learning_rate": 5.917721518987342e-07,
|
| 1136 |
+
"loss": 0.246,
|
| 1137 |
+
"step": 752
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 0.9,
|
| 1141 |
+
"learning_rate": 5.892405063291139e-07,
|
| 1142 |
+
"loss": 0.1832,
|
| 1143 |
+
"step": 756
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.9,
|
| 1147 |
+
"learning_rate": 5.867088607594937e-07,
|
| 1148 |
+
"loss": 0.2818,
|
| 1149 |
+
"step": 760
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 0.91,
|
| 1153 |
+
"learning_rate": 5.841772151898734e-07,
|
| 1154 |
+
"loss": 0.3186,
|
| 1155 |
+
"step": 764
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.91,
|
| 1159 |
+
"learning_rate": 5.816455696202531e-07,
|
| 1160 |
+
"loss": 0.1874,
|
| 1161 |
+
"step": 768
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.92,
|
| 1165 |
+
"learning_rate": 5.791139240506329e-07,
|
| 1166 |
+
"loss": 0.2877,
|
| 1167 |
+
"step": 772
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 0.92,
|
| 1171 |
+
"learning_rate": 5.765822784810126e-07,
|
| 1172 |
+
"loss": 0.2143,
|
| 1173 |
+
"step": 776
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 0.93,
|
| 1177 |
+
"learning_rate": 5.740506329113923e-07,
|
| 1178 |
+
"loss": 0.1791,
|
| 1179 |
+
"step": 780
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.93,
|
| 1183 |
+
"learning_rate": 5.715189873417722e-07,
|
| 1184 |
+
"loss": 0.2619,
|
| 1185 |
+
"step": 784
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.94,
|
| 1189 |
+
"learning_rate": 5.68987341772152e-07,
|
| 1190 |
+
"loss": 0.2721,
|
| 1191 |
+
"step": 788
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 0.94,
|
| 1195 |
+
"learning_rate": 5.664556962025316e-07,
|
| 1196 |
+
"loss": 0.2182,
|
| 1197 |
+
"step": 792
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.95,
|
| 1201 |
+
"learning_rate": 5.639240506329114e-07,
|
| 1202 |
+
"loss": 0.3081,
|
| 1203 |
+
"step": 796
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.95,
|
| 1207 |
+
"learning_rate": 5.613924050632911e-07,
|
| 1208 |
+
"loss": 0.1964,
|
| 1209 |
+
"step": 800
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 0.96,
|
| 1213 |
+
"learning_rate": 5.588607594936708e-07,
|
| 1214 |
+
"loss": 0.221,
|
| 1215 |
+
"step": 804
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 0.96,
|
| 1219 |
+
"learning_rate": 5.563291139240506e-07,
|
| 1220 |
+
"loss": 0.2069,
|
| 1221 |
+
"step": 808
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.97,
|
| 1225 |
+
"learning_rate": 5.537974683544303e-07,
|
| 1226 |
+
"loss": 0.294,
|
| 1227 |
+
"step": 812
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.97,
|
| 1231 |
+
"learning_rate": 5.512658227848101e-07,
|
| 1232 |
+
"loss": 0.1952,
|
| 1233 |
+
"step": 816
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 0.98,
|
| 1237 |
+
"learning_rate": 5.487341772151899e-07,
|
| 1238 |
+
"loss": 0.2268,
|
| 1239 |
+
"step": 820
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 0.98,
|
| 1243 |
+
"learning_rate": 5.462025316455696e-07,
|
| 1244 |
+
"loss": 0.232,
|
| 1245 |
+
"step": 824
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.99,
|
| 1249 |
+
"learning_rate": 5.436708860759493e-07,
|
| 1250 |
+
"loss": 0.2639,
|
| 1251 |
+
"step": 828
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 0.99,
|
| 1255 |
+
"learning_rate": 5.411392405063291e-07,
|
| 1256 |
+
"loss": 0.1895,
|
| 1257 |
+
"step": 832
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 1.0,
|
| 1261 |
+
"learning_rate": 5.386075949367088e-07,
|
| 1262 |
+
"loss": 0.2694,
|
| 1263 |
+
"step": 836
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.0,
|
| 1267 |
+
"learning_rate": 5.360759493670886e-07,
|
| 1268 |
+
"loss": 0.2513,
|
| 1269 |
+
"step": 840
|
| 1270 |
+
}
|
| 1271 |
+
],
|
| 1272 |
+
"logging_steps": 4,
|
| 1273 |
+
"max_steps": 1680,
|
| 1274 |
+
"num_input_tokens_seen": 0,
|
| 1275 |
+
"num_train_epochs": 2,
|
| 1276 |
+
"save_steps": 200,
|
| 1277 |
+
"total_flos": 1.371080632762368e+19,
|
| 1278 |
+
"train_batch_size": 16,
|
| 1279 |
+
"trial_name": null,
|
| 1280 |
+
"trial_params": null
|
| 1281 |
+
}
|
checkpoint-840/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:98168a42100d73ecdd098d767ba790511816ef1b8a26cde7cd1b47ec90c1d580
|
| 3 |
+
size 4984
|