Upload folder using huggingface_hub
Browse files- adapter_config.json +1 -1
- adapter_model.safetensors +1 -1
- optimizer.pt +2 -2
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- special_tokens_map.json +18 -2
- tokenizer.json +18 -0
- tokenizer_config.json +22 -2
- trainer_state.json +103 -403
- training_args.bin +1 -1
adapter_config.json
CHANGED
|
@@ -16,9 +16,9 @@
|
|
| 16 |
"rank_pattern": {},
|
| 17 |
"revision": null,
|
| 18 |
"target_modules": [
|
|
|
|
| 19 |
"dense_4h_to_h",
|
| 20 |
"query_key_value",
|
| 21 |
-
"dense",
|
| 22 |
"dense_h_to_4h"
|
| 23 |
],
|
| 24 |
"task_type": "CAUSAL_LM"
|
|
|
|
| 16 |
"rank_pattern": {},
|
| 17 |
"revision": null,
|
| 18 |
"target_modules": [
|
| 19 |
+
"dense",
|
| 20 |
"dense_4h_to_h",
|
| 21 |
"query_key_value",
|
|
|
|
| 22 |
"dense_h_to_4h"
|
| 23 |
],
|
| 24 |
"task_type": "CAUSAL_LM"
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 134235712
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1a303fc03478413846047c106f42c7364c8588fe0f79ae48b430a8e64620553c
|
| 3 |
size 134235712
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b4556ee4b472013f0fe859c60efc7f003ce618b375db68deb9167d83f13258c5
|
| 3 |
+
size 268514874
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:531fa2d6171a6f0be50b9b00866646dbca057c880008dd16007322fc07d26cba
|
| 3 |
size 14244
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3e3bcbfe9174fb2950786064200b81466459d490f88419b7e74e2f7cd9e9471c
|
| 3 |
size 1064
|
special_tokens_map.json
CHANGED
|
@@ -1,6 +1,22 @@
|
|
| 1 |
{
|
| 2 |
-
"
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
"pad_token": "<|endoftext|>",
|
| 5 |
"unk_token": "<|endoftext|>"
|
| 6 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
{
|
| 4 |
+
"content": "<|im_end|>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
{
|
| 11 |
+
"content": "<|im_start|>",
|
| 12 |
+
"lstrip": false,
|
| 13 |
+
"normalized": false,
|
| 14 |
+
"rstrip": false,
|
| 15 |
+
"single_word": false
|
| 16 |
+
}
|
| 17 |
+
],
|
| 18 |
+
"bos_token": "<|im_start|>",
|
| 19 |
+
"eos_token": "<|im_end|>",
|
| 20 |
"pad_token": "<|endoftext|>",
|
| 21 |
"unk_token": "<|endoftext|>"
|
| 22 |
}
|
tokenizer.json
CHANGED
|
@@ -232,6 +232,24 @@
|
|
| 232 |
"rstrip": false,
|
| 233 |
"normalized": true,
|
| 234 |
"special": false
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 235 |
}
|
| 236 |
],
|
| 237 |
"normalizer": {
|
|
|
|
| 232 |
"rstrip": false,
|
| 233 |
"normalized": true,
|
| 234 |
"special": false
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"id": 50277,
|
| 238 |
+
"content": "<|im_end|>",
|
| 239 |
+
"single_word": false,
|
| 240 |
+
"lstrip": false,
|
| 241 |
+
"rstrip": false,
|
| 242 |
+
"normalized": false,
|
| 243 |
+
"special": true
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"id": 50278,
|
| 247 |
+
"content": "<|im_start|>",
|
| 248 |
+
"single_word": false,
|
| 249 |
+
"lstrip": false,
|
| 250 |
+
"rstrip": false,
|
| 251 |
+
"normalized": false,
|
| 252 |
+
"special": true
|
| 253 |
}
|
| 254 |
],
|
| 255 |
"normalizer": {
|
tokenizer_config.json
CHANGED
|
@@ -200,11 +200,31 @@
|
|
| 200 |
"rstrip": false,
|
| 201 |
"single_word": false,
|
| 202 |
"special": false
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 203 |
}
|
| 204 |
},
|
| 205 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
| 206 |
"clean_up_tokenization_spaces": true,
|
| 207 |
-
"eos_token": "<|
|
| 208 |
"model_max_length": 1000000000000000019884624838656,
|
| 209 |
"pad_token": "<|endoftext|>",
|
| 210 |
"tokenizer_class": "GPTNeoXTokenizer",
|
|
|
|
| 200 |
"rstrip": false,
|
| 201 |
"single_word": false,
|
| 202 |
"special": false
|
| 203 |
+
},
|
| 204 |
+
"50277": {
|
| 205 |
+
"content": "<|im_end|>",
|
| 206 |
+
"lstrip": false,
|
| 207 |
+
"normalized": false,
|
| 208 |
+
"rstrip": false,
|
| 209 |
+
"single_word": false,
|
| 210 |
+
"special": true
|
| 211 |
+
},
|
| 212 |
+
"50278": {
|
| 213 |
+
"content": "<|im_start|>",
|
| 214 |
+
"lstrip": false,
|
| 215 |
+
"normalized": false,
|
| 216 |
+
"rstrip": false,
|
| 217 |
+
"single_word": false,
|
| 218 |
+
"special": true
|
| 219 |
}
|
| 220 |
},
|
| 221 |
+
"additional_special_tokens": [
|
| 222 |
+
"<|im_end|>",
|
| 223 |
+
"<|im_start|>"
|
| 224 |
+
],
|
| 225 |
+
"bos_token": "<|im_start|>",
|
| 226 |
"clean_up_tokenization_spaces": true,
|
| 227 |
+
"eos_token": "<|im_end|>",
|
| 228 |
"model_max_length": 1000000000000000019884624838656,
|
| 229 |
"pad_token": "<|endoftext|>",
|
| 230 |
"tokenizer_class": "GPTNeoXTokenizer",
|
trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -11,909 +11,609 @@
|
|
| 11 |
{
|
| 12 |
"epoch": 0.02,
|
| 13 |
"learning_rate": 4.444444444444444e-06,
|
| 14 |
-
"loss": 2.
|
| 15 |
"step": 2
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.04,
|
| 19 |
"learning_rate": 8.888888888888888e-06,
|
| 20 |
-
"loss": 2.
|
| 21 |
"step": 4
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.06,
|
| 25 |
"learning_rate": 1.3333333333333333e-05,
|
| 26 |
-
"loss": 2.
|
| 27 |
"step": 6
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.08,
|
| 31 |
"learning_rate": 1.7777777777777777e-05,
|
| 32 |
-
"loss": 2.
|
| 33 |
"step": 8
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.1,
|
| 37 |
"learning_rate": 1.9999417253661235e-05,
|
| 38 |
-
"loss": 2.
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.12,
|
| 43 |
"learning_rate": 1.9994755690455154e-05,
|
| 44 |
-
"loss": 2.
|
| 45 |
"step": 12
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.14,
|
| 49 |
"learning_rate": 1.998543473718677e-05,
|
| 50 |
-
"loss": 2.
|
| 51 |
"step": 14
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.16,
|
| 55 |
"learning_rate": 1.9971458739130598e-05,
|
| 56 |
-
"loss": 2.
|
| 57 |
"step": 16
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.18,
|
| 61 |
"learning_rate": 1.995283421166614e-05,
|
| 62 |
-
"loss": 2.
|
| 63 |
"step": 18
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.2,
|
| 67 |
"learning_rate": 1.9929569837240567e-05,
|
| 68 |
-
"loss": 2.
|
| 69 |
"step": 20
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.22,
|
| 73 |
"learning_rate": 1.990167646132107e-05,
|
| 74 |
-
"loss": 2.
|
| 75 |
"step": 22
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.24,
|
| 79 |
"learning_rate": 1.9869167087338908e-05,
|
| 80 |
-
"loss": 2.
|
| 81 |
"step": 24
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.26,
|
| 85 |
"learning_rate": 1.983205687062742e-05,
|
| 86 |
-
"loss": 2.
|
| 87 |
"step": 26
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.28,
|
| 91 |
"learning_rate": 1.9790363111356838e-05,
|
| 92 |
-
"loss": 2.
|
| 93 |
"step": 28
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.3,
|
| 97 |
"learning_rate": 1.9744105246469264e-05,
|
| 98 |
-
"loss": 2.
|
| 99 |
"step": 30
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.32,
|
| 103 |
"learning_rate": 1.9693304840617456e-05,
|
| 104 |
-
"loss": 2.
|
| 105 |
"step": 32
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.34,
|
| 109 |
"learning_rate": 1.963798557611178e-05,
|
| 110 |
-
"loss": 2.
|
| 111 |
"step": 34
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 0.36,
|
| 115 |
"learning_rate": 1.957817324187987e-05,
|
| 116 |
-
"loss": 2.
|
| 117 |
"step": 36
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"epoch": 0.38,
|
| 121 |
"learning_rate": 1.9513895721444286e-05,
|
| 122 |
-
"loss": 2.
|
| 123 |
"step": 38
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.4,
|
| 127 |
"learning_rate": 1.9445182979923657e-05,
|
| 128 |
-
"loss": 2.
|
| 129 |
"step": 40
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.42,
|
| 133 |
"learning_rate": 1.937206705006344e-05,
|
| 134 |
-
"loss": 2.
|
| 135 |
"step": 42
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.44,
|
| 139 |
"learning_rate": 1.9294582017302797e-05,
|
| 140 |
-
"loss": 2.
|
| 141 |
"step": 44
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.46,
|
| 145 |
"learning_rate": 1.921276400388451e-05,
|
| 146 |
-
"loss": 2.
|
| 147 |
"step": 46
|
| 148 |
},
|
| 149 |
{
|
| 150 |
"epoch": 0.48,
|
| 151 |
"learning_rate": 1.9126651152015404e-05,
|
| 152 |
-
"loss": 2.
|
| 153 |
"step": 48
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 0.5,
|
| 157 |
"learning_rate": 1.9036283606085057e-05,
|
| 158 |
-
"loss": 2.
|
| 159 |
"step": 50
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 0.52,
|
| 163 |
"learning_rate": 1.8941703493951163e-05,
|
| 164 |
-
"loss": 2.
|
| 165 |
"step": 52
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 0.54,
|
| 169 |
"learning_rate": 1.8842954907300236e-05,
|
| 170 |
-
"loss": 2.
|
| 171 |
"step": 54
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"epoch": 0.56,
|
| 175 |
"learning_rate": 1.874008388109276e-05,
|
| 176 |
-
"loss": 2.
|
| 177 |
"step": 56
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.58,
|
| 181 |
"learning_rate": 1.863313837210247e-05,
|
| 182 |
-
"loss": 2.
|
| 183 |
"step": 58
|
| 184 |
},
|
| 185 |
{
|
| 186 |
"epoch": 0.6,
|
| 187 |
"learning_rate": 1.8522168236559693e-05,
|
| 188 |
-
"loss": 2.
|
| 189 |
"step": 60
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 0.62,
|
| 193 |
"learning_rate": 1.840722520690921e-05,
|
| 194 |
-
"loss": 2.
|
| 195 |
"step": 62
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 0.64,
|
| 199 |
"learning_rate": 1.8288362867693414e-05,
|
| 200 |
-
"loss": 2.
|
| 201 |
"step": 64
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 0.66,
|
| 205 |
"learning_rate": 1.816563663057211e-05,
|
| 206 |
-
"loss": 2.
|
| 207 |
"step": 66
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 0.68,
|
| 211 |
"learning_rate": 1.8039103708490503e-05,
|
| 212 |
-
"loss": 2.
|
| 213 |
"step": 68
|
| 214 |
},
|
| 215 |
{
|
| 216 |
"epoch": 0.7,
|
| 217 |
"learning_rate": 1.790882308900746e-05,
|
| 218 |
-
"loss": 2.
|
| 219 |
"step": 70
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.72,
|
| 223 |
"learning_rate": 1.7774855506796497e-05,
|
| 224 |
-
"loss": 2.
|
| 225 |
"step": 72
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"epoch": 0.74,
|
| 229 |
"learning_rate": 1.7637263415332272e-05,
|
| 230 |
-
"loss": 2.
|
| 231 |
"step": 74
|
| 232 |
},
|
| 233 |
{
|
| 234 |
"epoch": 0.76,
|
| 235 |
"learning_rate": 1.749611095777581e-05,
|
| 236 |
-
"loss": 2.
|
| 237 |
"step": 76
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 0.78,
|
| 241 |
"learning_rate": 1.7351463937072008e-05,
|
| 242 |
-
"loss": 2.
|
| 243 |
"step": 78
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"epoch": 0.8,
|
| 247 |
"learning_rate": 1.7203389785273402e-05,
|
| 248 |
-
"loss": 2.
|
| 249 |
"step": 80
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.82,
|
| 253 |
"learning_rate": 1.705195753210446e-05,
|
| 254 |
-
"loss": 2.
|
| 255 |
"step": 82
|
| 256 |
},
|
| 257 |
{
|
| 258 |
"epoch": 0.84,
|
| 259 |
"learning_rate": 1.6897237772781046e-05,
|
| 260 |
-
"loss": 2.
|
| 261 |
"step": 84
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.86,
|
| 265 |
"learning_rate": 1.673930263510011e-05,
|
| 266 |
-
"loss": 2.
|
| 267 |
"step": 86
|
| 268 |
},
|
| 269 |
{
|
| 270 |
"epoch": 0.88,
|
| 271 |
"learning_rate": 1.6578225745814907e-05,
|
| 272 |
-
"loss": 2.
|
| 273 |
"step": 88
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 0.89,
|
| 277 |
"learning_rate": 1.6414082196311402e-05,
|
| 278 |
-
"loss": 2.
|
| 279 |
"step": 90
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"epoch": 0.91,
|
| 283 |
"learning_rate": 1.6246948507601915e-05,
|
| 284 |
-
"loss": 2.
|
| 285 |
"step": 92
|
| 286 |
},
|
| 287 |
{
|
| 288 |
"epoch": 0.93,
|
| 289 |
"learning_rate": 1.607690259465229e-05,
|
| 290 |
-
"loss": 2.
|
| 291 |
"step": 94
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 0.95,
|
| 295 |
"learning_rate": 1.5904023730059227e-05,
|
| 296 |
-
"loss": 2.
|
| 297 |
"step": 96
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"epoch": 0.97,
|
| 301 |
"learning_rate": 1.57283925070947e-05,
|
| 302 |
-
"loss": 2.
|
| 303 |
"step": 98
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.99,
|
| 307 |
"learning_rate": 1.55500908021347e-05,
|
| 308 |
-
"loss":
|
| 309 |
"step": 100
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 1.01,
|
| 313 |
"learning_rate": 1.536920173648984e-05,
|
| 314 |
-
"loss": 2.
|
| 315 |
"step": 102
|
| 316 |
},
|
| 317 |
{
|
| 318 |
"epoch": 1.03,
|
| 319 |
"learning_rate": 1.5185809637655548e-05,
|
| 320 |
-
"loss": 2.
|
| 321 |
"step": 104
|
| 322 |
},
|
| 323 |
{
|
| 324 |
"epoch": 1.05,
|
| 325 |
"learning_rate": 1.5000000000000002e-05,
|
| 326 |
-
"loss": 2.
|
| 327 |
"step": 106
|
| 328 |
},
|
| 329 |
{
|
| 330 |
"epoch": 1.07,
|
| 331 |
"learning_rate": 1.4811859444908053e-05,
|
| 332 |
-
"loss": 2.
|
| 333 |
"step": 108
|
| 334 |
},
|
| 335 |
{
|
| 336 |
"epoch": 1.09,
|
| 337 |
"learning_rate": 1.4621475680399771e-05,
|
| 338 |
-
"loss": 2.
|
| 339 |
"step": 110
|
| 340 |
},
|
| 341 |
{
|
| 342 |
"epoch": 1.11,
|
| 343 |
"learning_rate": 1.4428937460242417e-05,
|
| 344 |
-
"loss": 2.
|
| 345 |
"step": 112
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 1.13,
|
| 349 |
"learning_rate": 1.4234334542574906e-05,
|
| 350 |
-
"loss": 2.
|
| 351 |
"step": 114
|
| 352 |
},
|
| 353 |
{
|
| 354 |
"epoch": 1.15,
|
| 355 |
"learning_rate": 1.4037757648064019e-05,
|
| 356 |
-
"loss": 2.
|
| 357 |
"step": 116
|
| 358 |
},
|
| 359 |
{
|
| 360 |
"epoch": 1.17,
|
| 361 |
"learning_rate": 1.3839298417611964e-05,
|
| 362 |
-
"loss": 2.
|
| 363 |
"step": 118
|
| 364 |
},
|
| 365 |
{
|
| 366 |
"epoch": 1.19,
|
| 367 |
"learning_rate": 1.3639049369634878e-05,
|
| 368 |
-
"loss": 2.
|
| 369 |
"step": 120
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 1.21,
|
| 373 |
"learning_rate": 1.3437103856932266e-05,
|
| 374 |
-
"loss": 2.
|
| 375 |
"step": 122
|
| 376 |
},
|
| 377 |
{
|
| 378 |
"epoch": 1.23,
|
| 379 |
"learning_rate": 1.3233556023167487e-05,
|
| 380 |
-
"loss": 2.
|
| 381 |
"step": 124
|
| 382 |
},
|
| 383 |
{
|
| 384 |
"epoch": 1.25,
|
| 385 |
"learning_rate": 1.3028500758979507e-05,
|
| 386 |
-
"loss": 2.
|
| 387 |
"step": 126
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 1.27,
|
| 391 |
"learning_rate": 1.2822033657746478e-05,
|
| 392 |
-
"loss": 2.
|
| 393 |
"step": 128
|
| 394 |
},
|
| 395 |
{
|
| 396 |
"epoch": 1.29,
|
| 397 |
"learning_rate": 1.2614250971021658e-05,
|
| 398 |
-
"loss": 2.
|
| 399 |
"step": 130
|
| 400 |
},
|
| 401 |
{
|
| 402 |
"epoch": 1.31,
|
| 403 |
"learning_rate": 1.2405249563662539e-05,
|
| 404 |
-
"loss": 2.
|
| 405 |
"step": 132
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"epoch": 1.33,
|
| 409 |
"learning_rate": 1.2195126868674052e-05,
|
| 410 |
-
"loss": 2.
|
| 411 |
"step": 134
|
| 412 |
},
|
| 413 |
{
|
| 414 |
"epoch": 1.35,
|
| 415 |
"learning_rate": 1.1983980841786899e-05,
|
| 416 |
-
"loss": 2.
|
| 417 |
"step": 136
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"epoch": 1.37,
|
| 421 |
"learning_rate": 1.177190991579223e-05,
|
| 422 |
-
"loss": 2.
|
| 423 |
"step": 138
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"epoch": 1.39,
|
| 427 |
"learning_rate": 1.1559012954653865e-05,
|
| 428 |
-
"loss": 2.
|
| 429 |
"step": 140
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 1.41,
|
| 433 |
"learning_rate": 1.1345389207419588e-05,
|
| 434 |
-
"loss": 2.
|
| 435 |
"step": 142
|
| 436 |
},
|
| 437 |
{
|
| 438 |
"epoch": 1.43,
|
| 439 |
"learning_rate": 1.1131138261952845e-05,
|
| 440 |
-
"loss": 2.
|
| 441 |
"step": 144
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"epoch": 1.45,
|
| 445 |
"learning_rate": 1.0916359998506549e-05,
|
| 446 |
-
"loss": 2.
|
| 447 |
"step": 146
|
| 448 |
},
|
| 449 |
{
|
| 450 |
"epoch": 1.47,
|
| 451 |
"learning_rate": 1.070115454316054e-05,
|
| 452 |
-
"loss": 2.
|
| 453 |
"step": 148
|
| 454 |
},
|
| 455 |
{
|
| 456 |
"epoch": 1.49,
|
| 457 |
"learning_rate": 1.0485622221144485e-05,
|
| 458 |
-
"loss": 2.
|
| 459 |
"step": 150
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"epoch": 1.51,
|
| 463 |
"learning_rate": 1.0269863510067872e-05,
|
| 464 |
-
"loss": 2.
|
| 465 |
"step": 152
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"epoch": 1.53,
|
| 469 |
"learning_rate": 1.0053978993079046e-05,
|
| 470 |
-
"loss": 2.
|
| 471 |
"step": 154
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 1.55,
|
| 475 |
"learning_rate": 9.838069311974986e-06,
|
| 476 |
-
"loss": 2.
|
| 477 |
"step": 156
|
| 478 |
},
|
| 479 |
{
|
| 480 |
"epoch": 1.57,
|
| 481 |
"learning_rate": 9.622235120283769e-06,
|
| 482 |
-
"loss": 2.
|
| 483 |
"step": 158
|
| 484 |
},
|
| 485 |
{
|
| 486 |
"epoch": 1.59,
|
| 487 |
"learning_rate": 9.406577036341548e-06,
|
| 488 |
-
"loss": 2.
|
| 489 |
"step": 160
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 1.61,
|
| 493 |
"learning_rate": 9.19119559638596e-06,
|
| 494 |
-
"loss": 2.
|
| 495 |
"step": 162
|
| 496 |
},
|
| 497 |
{
|
| 498 |
"epoch": 1.63,
|
| 499 |
"learning_rate": 8.976191207687775e-06,
|
| 500 |
-
"loss": 2.
|
| 501 |
"step": 164
|
| 502 |
},
|
| 503 |
{
|
| 504 |
"epoch": 1.65,
|
| 505 |
"learning_rate": 8.7616641017427e-06,
|
| 506 |
-
"loss": 2.
|
| 507 |
"step": 166
|
| 508 |
},
|
| 509 |
{
|
| 510 |
"epoch": 1.67,
|
| 511 |
"learning_rate": 8.5477142875451e-06,
|
| 512 |
-
"loss": 2.
|
| 513 |
"step": 168
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 1.69,
|
| 517 |
"learning_rate": 8.334441504965456e-06,
|
| 518 |
-
"loss":
|
| 519 |
"step": 170
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 1.71,
|
| 523 |
"learning_rate": 8.1219451782533e-06,
|
| 524 |
-
"loss": 2.
|
| 525 |
"step": 172
|
| 526 |
},
|
| 527 |
{
|
| 528 |
"epoch": 1.73,
|
| 529 |
"learning_rate": 7.91032436968725e-06,
|
| 530 |
-
"loss": 2.
|
| 531 |
"step": 174
|
| 532 |
},
|
| 533 |
{
|
| 534 |
"epoch": 1.75,
|
| 535 |
"learning_rate": 7.699677733393827e-06,
|
| 536 |
-
"loss": 2.
|
| 537 |
"step": 176
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 1.77,
|
| 541 |
"learning_rate": 7.490103469356513e-06,
|
| 542 |
-
"loss": 2.
|
| 543 |
"step": 178
|
| 544 |
},
|
| 545 |
{
|
| 546 |
"epoch": 1.79,
|
| 547 |
"learning_rate": 7.2816992776365714e-06,
|
| 548 |
-
"loss": 2.
|
| 549 |
"step": 180
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"epoch": 1.81,
|
| 553 |
"learning_rate": 7.0745623128268605e-06,
|
| 554 |
-
"loss": 2.
|
| 555 |
"step": 182
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 1.83,
|
| 559 |
"learning_rate": 6.868789138759977e-06,
|
| 560 |
-
"loss": 2.
|
| 561 |
"step": 184
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 1.85,
|
| 565 |
"learning_rate": 6.664475683491797e-06,
|
| 566 |
-
"loss": 2.
|
| 567 |
"step": 186
|
| 568 |
},
|
| 569 |
{
|
| 570 |
"epoch": 1.87,
|
| 571 |
"learning_rate": 6.461717194581394e-06,
|
| 572 |
-
"loss": 2.
|
| 573 |
"step": 188
|
| 574 |
},
|
| 575 |
{
|
| 576 |
"epoch": 1.89,
|
| 577 |
"learning_rate": 6.260608194688207e-06,
|
| 578 |
-
"loss": 2.
|
| 579 |
"step": 190
|
| 580 |
},
|
| 581 |
{
|
| 582 |
"epoch": 1.91,
|
| 583 |
"learning_rate": 6.061242437507131e-06,
|
| 584 |
-
"loss": 2.
|
| 585 |
"step": 192
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 1.93,
|
| 589 |
"learning_rate": 5.863712864062089e-06,
|
| 590 |
-
"loss": 2.
|
| 591 |
"step": 194
|
| 592 |
},
|
| 593 |
{
|
| 594 |
"epoch": 1.95,
|
| 595 |
"learning_rate": 5.6681115593784705e-06,
|
| 596 |
-
"loss": 2.
|
| 597 |
"step": 196
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 1.97,
|
| 601 |
"learning_rate": 5.4745297095546125e-06,
|
| 602 |
-
"loss": 2.
|
| 603 |
"step": 198
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"epoch": 1.99,
|
| 607 |
"learning_rate": 5.2830575592523415e-06,
|
| 608 |
-
"loss":
|
| 609 |
"step": 200
|
| 610 |
-
},
|
| 611 |
-
{
|
| 612 |
-
"epoch": 2.01,
|
| 613 |
-
"learning_rate": 5.093784369626397e-06,
|
| 614 |
-
"loss": 1.9281,
|
| 615 |
-
"step": 202
|
| 616 |
-
},
|
| 617 |
-
{
|
| 618 |
-
"epoch": 2.03,
|
| 619 |
-
"learning_rate": 4.9067983767123736e-06,
|
| 620 |
-
"loss": 2.3076,
|
| 621 |
-
"step": 204
|
| 622 |
-
},
|
| 623 |
-
{
|
| 624 |
-
"epoch": 2.05,
|
| 625 |
-
"learning_rate": 4.722186750292511e-06,
|
| 626 |
-
"loss": 2.3358,
|
| 627 |
-
"step": 206
|
| 628 |
-
},
|
| 629 |
-
{
|
| 630 |
-
"epoch": 2.07,
|
| 631 |
-
"learning_rate": 4.54003555325862e-06,
|
| 632 |
-
"loss": 2.0644,
|
| 633 |
-
"step": 208
|
| 634 |
-
},
|
| 635 |
-
{
|
| 636 |
-
"epoch": 2.09,
|
| 637 |
-
"learning_rate": 4.360429701490935e-06,
|
| 638 |
-
"loss": 2.2204,
|
| 639 |
-
"step": 210
|
| 640 |
-
},
|
| 641 |
-
{
|
| 642 |
-
"epoch": 2.11,
|
| 643 |
-
"learning_rate": 4.183452924271776e-06,
|
| 644 |
-
"loss": 2.1468,
|
| 645 |
-
"step": 212
|
| 646 |
-
},
|
| 647 |
-
{
|
| 648 |
-
"epoch": 2.13,
|
| 649 |
-
"learning_rate": 4.009187725252309e-06,
|
| 650 |
-
"loss": 2.1537,
|
| 651 |
-
"step": 214
|
| 652 |
-
},
|
| 653 |
-
{
|
| 654 |
-
"epoch": 2.15,
|
| 655 |
-
"learning_rate": 3.837715343990727e-06,
|
| 656 |
-
"loss": 2.2264,
|
| 657 |
-
"step": 216
|
| 658 |
-
},
|
| 659 |
-
{
|
| 660 |
-
"epoch": 2.17,
|
| 661 |
-
"learning_rate": 3.669115718079702e-06,
|
| 662 |
-
"loss": 2.2227,
|
| 663 |
-
"step": 218
|
| 664 |
-
},
|
| 665 |
-
{
|
| 666 |
-
"epoch": 2.19,
|
| 667 |
-
"learning_rate": 3.5034674458807893e-06,
|
| 668 |
-
"loss": 2.0947,
|
| 669 |
-
"step": 220
|
| 670 |
-
},
|
| 671 |
-
{
|
| 672 |
-
"epoch": 2.21,
|
| 673 |
-
"learning_rate": 3.3408477498831917e-06,
|
| 674 |
-
"loss": 2.0725,
|
| 675 |
-
"step": 222
|
| 676 |
-
},
|
| 677 |
-
{
|
| 678 |
-
"epoch": 2.23,
|
| 679 |
-
"learning_rate": 3.1813324407038826e-06,
|
| 680 |
-
"loss": 2.0791,
|
| 681 |
-
"step": 224
|
| 682 |
-
},
|
| 683 |
-
{
|
| 684 |
-
"epoch": 2.25,
|
| 685 |
-
"learning_rate": 3.024995881745972e-06,
|
| 686 |
-
"loss": 2.0566,
|
| 687 |
-
"step": 226
|
| 688 |
-
},
|
| 689 |
-
{
|
| 690 |
-
"epoch": 2.27,
|
| 691 |
-
"learning_rate": 2.8719109545317102e-06,
|
| 692 |
-
"loss": 2.1872,
|
| 693 |
-
"step": 228
|
| 694 |
-
},
|
| 695 |
-
{
|
| 696 |
-
"epoch": 2.29,
|
| 697 |
-
"learning_rate": 2.722149024726307e-06,
|
| 698 |
-
"loss": 2.2038,
|
| 699 |
-
"step": 230
|
| 700 |
-
},
|
| 701 |
-
{
|
| 702 |
-
"epoch": 2.31,
|
| 703 |
-
"learning_rate": 2.5757799088684654e-06,
|
| 704 |
-
"loss": 2.206,
|
| 705 |
-
"step": 232
|
| 706 |
-
},
|
| 707 |
-
{
|
| 708 |
-
"epoch": 2.33,
|
| 709 |
-
"learning_rate": 2.432871841823047e-06,
|
| 710 |
-
"loss": 2.225,
|
| 711 |
-
"step": 234
|
| 712 |
-
},
|
| 713 |
-
{
|
| 714 |
-
"epoch": 2.35,
|
| 715 |
-
"learning_rate": 2.293491444971109e-06,
|
| 716 |
-
"loss": 2.2296,
|
| 717 |
-
"step": 236
|
| 718 |
-
},
|
| 719 |
-
{
|
| 720 |
-
"epoch": 2.37,
|
| 721 |
-
"learning_rate": 2.157703695152109e-06,
|
| 722 |
-
"loss": 2.3258,
|
| 723 |
-
"step": 238
|
| 724 |
-
},
|
| 725 |
-
{
|
| 726 |
-
"epoch": 2.39,
|
| 727 |
-
"learning_rate": 2.025571894372794e-06,
|
| 728 |
-
"loss": 2.214,
|
| 729 |
-
"step": 240
|
| 730 |
-
},
|
| 731 |
-
{
|
| 732 |
-
"epoch": 2.41,
|
| 733 |
-
"learning_rate": 1.897157640296825e-06,
|
| 734 |
-
"loss": 2.2363,
|
| 735 |
-
"step": 242
|
| 736 |
-
},
|
| 737 |
-
{
|
| 738 |
-
"epoch": 2.43,
|
| 739 |
-
"learning_rate": 1.7725207975289883e-06,
|
| 740 |
-
"loss": 2.0114,
|
| 741 |
-
"step": 244
|
| 742 |
-
},
|
| 743 |
-
{
|
| 744 |
-
"epoch": 2.45,
|
| 745 |
-
"learning_rate": 1.6517194697072903e-06,
|
| 746 |
-
"loss": 2.1286,
|
| 747 |
-
"step": 246
|
| 748 |
-
},
|
| 749 |
-
{
|
| 750 |
-
"epoch": 2.47,
|
| 751 |
-
"learning_rate": 1.534809972415998e-06,
|
| 752 |
-
"loss": 2.0696,
|
| 753 |
-
"step": 248
|
| 754 |
-
},
|
| 755 |
-
{
|
| 756 |
-
"epoch": 2.49,
|
| 757 |
-
"learning_rate": 1.4218468069322576e-06,
|
| 758 |
-
"loss": 2.1258,
|
| 759 |
-
"step": 250
|
| 760 |
-
},
|
| 761 |
-
{
|
| 762 |
-
"epoch": 2.51,
|
| 763 |
-
"learning_rate": 1.3128826348184886e-06,
|
| 764 |
-
"loss": 1.9143,
|
| 765 |
-
"step": 252
|
| 766 |
-
},
|
| 767 |
-
{
|
| 768 |
-
"epoch": 2.53,
|
| 769 |
-
"learning_rate": 1.207968253372438e-06,
|
| 770 |
-
"loss": 2.197,
|
| 771 |
-
"step": 254
|
| 772 |
-
},
|
| 773 |
-
{
|
| 774 |
-
"epoch": 2.55,
|
| 775 |
-
"learning_rate": 1.1071525719463094e-06,
|
| 776 |
-
"loss": 2.3897,
|
| 777 |
-
"step": 256
|
| 778 |
-
},
|
| 779 |
-
{
|
| 780 |
-
"epoch": 2.57,
|
| 781 |
-
"learning_rate": 1.010482589146048e-06,
|
| 782 |
-
"loss": 2.2106,
|
| 783 |
-
"step": 258
|
| 784 |
-
},
|
| 785 |
-
{
|
| 786 |
-
"epoch": 2.59,
|
| 787 |
-
"learning_rate": 9.180033709213454e-07,
|
| 788 |
-
"loss": 2.2753,
|
| 789 |
-
"step": 260
|
| 790 |
-
},
|
| 791 |
-
{
|
| 792 |
-
"epoch": 2.61,
|
| 793 |
-
"learning_rate": 8.297580295566576e-07,
|
| 794 |
-
"loss": 2.1057,
|
| 795 |
-
"step": 262
|
| 796 |
-
},
|
| 797 |
-
{
|
| 798 |
-
"epoch": 2.63,
|
| 799 |
-
"learning_rate": 7.457877035729588e-07,
|
| 800 |
-
"loss": 2.2365,
|
| 801 |
-
"step": 264
|
| 802 |
-
},
|
| 803 |
-
{
|
| 804 |
-
"epoch": 2.65,
|
| 805 |
-
"learning_rate": 6.661315385496426e-07,
|
| 806 |
-
"loss": 2.2341,
|
| 807 |
-
"step": 266
|
| 808 |
-
},
|
| 809 |
-
{
|
| 810 |
-
"epoch": 2.67,
|
| 811 |
-
"learning_rate": 5.908266688755049e-07,
|
| 812 |
-
"loss": 2.1549,
|
| 813 |
-
"step": 268
|
| 814 |
-
},
|
| 815 |
-
{
|
| 816 |
-
"epoch": 2.68,
|
| 817 |
-
"learning_rate": 5.199082004372958e-07,
|
| 818 |
-
"loss": 2.2376,
|
| 819 |
-
"step": 270
|
| 820 |
-
},
|
| 821 |
-
{
|
| 822 |
-
"epoch": 2.7,
|
| 823 |
-
"learning_rate": 4.534091942539476e-07,
|
| 824 |
-
"loss": 2.0085,
|
| 825 |
-
"step": 272
|
| 826 |
-
},
|
| 827 |
-
{
|
| 828 |
-
"epoch": 2.72,
|
| 829 |
-
"learning_rate": 3.913606510640644e-07,
|
| 830 |
-
"loss": 2.1783,
|
| 831 |
-
"step": 274
|
| 832 |
-
},
|
| 833 |
-
{
|
| 834 |
-
"epoch": 2.74,
|
| 835 |
-
"learning_rate": 3.3379149687388866e-07,
|
| 836 |
-
"loss": 2.0287,
|
| 837 |
-
"step": 276
|
| 838 |
-
},
|
| 839 |
-
{
|
| 840 |
-
"epoch": 2.76,
|
| 841 |
-
"learning_rate": 2.807285694724804e-07,
|
| 842 |
-
"loss": 2.2411,
|
| 843 |
-
"step": 278
|
| 844 |
-
},
|
| 845 |
-
{
|
| 846 |
-
"epoch": 2.78,
|
| 847 |
-
"learning_rate": 2.3219660592038285e-07,
|
| 848 |
-
"loss": 2.1422,
|
| 849 |
-
"step": 280
|
| 850 |
-
},
|
| 851 |
-
{
|
| 852 |
-
"epoch": 2.8,
|
| 853 |
-
"learning_rate": 1.8821823101760949e-07,
|
| 854 |
-
"loss": 2.1875,
|
| 855 |
-
"step": 282
|
| 856 |
-
},
|
| 857 |
-
{
|
| 858 |
-
"epoch": 2.82,
|
| 859 |
-
"learning_rate": 1.4881394675633543e-07,
|
| 860 |
-
"loss": 2.0923,
|
| 861 |
-
"step": 284
|
| 862 |
-
},
|
| 863 |
-
{
|
| 864 |
-
"epoch": 2.84,
|
| 865 |
-
"learning_rate": 1.1400212276321377e-07,
|
| 866 |
-
"loss": 2.3118,
|
| 867 |
-
"step": 286
|
| 868 |
-
},
|
| 869 |
-
{
|
| 870 |
-
"epoch": 2.86,
|
| 871 |
-
"learning_rate": 8.379898773574924e-08,
|
| 872 |
-
"loss": 2.3483,
|
| 873 |
-
"step": 288
|
| 874 |
-
},
|
| 875 |
-
{
|
| 876 |
-
"epoch": 2.88,
|
| 877 |
-
"learning_rate": 5.821862187675775e-08,
|
| 878 |
-
"loss": 2.2267,
|
| 879 |
-
"step": 290
|
| 880 |
-
},
|
| 881 |
-
{
|
| 882 |
-
"epoch": 2.9,
|
| 883 |
-
"learning_rate": 3.727295033040035e-08,
|
| 884 |
-
"loss": 2.2736,
|
| 885 |
-
"step": 292
|
| 886 |
-
},
|
| 887 |
-
{
|
| 888 |
-
"epoch": 2.92,
|
| 889 |
-
"learning_rate": 2.0971737622883515e-08,
|
| 890 |
-
"loss": 1.9528,
|
| 891 |
-
"step": 294
|
| 892 |
-
},
|
| 893 |
-
{
|
| 894 |
-
"epoch": 2.94,
|
| 895 |
-
"learning_rate": 9.322583110392692e-09,
|
| 896 |
-
"loss": 1.9981,
|
| 897 |
-
"step": 296
|
| 898 |
-
},
|
| 899 |
-
{
|
| 900 |
-
"epoch": 2.96,
|
| 901 |
-
"learning_rate": 2.330917436402791e-09,
|
| 902 |
-
"loss": 2.0332,
|
| 903 |
-
"step": 298
|
| 904 |
-
},
|
| 905 |
-
{
|
| 906 |
-
"epoch": 2.98,
|
| 907 |
-
"learning_rate": 0.0,
|
| 908 |
-
"loss": 2.0429,
|
| 909 |
-
"step": 300
|
| 910 |
}
|
| 911 |
],
|
| 912 |
"logging_steps": 2,
|
| 913 |
"max_steps": 300,
|
| 914 |
"num_train_epochs": 3,
|
| 915 |
"save_steps": 500,
|
| 916 |
-
"total_flos":
|
| 917 |
"trial_name": null,
|
| 918 |
"trial_params": null
|
| 919 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.9987569919204475,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 201,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 11 |
{
|
| 12 |
"epoch": 0.02,
|
| 13 |
"learning_rate": 4.444444444444444e-06,
|
| 14 |
+
"loss": 2.687,
|
| 15 |
"step": 2
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.04,
|
| 19 |
"learning_rate": 8.888888888888888e-06,
|
| 20 |
+
"loss": 2.7029,
|
| 21 |
"step": 4
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.06,
|
| 25 |
"learning_rate": 1.3333333333333333e-05,
|
| 26 |
+
"loss": 2.6189,
|
| 27 |
"step": 6
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.08,
|
| 31 |
"learning_rate": 1.7777777777777777e-05,
|
| 32 |
+
"loss": 2.5673,
|
| 33 |
"step": 8
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.1,
|
| 37 |
"learning_rate": 1.9999417253661235e-05,
|
| 38 |
+
"loss": 2.6342,
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.12,
|
| 43 |
"learning_rate": 1.9994755690455154e-05,
|
| 44 |
+
"loss": 2.7195,
|
| 45 |
"step": 12
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.14,
|
| 49 |
"learning_rate": 1.998543473718677e-05,
|
| 50 |
+
"loss": 2.6762,
|
| 51 |
"step": 14
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.16,
|
| 55 |
"learning_rate": 1.9971458739130598e-05,
|
| 56 |
+
"loss": 2.8015,
|
| 57 |
"step": 16
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.18,
|
| 61 |
"learning_rate": 1.995283421166614e-05,
|
| 62 |
+
"loss": 2.4821,
|
| 63 |
"step": 18
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.2,
|
| 67 |
"learning_rate": 1.9929569837240567e-05,
|
| 68 |
+
"loss": 2.3197,
|
| 69 |
"step": 20
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.22,
|
| 73 |
"learning_rate": 1.990167646132107e-05,
|
| 74 |
+
"loss": 2.4756,
|
| 75 |
"step": 22
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.24,
|
| 79 |
"learning_rate": 1.9869167087338908e-05,
|
| 80 |
+
"loss": 2.5277,
|
| 81 |
"step": 24
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.26,
|
| 85 |
"learning_rate": 1.983205687062742e-05,
|
| 86 |
+
"loss": 2.7706,
|
| 87 |
"step": 26
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.28,
|
| 91 |
"learning_rate": 1.9790363111356838e-05,
|
| 92 |
+
"loss": 2.6553,
|
| 93 |
"step": 28
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.3,
|
| 97 |
"learning_rate": 1.9744105246469264e-05,
|
| 98 |
+
"loss": 2.6354,
|
| 99 |
"step": 30
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.32,
|
| 103 |
"learning_rate": 1.9693304840617456e-05,
|
| 104 |
+
"loss": 2.4762,
|
| 105 |
"step": 32
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.34,
|
| 109 |
"learning_rate": 1.963798557611178e-05,
|
| 110 |
+
"loss": 2.6096,
|
| 111 |
"step": 34
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 0.36,
|
| 115 |
"learning_rate": 1.957817324187987e-05,
|
| 116 |
+
"loss": 2.4001,
|
| 117 |
"step": 36
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"epoch": 0.38,
|
| 121 |
"learning_rate": 1.9513895721444286e-05,
|
| 122 |
+
"loss": 2.4426,
|
| 123 |
"step": 38
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.4,
|
| 127 |
"learning_rate": 1.9445182979923657e-05,
|
| 128 |
+
"loss": 2.5961,
|
| 129 |
"step": 40
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.42,
|
| 133 |
"learning_rate": 1.937206705006344e-05,
|
| 134 |
+
"loss": 2.5445,
|
| 135 |
"step": 42
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.44,
|
| 139 |
"learning_rate": 1.9294582017302797e-05,
|
| 140 |
+
"loss": 2.4417,
|
| 141 |
"step": 44
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.46,
|
| 145 |
"learning_rate": 1.921276400388451e-05,
|
| 146 |
+
"loss": 2.4769,
|
| 147 |
"step": 46
|
| 148 |
},
|
| 149 |
{
|
| 150 |
"epoch": 0.48,
|
| 151 |
"learning_rate": 1.9126651152015404e-05,
|
| 152 |
+
"loss": 2.532,
|
| 153 |
"step": 48
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 0.5,
|
| 157 |
"learning_rate": 1.9036283606085057e-05,
|
| 158 |
+
"loss": 2.698,
|
| 159 |
"step": 50
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 0.52,
|
| 163 |
"learning_rate": 1.8941703493951163e-05,
|
| 164 |
+
"loss": 2.5106,
|
| 165 |
"step": 52
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 0.54,
|
| 169 |
"learning_rate": 1.8842954907300236e-05,
|
| 170 |
+
"loss": 2.4174,
|
| 171 |
"step": 54
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"epoch": 0.56,
|
| 175 |
"learning_rate": 1.874008388109276e-05,
|
| 176 |
+
"loss": 2.4691,
|
| 177 |
"step": 56
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.58,
|
| 181 |
"learning_rate": 1.863313837210247e-05,
|
| 182 |
+
"loss": 2.3981,
|
| 183 |
"step": 58
|
| 184 |
},
|
| 185 |
{
|
| 186 |
"epoch": 0.6,
|
| 187 |
"learning_rate": 1.8522168236559693e-05,
|
| 188 |
+
"loss": 2.5144,
|
| 189 |
"step": 60
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 0.62,
|
| 193 |
"learning_rate": 1.840722520690921e-05,
|
| 194 |
+
"loss": 2.3611,
|
| 195 |
"step": 62
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 0.64,
|
| 199 |
"learning_rate": 1.8288362867693414e-05,
|
| 200 |
+
"loss": 2.4442,
|
| 201 |
"step": 64
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 0.66,
|
| 205 |
"learning_rate": 1.816563663057211e-05,
|
| 206 |
+
"loss": 2.4351,
|
| 207 |
"step": 66
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 0.68,
|
| 211 |
"learning_rate": 1.8039103708490503e-05,
|
| 212 |
+
"loss": 2.3272,
|
| 213 |
"step": 68
|
| 214 |
},
|
| 215 |
{
|
| 216 |
"epoch": 0.7,
|
| 217 |
"learning_rate": 1.790882308900746e-05,
|
| 218 |
+
"loss": 2.3544,
|
| 219 |
"step": 70
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.72,
|
| 223 |
"learning_rate": 1.7774855506796497e-05,
|
| 224 |
+
"loss": 2.3661,
|
| 225 |
"step": 72
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"epoch": 0.74,
|
| 229 |
"learning_rate": 1.7637263415332272e-05,
|
| 230 |
+
"loss": 2.4607,
|
| 231 |
"step": 74
|
| 232 |
},
|
| 233 |
{
|
| 234 |
"epoch": 0.76,
|
| 235 |
"learning_rate": 1.749611095777581e-05,
|
| 236 |
+
"loss": 2.4305,
|
| 237 |
"step": 76
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 0.78,
|
| 241 |
"learning_rate": 1.7351463937072008e-05,
|
| 242 |
+
"loss": 2.5995,
|
| 243 |
"step": 78
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"epoch": 0.8,
|
| 247 |
"learning_rate": 1.7203389785273402e-05,
|
| 248 |
+
"loss": 2.3701,
|
| 249 |
"step": 80
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.82,
|
| 253 |
"learning_rate": 1.705195753210446e-05,
|
| 254 |
+
"loss": 2.3289,
|
| 255 |
"step": 82
|
| 256 |
},
|
| 257 |
{
|
| 258 |
"epoch": 0.84,
|
| 259 |
"learning_rate": 1.6897237772781046e-05,
|
| 260 |
+
"loss": 2.4084,
|
| 261 |
"step": 84
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.86,
|
| 265 |
"learning_rate": 1.673930263510011e-05,
|
| 266 |
+
"loss": 2.3949,
|
| 267 |
"step": 86
|
| 268 |
},
|
| 269 |
{
|
| 270 |
"epoch": 0.88,
|
| 271 |
"learning_rate": 1.6578225745814907e-05,
|
| 272 |
+
"loss": 2.4651,
|
| 273 |
"step": 88
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 0.89,
|
| 277 |
"learning_rate": 1.6414082196311402e-05,
|
| 278 |
+
"loss": 2.4417,
|
| 279 |
"step": 90
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"epoch": 0.91,
|
| 283 |
"learning_rate": 1.6246948507601915e-05,
|
| 284 |
+
"loss": 2.3136,
|
| 285 |
"step": 92
|
| 286 |
},
|
| 287 |
{
|
| 288 |
"epoch": 0.93,
|
| 289 |
"learning_rate": 1.607690259465229e-05,
|
| 290 |
+
"loss": 2.2265,
|
| 291 |
"step": 94
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 0.95,
|
| 295 |
"learning_rate": 1.5904023730059227e-05,
|
| 296 |
+
"loss": 2.4076,
|
| 297 |
"step": 96
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"epoch": 0.97,
|
| 301 |
"learning_rate": 1.57283925070947e-05,
|
| 302 |
+
"loss": 2.3468,
|
| 303 |
"step": 98
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.99,
|
| 307 |
"learning_rate": 1.55500908021347e-05,
|
| 308 |
+
"loss": 2.2458,
|
| 309 |
"step": 100
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 1.01,
|
| 313 |
"learning_rate": 1.536920173648984e-05,
|
| 314 |
+
"loss": 2.2984,
|
| 315 |
"step": 102
|
| 316 |
},
|
| 317 |
{
|
| 318 |
"epoch": 1.03,
|
| 319 |
"learning_rate": 1.5185809637655548e-05,
|
| 320 |
+
"loss": 2.3769,
|
| 321 |
"step": 104
|
| 322 |
},
|
| 323 |
{
|
| 324 |
"epoch": 1.05,
|
| 325 |
"learning_rate": 1.5000000000000002e-05,
|
| 326 |
+
"loss": 2.284,
|
| 327 |
"step": 106
|
| 328 |
},
|
| 329 |
{
|
| 330 |
"epoch": 1.07,
|
| 331 |
"learning_rate": 1.4811859444908053e-05,
|
| 332 |
+
"loss": 2.4381,
|
| 333 |
"step": 108
|
| 334 |
},
|
| 335 |
{
|
| 336 |
"epoch": 1.09,
|
| 337 |
"learning_rate": 1.4621475680399771e-05,
|
| 338 |
+
"loss": 2.409,
|
| 339 |
"step": 110
|
| 340 |
},
|
| 341 |
{
|
| 342 |
"epoch": 1.11,
|
| 343 |
"learning_rate": 1.4428937460242417e-05,
|
| 344 |
+
"loss": 2.3612,
|
| 345 |
"step": 112
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 1.13,
|
| 349 |
"learning_rate": 1.4234334542574906e-05,
|
| 350 |
+
"loss": 2.3199,
|
| 351 |
"step": 114
|
| 352 |
},
|
| 353 |
{
|
| 354 |
"epoch": 1.15,
|
| 355 |
"learning_rate": 1.4037757648064019e-05,
|
| 356 |
+
"loss": 2.2612,
|
| 357 |
"step": 116
|
| 358 |
},
|
| 359 |
{
|
| 360 |
"epoch": 1.17,
|
| 361 |
"learning_rate": 1.3839298417611964e-05,
|
| 362 |
+
"loss": 2.2587,
|
| 363 |
"step": 118
|
| 364 |
},
|
| 365 |
{
|
| 366 |
"epoch": 1.19,
|
| 367 |
"learning_rate": 1.3639049369634878e-05,
|
| 368 |
+
"loss": 2.2673,
|
| 369 |
"step": 120
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 1.21,
|
| 373 |
"learning_rate": 1.3437103856932266e-05,
|
| 374 |
+
"loss": 2.2712,
|
| 375 |
"step": 122
|
| 376 |
},
|
| 377 |
{
|
| 378 |
"epoch": 1.23,
|
| 379 |
"learning_rate": 1.3233556023167487e-05,
|
| 380 |
+
"loss": 2.2305,
|
| 381 |
"step": 124
|
| 382 |
},
|
| 383 |
{
|
| 384 |
"epoch": 1.25,
|
| 385 |
"learning_rate": 1.3028500758979507e-05,
|
| 386 |
+
"loss": 2.4667,
|
| 387 |
"step": 126
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 1.27,
|
| 391 |
"learning_rate": 1.2822033657746478e-05,
|
| 392 |
+
"loss": 2.4123,
|
| 393 |
"step": 128
|
| 394 |
},
|
| 395 |
{
|
| 396 |
"epoch": 1.29,
|
| 397 |
"learning_rate": 1.2614250971021658e-05,
|
| 398 |
+
"loss": 2.2408,
|
| 399 |
"step": 130
|
| 400 |
},
|
| 401 |
{
|
| 402 |
"epoch": 1.31,
|
| 403 |
"learning_rate": 1.2405249563662539e-05,
|
| 404 |
+
"loss": 2.3962,
|
| 405 |
"step": 132
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"epoch": 1.33,
|
| 409 |
"learning_rate": 1.2195126868674052e-05,
|
| 410 |
+
"loss": 2.3659,
|
| 411 |
"step": 134
|
| 412 |
},
|
| 413 |
{
|
| 414 |
"epoch": 1.35,
|
| 415 |
"learning_rate": 1.1983980841786899e-05,
|
| 416 |
+
"loss": 2.4564,
|
| 417 |
"step": 136
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"epoch": 1.37,
|
| 421 |
"learning_rate": 1.177190991579223e-05,
|
| 422 |
+
"loss": 2.2792,
|
| 423 |
"step": 138
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"epoch": 1.39,
|
| 427 |
"learning_rate": 1.1559012954653865e-05,
|
| 428 |
+
"loss": 2.2928,
|
| 429 |
"step": 140
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 1.41,
|
| 433 |
"learning_rate": 1.1345389207419588e-05,
|
| 434 |
+
"loss": 2.4787,
|
| 435 |
"step": 142
|
| 436 |
},
|
| 437 |
{
|
| 438 |
"epoch": 1.43,
|
| 439 |
"learning_rate": 1.1131138261952845e-05,
|
| 440 |
+
"loss": 2.1273,
|
| 441 |
"step": 144
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"epoch": 1.45,
|
| 445 |
"learning_rate": 1.0916359998506549e-05,
|
| 446 |
+
"loss": 2.1601,
|
| 447 |
"step": 146
|
| 448 |
},
|
| 449 |
{
|
| 450 |
"epoch": 1.47,
|
| 451 |
"learning_rate": 1.070115454316054e-05,
|
| 452 |
+
"loss": 2.2203,
|
| 453 |
"step": 148
|
| 454 |
},
|
| 455 |
{
|
| 456 |
"epoch": 1.49,
|
| 457 |
"learning_rate": 1.0485622221144485e-05,
|
| 458 |
+
"loss": 2.3428,
|
| 459 |
"step": 150
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"epoch": 1.51,
|
| 463 |
"learning_rate": 1.0269863510067872e-05,
|
| 464 |
+
"loss": 2.2332,
|
| 465 |
"step": 152
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"epoch": 1.53,
|
| 469 |
"learning_rate": 1.0053978993079046e-05,
|
| 470 |
+
"loss": 2.4753,
|
| 471 |
"step": 154
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 1.55,
|
| 475 |
"learning_rate": 9.838069311974986e-06,
|
| 476 |
+
"loss": 2.3579,
|
| 477 |
"step": 156
|
| 478 |
},
|
| 479 |
{
|
| 480 |
"epoch": 1.57,
|
| 481 |
"learning_rate": 9.622235120283769e-06,
|
| 482 |
+
"loss": 2.2148,
|
| 483 |
"step": 158
|
| 484 |
},
|
| 485 |
{
|
| 486 |
"epoch": 1.59,
|
| 487 |
"learning_rate": 9.406577036341548e-06,
|
| 488 |
+
"loss": 2.3775,
|
| 489 |
"step": 160
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 1.61,
|
| 493 |
"learning_rate": 9.19119559638596e-06,
|
| 494 |
+
"loss": 2.321,
|
| 495 |
"step": 162
|
| 496 |
},
|
| 497 |
{
|
| 498 |
"epoch": 1.63,
|
| 499 |
"learning_rate": 8.976191207687775e-06,
|
| 500 |
+
"loss": 2.2917,
|
| 501 |
"step": 164
|
| 502 |
},
|
| 503 |
{
|
| 504 |
"epoch": 1.65,
|
| 505 |
"learning_rate": 8.7616641017427e-06,
|
| 506 |
+
"loss": 2.191,
|
| 507 |
"step": 166
|
| 508 |
},
|
| 509 |
{
|
| 510 |
"epoch": 1.67,
|
| 511 |
"learning_rate": 8.5477142875451e-06,
|
| 512 |
+
"loss": 2.3916,
|
| 513 |
"step": 168
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 1.69,
|
| 517 |
"learning_rate": 8.334441504965456e-06,
|
| 518 |
+
"loss": 2.0293,
|
| 519 |
"step": 170
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 1.71,
|
| 523 |
"learning_rate": 8.1219451782533e-06,
|
| 524 |
+
"loss": 2.2324,
|
| 525 |
"step": 172
|
| 526 |
},
|
| 527 |
{
|
| 528 |
"epoch": 1.73,
|
| 529 |
"learning_rate": 7.91032436968725e-06,
|
| 530 |
+
"loss": 2.3022,
|
| 531 |
"step": 174
|
| 532 |
},
|
| 533 |
{
|
| 534 |
"epoch": 1.75,
|
| 535 |
"learning_rate": 7.699677733393827e-06,
|
| 536 |
+
"loss": 2.4032,
|
| 537 |
"step": 176
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 1.77,
|
| 541 |
"learning_rate": 7.490103469356513e-06,
|
| 542 |
+
"loss": 2.3939,
|
| 543 |
"step": 178
|
| 544 |
},
|
| 545 |
{
|
| 546 |
"epoch": 1.79,
|
| 547 |
"learning_rate": 7.2816992776365714e-06,
|
| 548 |
+
"loss": 2.318,
|
| 549 |
"step": 180
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"epoch": 1.81,
|
| 553 |
"learning_rate": 7.0745623128268605e-06,
|
| 554 |
+
"loss": 2.4244,
|
| 555 |
"step": 182
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 1.83,
|
| 559 |
"learning_rate": 6.868789138759977e-06,
|
| 560 |
+
"loss": 2.4398,
|
| 561 |
"step": 184
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 1.85,
|
| 565 |
"learning_rate": 6.664475683491797e-06,
|
| 566 |
+
"loss": 2.2545,
|
| 567 |
"step": 186
|
| 568 |
},
|
| 569 |
{
|
| 570 |
"epoch": 1.87,
|
| 571 |
"learning_rate": 6.461717194581394e-06,
|
| 572 |
+
"loss": 2.3647,
|
| 573 |
"step": 188
|
| 574 |
},
|
| 575 |
{
|
| 576 |
"epoch": 1.89,
|
| 577 |
"learning_rate": 6.260608194688207e-06,
|
| 578 |
+
"loss": 2.3776,
|
| 579 |
"step": 190
|
| 580 |
},
|
| 581 |
{
|
| 582 |
"epoch": 1.91,
|
| 583 |
"learning_rate": 6.061242437507131e-06,
|
| 584 |
+
"loss": 2.2619,
|
| 585 |
"step": 192
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 1.93,
|
| 589 |
"learning_rate": 5.863712864062089e-06,
|
| 590 |
+
"loss": 2.1638,
|
| 591 |
"step": 194
|
| 592 |
},
|
| 593 |
{
|
| 594 |
"epoch": 1.95,
|
| 595 |
"learning_rate": 5.6681115593784705e-06,
|
| 596 |
+
"loss": 2.2278,
|
| 597 |
"step": 196
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 1.97,
|
| 601 |
"learning_rate": 5.4745297095546125e-06,
|
| 602 |
+
"loss": 2.2256,
|
| 603 |
"step": 198
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"epoch": 1.99,
|
| 607 |
"learning_rate": 5.2830575592523415e-06,
|
| 608 |
+
"loss": 2.1212,
|
| 609 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 610 |
}
|
| 611 |
],
|
| 612 |
"logging_steps": 2,
|
| 613 |
"max_steps": 300,
|
| 614 |
"num_train_epochs": 3,
|
| 615 |
"save_steps": 500,
|
| 616 |
+
"total_flos": 1.734000020127744e+16,
|
| 617 |
"trial_name": null,
|
| 618 |
"trial_params": null
|
| 619 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4600
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e15653e523dcd20b017e26f004b5148445019f8bee34ceb24ad06ffc2133c816
|
| 3 |
size 4600
|