🍻 cheers
Browse files- all_results.json +12 -0
- eval_results.json +8 -0
- runs/Jul23_12-34-19_bernard/events.out.tfevents.1658572919.bernard.2664185.2 +3 -0
- train_results.json +7 -0
- trainer_state.json +319 -0
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_accuracy": 0.5076923076923077,
|
| 4 |
+
"eval_loss": 2.5673868656158447,
|
| 5 |
+
"eval_runtime": 30.9749,
|
| 6 |
+
"eval_samples_per_second": 54.56,
|
| 7 |
+
"eval_steps_per_second": 6.844,
|
| 8 |
+
"train_loss": 3.3443971855695858,
|
| 9 |
+
"train_runtime": 396.8195,
|
| 10 |
+
"train_samples_per_second": 17.338,
|
| 11 |
+
"train_steps_per_second": 1.084
|
| 12 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_accuracy": 0.5076923076923077,
|
| 4 |
+
"eval_loss": 2.5673868656158447,
|
| 5 |
+
"eval_runtime": 30.9749,
|
| 6 |
+
"eval_samples_per_second": 54.56,
|
| 7 |
+
"eval_steps_per_second": 6.844
|
| 8 |
+
}
|
runs/Jul23_12-34-19_bernard/events.out.tfevents.1658572919.bernard.2664185.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:336d56598451455a3d1ed20dcb410c214ff910e51ca2ad9907982b562218bd9d
|
| 3 |
+
size 363
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 3.3443971855695858,
|
| 4 |
+
"train_runtime": 396.8195,
|
| 5 |
+
"train_samples_per_second": 17.338,
|
| 6 |
+
"train_steps_per_second": 1.084
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,319 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 2.5673868656158447,
|
| 3 |
+
"best_model_checkpoint": "./robot22/checkpoint-400",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"global_step": 430,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 0.00019534883720930232,
|
| 13 |
+
"loss": 4.9495,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 0.00019069767441860466,
|
| 19 |
+
"loss": 4.8535,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 0.000186046511627907,
|
| 25 |
+
"loss": 4.6491,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.09,
|
| 30 |
+
"learning_rate": 0.0001813953488372093,
|
| 31 |
+
"loss": 4.5284,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.12,
|
| 36 |
+
"learning_rate": 0.00017674418604651164,
|
| 37 |
+
"loss": 4.4311,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.14,
|
| 42 |
+
"learning_rate": 0.00017209302325581395,
|
| 43 |
+
"loss": 4.2551,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.16,
|
| 48 |
+
"learning_rate": 0.00016744186046511629,
|
| 49 |
+
"loss": 4.2248,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.19,
|
| 54 |
+
"learning_rate": 0.00016279069767441862,
|
| 55 |
+
"loss": 4.0917,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.21,
|
| 60 |
+
"learning_rate": 0.00015813953488372093,
|
| 61 |
+
"loss": 3.964,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.23,
|
| 66 |
+
"learning_rate": 0.00015348837209302327,
|
| 67 |
+
"loss": 3.9154,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.23,
|
| 72 |
+
"eval_accuracy": 0.22130177514792898,
|
| 73 |
+
"eval_loss": 3.8417413234710693,
|
| 74 |
+
"eval_runtime": 33.9476,
|
| 75 |
+
"eval_samples_per_second": 49.783,
|
| 76 |
+
"eval_steps_per_second": 6.245,
|
| 77 |
+
"step": 100
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 0.26,
|
| 81 |
+
"learning_rate": 0.00014883720930232558,
|
| 82 |
+
"loss": 3.8266,
|
| 83 |
+
"step": 110
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.28,
|
| 87 |
+
"learning_rate": 0.00014418604651162791,
|
| 88 |
+
"loss": 3.721,
|
| 89 |
+
"step": 120
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.3,
|
| 93 |
+
"learning_rate": 0.00013953488372093025,
|
| 94 |
+
"loss": 3.5688,
|
| 95 |
+
"step": 130
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 0.33,
|
| 99 |
+
"learning_rate": 0.00013488372093023256,
|
| 100 |
+
"loss": 3.5956,
|
| 101 |
+
"step": 140
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.35,
|
| 105 |
+
"learning_rate": 0.0001302325581395349,
|
| 106 |
+
"loss": 3.5324,
|
| 107 |
+
"step": 150
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.37,
|
| 111 |
+
"learning_rate": 0.0001255813953488372,
|
| 112 |
+
"loss": 3.4582,
|
| 113 |
+
"step": 160
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.4,
|
| 117 |
+
"learning_rate": 0.00012093023255813953,
|
| 118 |
+
"loss": 3.4678,
|
| 119 |
+
"step": 170
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.42,
|
| 123 |
+
"learning_rate": 0.00011627906976744187,
|
| 124 |
+
"loss": 3.3492,
|
| 125 |
+
"step": 180
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.44,
|
| 129 |
+
"learning_rate": 0.00011162790697674419,
|
| 130 |
+
"loss": 3.4026,
|
| 131 |
+
"step": 190
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 0.47,
|
| 135 |
+
"learning_rate": 0.00010697674418604651,
|
| 136 |
+
"loss": 3.1764,
|
| 137 |
+
"step": 200
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.47,
|
| 141 |
+
"eval_accuracy": 0.3201183431952663,
|
| 142 |
+
"eval_loss": 3.2242767810821533,
|
| 143 |
+
"eval_runtime": 31.328,
|
| 144 |
+
"eval_samples_per_second": 53.945,
|
| 145 |
+
"eval_steps_per_second": 6.767,
|
| 146 |
+
"step": 200
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.49,
|
| 150 |
+
"learning_rate": 0.00010232558139534885,
|
| 151 |
+
"loss": 3.1029,
|
| 152 |
+
"step": 210
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.51,
|
| 156 |
+
"learning_rate": 9.767441860465116e-05,
|
| 157 |
+
"loss": 3.144,
|
| 158 |
+
"step": 220
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.53,
|
| 162 |
+
"learning_rate": 9.30232558139535e-05,
|
| 163 |
+
"loss": 3.092,
|
| 164 |
+
"step": 230
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.56,
|
| 168 |
+
"learning_rate": 8.837209302325582e-05,
|
| 169 |
+
"loss": 3.1839,
|
| 170 |
+
"step": 240
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.58,
|
| 174 |
+
"learning_rate": 8.372093023255814e-05,
|
| 175 |
+
"loss": 3.0839,
|
| 176 |
+
"step": 250
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.6,
|
| 180 |
+
"learning_rate": 7.906976744186047e-05,
|
| 181 |
+
"loss": 2.9506,
|
| 182 |
+
"step": 260
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.63,
|
| 186 |
+
"learning_rate": 7.441860465116279e-05,
|
| 187 |
+
"loss": 2.8607,
|
| 188 |
+
"step": 270
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.65,
|
| 192 |
+
"learning_rate": 6.976744186046513e-05,
|
| 193 |
+
"loss": 3.0074,
|
| 194 |
+
"step": 280
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.67,
|
| 198 |
+
"learning_rate": 6.511627906976745e-05,
|
| 199 |
+
"loss": 2.8763,
|
| 200 |
+
"step": 290
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.7,
|
| 204 |
+
"learning_rate": 6.0465116279069765e-05,
|
| 205 |
+
"loss": 2.8186,
|
| 206 |
+
"step": 300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.7,
|
| 210 |
+
"eval_accuracy": 0.4284023668639053,
|
| 211 |
+
"eval_loss": 2.7973387241363525,
|
| 212 |
+
"eval_runtime": 30.8947,
|
| 213 |
+
"eval_samples_per_second": 54.702,
|
| 214 |
+
"eval_steps_per_second": 6.862,
|
| 215 |
+
"step": 300
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.72,
|
| 219 |
+
"learning_rate": 5.5813953488372095e-05,
|
| 220 |
+
"loss": 2.8303,
|
| 221 |
+
"step": 310
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.74,
|
| 225 |
+
"learning_rate": 5.1162790697674425e-05,
|
| 226 |
+
"loss": 2.8085,
|
| 227 |
+
"step": 320
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.77,
|
| 231 |
+
"learning_rate": 4.651162790697675e-05,
|
| 232 |
+
"loss": 2.8264,
|
| 233 |
+
"step": 330
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.79,
|
| 237 |
+
"learning_rate": 4.186046511627907e-05,
|
| 238 |
+
"loss": 2.7512,
|
| 239 |
+
"step": 340
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 0.81,
|
| 243 |
+
"learning_rate": 3.7209302325581394e-05,
|
| 244 |
+
"loss": 2.6599,
|
| 245 |
+
"step": 350
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 0.84,
|
| 249 |
+
"learning_rate": 3.2558139534883724e-05,
|
| 250 |
+
"loss": 2.5858,
|
| 251 |
+
"step": 360
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.86,
|
| 255 |
+
"learning_rate": 2.7906976744186048e-05,
|
| 256 |
+
"loss": 2.6807,
|
| 257 |
+
"step": 370
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.88,
|
| 261 |
+
"learning_rate": 2.3255813953488374e-05,
|
| 262 |
+
"loss": 2.6653,
|
| 263 |
+
"step": 380
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.91,
|
| 267 |
+
"learning_rate": 1.8604651162790697e-05,
|
| 268 |
+
"loss": 2.4515,
|
| 269 |
+
"step": 390
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.93,
|
| 273 |
+
"learning_rate": 1.3953488372093024e-05,
|
| 274 |
+
"loss": 2.632,
|
| 275 |
+
"step": 400
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.93,
|
| 279 |
+
"eval_accuracy": 0.5076923076923077,
|
| 280 |
+
"eval_loss": 2.5673868656158447,
|
| 281 |
+
"eval_runtime": 33.1328,
|
| 282 |
+
"eval_samples_per_second": 51.007,
|
| 283 |
+
"eval_steps_per_second": 6.398,
|
| 284 |
+
"step": 400
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.95,
|
| 288 |
+
"learning_rate": 9.302325581395349e-06,
|
| 289 |
+
"loss": 2.6872,
|
| 290 |
+
"step": 410
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.98,
|
| 294 |
+
"learning_rate": 4.651162790697674e-06,
|
| 295 |
+
"loss": 2.556,
|
| 296 |
+
"step": 420
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.0,
|
| 300 |
+
"learning_rate": 0.0,
|
| 301 |
+
"loss": 2.5929,
|
| 302 |
+
"step": 430
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.0,
|
| 306 |
+
"step": 430,
|
| 307 |
+
"total_flos": 5.338473146715341e+17,
|
| 308 |
+
"train_loss": 3.3443971855695858,
|
| 309 |
+
"train_runtime": 396.8195,
|
| 310 |
+
"train_samples_per_second": 17.338,
|
| 311 |
+
"train_steps_per_second": 1.084
|
| 312 |
+
}
|
| 313 |
+
],
|
| 314 |
+
"max_steps": 430,
|
| 315 |
+
"num_train_epochs": 1,
|
| 316 |
+
"total_flos": 5.338473146715341e+17,
|
| 317 |
+
"trial_name": null,
|
| 318 |
+
"trial_params": null
|
| 319 |
+
}
|