🍻 cheers
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_accuracy": 0.7224363916730917,
|
| 4 |
+
"eval_loss": 1.388809323310852,
|
| 5 |
+
"eval_runtime": 31.0004,
|
| 6 |
+
"eval_samples_per_second": 83.676,
|
| 7 |
+
"eval_steps_per_second": 10.484,
|
| 8 |
+
"total_flos": 7.922356224994714e+17,
|
| 9 |
+
"train_loss": 2.18192570049065,
|
| 10 |
+
"train_runtime": 471.371,
|
| 11 |
+
"train_samples_per_second": 21.66,
|
| 12 |
+
"train_steps_per_second": 1.356
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_accuracy": 0.7224363916730917,
|
| 4 |
+
"eval_loss": 1.388809323310852,
|
| 5 |
+
"eval_runtime": 31.0004,
|
| 6 |
+
"eval_samples_per_second": 83.676,
|
| 7 |
+
"eval_steps_per_second": 10.484
|
| 8 |
+
}
|
runs/Jun19_14-29-47_7a3eefb8a690/events.out.tfevents.1655649804.7a3eefb8a690.72.4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:504edc6e54352b00e8304968a7784db6437efbae8e60fdcfeec2d2df9f952b35
|
| 3 |
+
size 363
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 7.922356224994714e+17,
|
| 4 |
+
"train_loss": 2.18192570049065,
|
| 5 |
+
"train_runtime": 471.371,
|
| 6 |
+
"train_samples_per_second": 21.66,
|
| 7 |
+
"train_steps_per_second": 1.356
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,457 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 1.388809323310852,
|
| 3 |
+
"best_model_checkpoint": "./modelversion01/checkpoint-600",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"global_step": 639,
|
| 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": 0.0001968701095461659,
|
| 13 |
+
"loss": 3.6977,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.03,
|
| 18 |
+
"learning_rate": 0.00019374021909233179,
|
| 19 |
+
"loss": 3.7426,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.05,
|
| 24 |
+
"learning_rate": 0.00019061032863849765,
|
| 25 |
+
"loss": 3.5577,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.06,
|
| 30 |
+
"learning_rate": 0.00018748043818466354,
|
| 31 |
+
"loss": 3.3304,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.08,
|
| 36 |
+
"learning_rate": 0.00018435054773082942,
|
| 37 |
+
"loss": 3.3761,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.09,
|
| 42 |
+
"learning_rate": 0.0001812206572769953,
|
| 43 |
+
"loss": 3.2324,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.11,
|
| 48 |
+
"learning_rate": 0.0001780907668231612,
|
| 49 |
+
"loss": 3.1829,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.13,
|
| 54 |
+
"learning_rate": 0.0001749608763693271,
|
| 55 |
+
"loss": 3.0992,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.14,
|
| 60 |
+
"learning_rate": 0.00017183098591549295,
|
| 61 |
+
"loss": 3.2535,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.16,
|
| 66 |
+
"learning_rate": 0.00016870109546165884,
|
| 67 |
+
"loss": 3.1304,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.16,
|
| 72 |
+
"eval_accuracy": 0.3253662297609869,
|
| 73 |
+
"eval_loss": 3.154646873474121,
|
| 74 |
+
"eval_runtime": 30.5027,
|
| 75 |
+
"eval_samples_per_second": 85.042,
|
| 76 |
+
"eval_steps_per_second": 10.655,
|
| 77 |
+
"step": 100
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 0.17,
|
| 81 |
+
"learning_rate": 0.00016557120500782473,
|
| 82 |
+
"loss": 3.1477,
|
| 83 |
+
"step": 110
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.19,
|
| 87 |
+
"learning_rate": 0.00016244131455399061,
|
| 88 |
+
"loss": 2.9492,
|
| 89 |
+
"step": 120
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.2,
|
| 93 |
+
"learning_rate": 0.0001593114241001565,
|
| 94 |
+
"loss": 2.7843,
|
| 95 |
+
"step": 130
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 0.22,
|
| 99 |
+
"learning_rate": 0.00015618153364632236,
|
| 100 |
+
"loss": 2.919,
|
| 101 |
+
"step": 140
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.23,
|
| 105 |
+
"learning_rate": 0.00015305164319248828,
|
| 106 |
+
"loss": 2.703,
|
| 107 |
+
"step": 150
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.25,
|
| 111 |
+
"learning_rate": 0.00014992175273865417,
|
| 112 |
+
"loss": 2.7079,
|
| 113 |
+
"step": 160
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.27,
|
| 117 |
+
"learning_rate": 0.00014679186228482006,
|
| 118 |
+
"loss": 2.5964,
|
| 119 |
+
"step": 170
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.28,
|
| 123 |
+
"learning_rate": 0.00014366197183098594,
|
| 124 |
+
"loss": 2.7854,
|
| 125 |
+
"step": 180
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.3,
|
| 129 |
+
"learning_rate": 0.0001405320813771518,
|
| 130 |
+
"loss": 2.4903,
|
| 131 |
+
"step": 190
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 0.31,
|
| 135 |
+
"learning_rate": 0.0001374021909233177,
|
| 136 |
+
"loss": 2.6514,
|
| 137 |
+
"step": 200
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.31,
|
| 141 |
+
"eval_accuracy": 0.4853508095605243,
|
| 142 |
+
"eval_loss": 2.5057504177093506,
|
| 143 |
+
"eval_runtime": 34.1479,
|
| 144 |
+
"eval_samples_per_second": 75.964,
|
| 145 |
+
"eval_steps_per_second": 9.517,
|
| 146 |
+
"step": 200
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.33,
|
| 150 |
+
"learning_rate": 0.00013427230046948358,
|
| 151 |
+
"loss": 2.4657,
|
| 152 |
+
"step": 210
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.34,
|
| 156 |
+
"learning_rate": 0.00013114241001564947,
|
| 157 |
+
"loss": 2.4503,
|
| 158 |
+
"step": 220
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.36,
|
| 162 |
+
"learning_rate": 0.00012801251956181536,
|
| 163 |
+
"loss": 2.2323,
|
| 164 |
+
"step": 230
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.38,
|
| 168 |
+
"learning_rate": 0.00012488262910798122,
|
| 169 |
+
"loss": 2.1866,
|
| 170 |
+
"step": 240
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.39,
|
| 174 |
+
"learning_rate": 0.00012175273865414711,
|
| 175 |
+
"loss": 2.3176,
|
| 176 |
+
"step": 250
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.41,
|
| 180 |
+
"learning_rate": 0.000118622848200313,
|
| 181 |
+
"loss": 2.2193,
|
| 182 |
+
"step": 260
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.42,
|
| 186 |
+
"learning_rate": 0.00011549295774647888,
|
| 187 |
+
"loss": 2.1877,
|
| 188 |
+
"step": 270
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.44,
|
| 192 |
+
"learning_rate": 0.00011236306729264476,
|
| 193 |
+
"loss": 2.2086,
|
| 194 |
+
"step": 280
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.45,
|
| 198 |
+
"learning_rate": 0.00010923317683881065,
|
| 199 |
+
"loss": 2.2096,
|
| 200 |
+
"step": 290
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.47,
|
| 204 |
+
"learning_rate": 0.00010610328638497653,
|
| 205 |
+
"loss": 2.0636,
|
| 206 |
+
"step": 300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.47,
|
| 210 |
+
"eval_accuracy": 0.5771010023130301,
|
| 211 |
+
"eval_loss": 2.0646958351135254,
|
| 212 |
+
"eval_runtime": 34.5408,
|
| 213 |
+
"eval_samples_per_second": 75.1,
|
| 214 |
+
"eval_steps_per_second": 9.409,
|
| 215 |
+
"step": 300
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.49,
|
| 219 |
+
"learning_rate": 0.00010297339593114241,
|
| 220 |
+
"loss": 2.1629,
|
| 221 |
+
"step": 310
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.5,
|
| 225 |
+
"learning_rate": 9.98435054773083e-05,
|
| 226 |
+
"loss": 1.9858,
|
| 227 |
+
"step": 320
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.52,
|
| 231 |
+
"learning_rate": 9.671361502347419e-05,
|
| 232 |
+
"loss": 2.1714,
|
| 233 |
+
"step": 330
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.53,
|
| 237 |
+
"learning_rate": 9.358372456964006e-05,
|
| 238 |
+
"loss": 1.9333,
|
| 239 |
+
"step": 340
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 0.55,
|
| 243 |
+
"learning_rate": 9.045383411580595e-05,
|
| 244 |
+
"loss": 1.9808,
|
| 245 |
+
"step": 350
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 0.56,
|
| 249 |
+
"learning_rate": 8.732394366197182e-05,
|
| 250 |
+
"loss": 1.8507,
|
| 251 |
+
"step": 360
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.58,
|
| 255 |
+
"learning_rate": 8.419405320813773e-05,
|
| 256 |
+
"loss": 1.9109,
|
| 257 |
+
"step": 370
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.59,
|
| 261 |
+
"learning_rate": 8.106416275430361e-05,
|
| 262 |
+
"loss": 1.7645,
|
| 263 |
+
"step": 380
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.61,
|
| 267 |
+
"learning_rate": 7.793427230046949e-05,
|
| 268 |
+
"loss": 1.7828,
|
| 269 |
+
"step": 390
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.63,
|
| 273 |
+
"learning_rate": 7.480438184663538e-05,
|
| 274 |
+
"loss": 1.7812,
|
| 275 |
+
"step": 400
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.63,
|
| 279 |
+
"eval_accuracy": 0.6422513492675405,
|
| 280 |
+
"eval_loss": 1.7535737752914429,
|
| 281 |
+
"eval_runtime": 34.377,
|
| 282 |
+
"eval_samples_per_second": 75.457,
|
| 283 |
+
"eval_steps_per_second": 9.454,
|
| 284 |
+
"step": 400
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.64,
|
| 288 |
+
"learning_rate": 7.167449139280125e-05,
|
| 289 |
+
"loss": 1.7357,
|
| 290 |
+
"step": 410
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.66,
|
| 294 |
+
"learning_rate": 6.854460093896714e-05,
|
| 295 |
+
"loss": 1.8218,
|
| 296 |
+
"step": 420
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.67,
|
| 300 |
+
"learning_rate": 6.541471048513303e-05,
|
| 301 |
+
"loss": 1.504,
|
| 302 |
+
"step": 430
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.69,
|
| 306 |
+
"learning_rate": 6.22848200312989e-05,
|
| 307 |
+
"loss": 1.665,
|
| 308 |
+
"step": 440
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.7,
|
| 312 |
+
"learning_rate": 5.915492957746479e-05,
|
| 313 |
+
"loss": 1.6636,
|
| 314 |
+
"step": 450
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.72,
|
| 318 |
+
"learning_rate": 5.602503912363067e-05,
|
| 319 |
+
"loss": 1.7505,
|
| 320 |
+
"step": 460
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.74,
|
| 324 |
+
"learning_rate": 5.2895148669796554e-05,
|
| 325 |
+
"loss": 1.528,
|
| 326 |
+
"step": 470
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.75,
|
| 330 |
+
"learning_rate": 4.976525821596245e-05,
|
| 331 |
+
"loss": 1.4391,
|
| 332 |
+
"step": 480
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.77,
|
| 336 |
+
"learning_rate": 4.663536776212833e-05,
|
| 337 |
+
"loss": 1.6142,
|
| 338 |
+
"step": 490
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.78,
|
| 342 |
+
"learning_rate": 4.350547730829421e-05,
|
| 343 |
+
"loss": 1.5857,
|
| 344 |
+
"step": 500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.78,
|
| 348 |
+
"eval_accuracy": 0.6973785659213569,
|
| 349 |
+
"eval_loss": 1.5272208452224731,
|
| 350 |
+
"eval_runtime": 33.0601,
|
| 351 |
+
"eval_samples_per_second": 78.463,
|
| 352 |
+
"eval_steps_per_second": 9.831,
|
| 353 |
+
"step": 500
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.8,
|
| 357 |
+
"learning_rate": 4.037558685446009e-05,
|
| 358 |
+
"loss": 1.5904,
|
| 359 |
+
"step": 510
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.81,
|
| 363 |
+
"learning_rate": 3.724569640062598e-05,
|
| 364 |
+
"loss": 1.4461,
|
| 365 |
+
"step": 520
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 0.83,
|
| 369 |
+
"learning_rate": 3.411580594679186e-05,
|
| 370 |
+
"loss": 1.5556,
|
| 371 |
+
"step": 530
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.85,
|
| 375 |
+
"learning_rate": 3.0985915492957744e-05,
|
| 376 |
+
"loss": 1.6289,
|
| 377 |
+
"step": 540
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.86,
|
| 381 |
+
"learning_rate": 2.7856025039123636e-05,
|
| 382 |
+
"loss": 1.4632,
|
| 383 |
+
"step": 550
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 0.88,
|
| 387 |
+
"learning_rate": 2.4726134585289514e-05,
|
| 388 |
+
"loss": 1.3283,
|
| 389 |
+
"step": 560
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.89,
|
| 393 |
+
"learning_rate": 2.1596244131455402e-05,
|
| 394 |
+
"loss": 1.3679,
|
| 395 |
+
"step": 570
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.91,
|
| 399 |
+
"learning_rate": 1.8466353677621284e-05,
|
| 400 |
+
"loss": 1.4145,
|
| 401 |
+
"step": 580
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.92,
|
| 405 |
+
"learning_rate": 1.533646322378717e-05,
|
| 406 |
+
"loss": 1.3822,
|
| 407 |
+
"step": 590
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 0.94,
|
| 411 |
+
"learning_rate": 1.2206572769953052e-05,
|
| 412 |
+
"loss": 1.3055,
|
| 413 |
+
"step": 600
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 0.94,
|
| 417 |
+
"eval_accuracy": 0.7224363916730917,
|
| 418 |
+
"eval_loss": 1.388809323310852,
|
| 419 |
+
"eval_runtime": 29.6379,
|
| 420 |
+
"eval_samples_per_second": 87.523,
|
| 421 |
+
"eval_steps_per_second": 10.966,
|
| 422 |
+
"step": 600
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.95,
|
| 426 |
+
"learning_rate": 9.076682316118937e-06,
|
| 427 |
+
"loss": 1.4099,
|
| 428 |
+
"step": 610
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.97,
|
| 432 |
+
"learning_rate": 5.94679186228482e-06,
|
| 433 |
+
"loss": 1.3346,
|
| 434 |
+
"step": 620
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.99,
|
| 438 |
+
"learning_rate": 2.8169014084507042e-06,
|
| 439 |
+
"loss": 1.4532,
|
| 440 |
+
"step": 630
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 1.0,
|
| 444 |
+
"step": 639,
|
| 445 |
+
"total_flos": 7.922356224994714e+17,
|
| 446 |
+
"train_loss": 2.18192570049065,
|
| 447 |
+
"train_runtime": 471.371,
|
| 448 |
+
"train_samples_per_second": 21.66,
|
| 449 |
+
"train_steps_per_second": 1.356
|
| 450 |
+
}
|
| 451 |
+
],
|
| 452 |
+
"max_steps": 639,
|
| 453 |
+
"num_train_epochs": 1,
|
| 454 |
+
"total_flos": 7.922356224994714e+17,
|
| 455 |
+
"trial_name": null,
|
| 456 |
+
"trial_params": null
|
| 457 |
+
}
|