🍻 cheers
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_accuracy": 0.9751972942502819,
|
| 4 |
+
"eval_loss": 0.08002225309610367,
|
| 5 |
+
"eval_runtime": 240.0574,
|
| 6 |
+
"eval_samples_per_second": 55.424,
|
| 7 |
+
"eval_steps_per_second": 6.932,
|
| 8 |
+
"total_flos": 8.247627438660256e+18,
|
| 9 |
+
"train_loss": 0.09209114413482199,
|
| 10 |
+
"train_runtime": 3805.7384,
|
| 11 |
+
"train_samples_per_second": 27.966,
|
| 12 |
+
"train_steps_per_second": 1.748
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_accuracy": 0.9751972942502819,
|
| 4 |
+
"eval_loss": 0.08002225309610367,
|
| 5 |
+
"eval_runtime": 240.0574,
|
| 6 |
+
"eval_samples_per_second": 55.424,
|
| 7 |
+
"eval_steps_per_second": 6.932
|
| 8 |
+
}
|
runs/Apr17_14-31-33_8e31fd4d164d/events.out.tfevents.1681745959.8e31fd4d164d.23.4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e05109fbaf07aa6b20257e5df98c448875f16db0514ce91268115b6a23e8cc61
|
| 3 |
+
size 363
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 8.247627438660256e+18,
|
| 4 |
+
"train_loss": 0.09209114413482199,
|
| 5 |
+
"train_runtime": 3805.7384,
|
| 6 |
+
"train_samples_per_second": 27.966,
|
| 7 |
+
"train_steps_per_second": 1.748
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,439 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.08002225309610367,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-nthu-ddd/checkpoint-6000",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 6652,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 0.00019702345159350573,
|
| 13 |
+
"loss": 0.1091,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 0.00019404690318701145,
|
| 19 |
+
"loss": 0.1244,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.09,
|
| 24 |
+
"learning_rate": 0.00019104028863499702,
|
| 25 |
+
"loss": 0.114,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.12,
|
| 30 |
+
"learning_rate": 0.00018803367408298257,
|
| 31 |
+
"loss": 0.1259,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.15,
|
| 36 |
+
"learning_rate": 0.00018502705953096814,
|
| 37 |
+
"loss": 0.1257,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.18,
|
| 42 |
+
"learning_rate": 0.0001820204449789537,
|
| 43 |
+
"loss": 0.1064,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.21,
|
| 48 |
+
"learning_rate": 0.0001790438965724594,
|
| 49 |
+
"loss": 0.1172,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.24,
|
| 54 |
+
"learning_rate": 0.00017603728202044498,
|
| 55 |
+
"loss": 0.1186,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.27,
|
| 60 |
+
"learning_rate": 0.00017303066746843056,
|
| 61 |
+
"loss": 0.0997,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.3,
|
| 66 |
+
"learning_rate": 0.00017002405291641613,
|
| 67 |
+
"loss": 0.1223,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.33,
|
| 72 |
+
"learning_rate": 0.00016701743836440168,
|
| 73 |
+
"loss": 0.1081,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.36,
|
| 78 |
+
"learning_rate": 0.00016401082381238725,
|
| 79 |
+
"loss": 0.1167,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.39,
|
| 84 |
+
"learning_rate": 0.00016100420926037283,
|
| 85 |
+
"loss": 0.0867,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.42,
|
| 90 |
+
"learning_rate": 0.0001579975947083584,
|
| 91 |
+
"loss": 0.1048,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.45,
|
| 96 |
+
"learning_rate": 0.00015499098015634398,
|
| 97 |
+
"loss": 0.1142,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.48,
|
| 102 |
+
"learning_rate": 0.00015198436560432953,
|
| 103 |
+
"loss": 0.1033,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.51,
|
| 108 |
+
"learning_rate": 0.00014900781719783524,
|
| 109 |
+
"loss": 0.1123,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.54,
|
| 114 |
+
"learning_rate": 0.00014600120264582082,
|
| 115 |
+
"loss": 0.1353,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.57,
|
| 120 |
+
"learning_rate": 0.00014299458809380637,
|
| 121 |
+
"loss": 0.0924,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.6,
|
| 126 |
+
"learning_rate": 0.00013998797354179194,
|
| 127 |
+
"loss": 0.1173,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.63,
|
| 132 |
+
"learning_rate": 0.00013698135898977752,
|
| 133 |
+
"loss": 0.0863,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.66,
|
| 138 |
+
"learning_rate": 0.0001339747444377631,
|
| 139 |
+
"loss": 0.1035,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.69,
|
| 144 |
+
"learning_rate": 0.00013096812988574864,
|
| 145 |
+
"loss": 0.0897,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.72,
|
| 150 |
+
"learning_rate": 0.0001279615153337342,
|
| 151 |
+
"loss": 0.0943,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.75,
|
| 156 |
+
"learning_rate": 0.0001249549007817198,
|
| 157 |
+
"loss": 0.1159,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.78,
|
| 162 |
+
"learning_rate": 0.00012194828622970536,
|
| 163 |
+
"loss": 0.0931,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.81,
|
| 168 |
+
"learning_rate": 0.00011894167167769092,
|
| 169 |
+
"loss": 0.1131,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.84,
|
| 174 |
+
"learning_rate": 0.0001159350571256765,
|
| 175 |
+
"loss": 0.1052,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.87,
|
| 180 |
+
"learning_rate": 0.00011292844257366206,
|
| 181 |
+
"loss": 0.0944,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.9,
|
| 186 |
+
"learning_rate": 0.00010992182802164763,
|
| 187 |
+
"loss": 0.0873,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.9,
|
| 192 |
+
"eval_accuracy": 0.9549793310785419,
|
| 193 |
+
"eval_loss": 0.1255834847688675,
|
| 194 |
+
"eval_runtime": 240.5647,
|
| 195 |
+
"eval_samples_per_second": 55.307,
|
| 196 |
+
"eval_steps_per_second": 6.917,
|
| 197 |
+
"step": 3000
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 0.93,
|
| 201 |
+
"learning_rate": 0.00010691521346963321,
|
| 202 |
+
"loss": 0.0907,
|
| 203 |
+
"step": 3100
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 0.96,
|
| 207 |
+
"learning_rate": 0.00010390859891761877,
|
| 208 |
+
"loss": 0.0836,
|
| 209 |
+
"step": 3200
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 0.99,
|
| 213 |
+
"learning_rate": 0.00010090198436560434,
|
| 214 |
+
"loss": 0.0791,
|
| 215 |
+
"step": 3300
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 1.02,
|
| 219 |
+
"learning_rate": 9.78953698135899e-05,
|
| 220 |
+
"loss": 0.0953,
|
| 221 |
+
"step": 3400
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.05,
|
| 225 |
+
"learning_rate": 9.488875526157547e-05,
|
| 226 |
+
"loss": 0.0859,
|
| 227 |
+
"step": 3500
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 1.08,
|
| 231 |
+
"learning_rate": 9.191220685508118e-05,
|
| 232 |
+
"loss": 0.1059,
|
| 233 |
+
"step": 3600
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.11,
|
| 237 |
+
"learning_rate": 8.890559230306676e-05,
|
| 238 |
+
"loss": 0.0859,
|
| 239 |
+
"step": 3700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 1.14,
|
| 243 |
+
"learning_rate": 8.589897775105232e-05,
|
| 244 |
+
"loss": 0.0965,
|
| 245 |
+
"step": 3800
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.17,
|
| 249 |
+
"learning_rate": 8.28923631990379e-05,
|
| 250 |
+
"loss": 0.0801,
|
| 251 |
+
"step": 3900
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 1.2,
|
| 255 |
+
"learning_rate": 7.988574864702346e-05,
|
| 256 |
+
"loss": 0.1072,
|
| 257 |
+
"step": 4000
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 1.23,
|
| 261 |
+
"learning_rate": 7.687913409500902e-05,
|
| 262 |
+
"loss": 0.0771,
|
| 263 |
+
"step": 4100
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.26,
|
| 267 |
+
"learning_rate": 7.387251954299459e-05,
|
| 268 |
+
"loss": 0.0854,
|
| 269 |
+
"step": 4200
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.29,
|
| 273 |
+
"learning_rate": 7.086590499098017e-05,
|
| 274 |
+
"loss": 0.0721,
|
| 275 |
+
"step": 4300
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.32,
|
| 279 |
+
"learning_rate": 6.785929043896573e-05,
|
| 280 |
+
"loss": 0.0929,
|
| 281 |
+
"step": 4400
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.35,
|
| 285 |
+
"learning_rate": 6.48526758869513e-05,
|
| 286 |
+
"loss": 0.0649,
|
| 287 |
+
"step": 4500
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.38,
|
| 291 |
+
"learning_rate": 6.184606133493686e-05,
|
| 292 |
+
"loss": 0.085,
|
| 293 |
+
"step": 4600
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.41,
|
| 297 |
+
"learning_rate": 5.883944678292244e-05,
|
| 298 |
+
"loss": 0.0838,
|
| 299 |
+
"step": 4700
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.44,
|
| 303 |
+
"learning_rate": 5.5832832230908004e-05,
|
| 304 |
+
"loss": 0.0629,
|
| 305 |
+
"step": 4800
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.47,
|
| 309 |
+
"learning_rate": 5.2826217678893566e-05,
|
| 310 |
+
"loss": 0.0985,
|
| 311 |
+
"step": 4900
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.5,
|
| 315 |
+
"learning_rate": 4.981960312687914e-05,
|
| 316 |
+
"loss": 0.0656,
|
| 317 |
+
"step": 5000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.53,
|
| 321 |
+
"learning_rate": 4.68129885748647e-05,
|
| 322 |
+
"loss": 0.0814,
|
| 323 |
+
"step": 5100
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.56,
|
| 327 |
+
"learning_rate": 4.380637402285027e-05,
|
| 328 |
+
"loss": 0.0667,
|
| 329 |
+
"step": 5200
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.59,
|
| 333 |
+
"learning_rate": 4.0799759470835844e-05,
|
| 334 |
+
"loss": 0.0944,
|
| 335 |
+
"step": 5300
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.62,
|
| 339 |
+
"learning_rate": 3.779314491882141e-05,
|
| 340 |
+
"loss": 0.0566,
|
| 341 |
+
"step": 5400
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.65,
|
| 345 |
+
"learning_rate": 3.478653036680698e-05,
|
| 346 |
+
"loss": 0.0867,
|
| 347 |
+
"step": 5500
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.68,
|
| 351 |
+
"learning_rate": 3.177991581479254e-05,
|
| 352 |
+
"loss": 0.0705,
|
| 353 |
+
"step": 5600
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.71,
|
| 357 |
+
"learning_rate": 2.877330126277811e-05,
|
| 358 |
+
"loss": 0.0734,
|
| 359 |
+
"step": 5700
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.74,
|
| 363 |
+
"learning_rate": 2.5766686710763683e-05,
|
| 364 |
+
"loss": 0.0781,
|
| 365 |
+
"step": 5800
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.77,
|
| 369 |
+
"learning_rate": 2.2760072158749247e-05,
|
| 370 |
+
"loss": 0.0712,
|
| 371 |
+
"step": 5900
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.8,
|
| 375 |
+
"learning_rate": 1.975345760673482e-05,
|
| 376 |
+
"loss": 0.06,
|
| 377 |
+
"step": 6000
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.8,
|
| 381 |
+
"eval_accuracy": 0.9751972942502819,
|
| 382 |
+
"eval_loss": 0.08002225309610367,
|
| 383 |
+
"eval_runtime": 240.0279,
|
| 384 |
+
"eval_samples_per_second": 55.431,
|
| 385 |
+
"eval_steps_per_second": 6.933,
|
| 386 |
+
"step": 6000
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 1.83,
|
| 390 |
+
"learning_rate": 1.6746843054720386e-05,
|
| 391 |
+
"loss": 0.0698,
|
| 392 |
+
"step": 6100
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 1.86,
|
| 396 |
+
"learning_rate": 1.3740228502705954e-05,
|
| 397 |
+
"loss": 0.0705,
|
| 398 |
+
"step": 6200
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 1.89,
|
| 402 |
+
"learning_rate": 1.0733613950691522e-05,
|
| 403 |
+
"loss": 0.0553,
|
| 404 |
+
"step": 6300
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.92,
|
| 408 |
+
"learning_rate": 7.72699939867709e-06,
|
| 409 |
+
"loss": 0.0743,
|
| 410 |
+
"step": 6400
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 1.95,
|
| 414 |
+
"learning_rate": 4.750450992182803e-06,
|
| 415 |
+
"loss": 0.0636,
|
| 416 |
+
"step": 6500
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.98,
|
| 420 |
+
"learning_rate": 1.7438364401683706e-06,
|
| 421 |
+
"loss": 0.083,
|
| 422 |
+
"step": 6600
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.0,
|
| 426 |
+
"step": 6652,
|
| 427 |
+
"total_flos": 8.247627438660256e+18,
|
| 428 |
+
"train_loss": 0.09209114413482199,
|
| 429 |
+
"train_runtime": 3805.7384,
|
| 430 |
+
"train_samples_per_second": 27.966,
|
| 431 |
+
"train_steps_per_second": 1.748
|
| 432 |
+
}
|
| 433 |
+
],
|
| 434 |
+
"max_steps": 6652,
|
| 435 |
+
"num_train_epochs": 2,
|
| 436 |
+
"total_flos": 8.247627438660256e+18,
|
| 437 |
+
"trial_name": null,
|
| 438 |
+
"trial_params": null
|
| 439 |
+
}
|