End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.2442307692307693,
|
| 3 |
+
"eval_accuracy": 0.6320582877959927,
|
| 4 |
+
"eval_loss": 1.9572811126708984,
|
| 5 |
+
"eval_runtime": 146.1218,
|
| 6 |
+
"eval_samples_per_second": 3.757,
|
| 7 |
+
"eval_steps_per_second": 0.472
|
| 8 |
+
}
|
runs/Jan08_18-46-08_5a3016f64bbf/events.out.tfevents.1736364948.5a3016f64bbf.332.1
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:c91ac5ccebfda156917300782837c7154abc64899b7141d6e80ed6808e3a5c55
|
| 3 |
+
size 734
|
test_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.2442307692307693,
|
| 3 |
+
"eval_accuracy": 0.6320582877959927,
|
| 4 |
+
"eval_loss": 1.9572811126708984,
|
| 5 |
+
"eval_runtime": 146.1218,
|
| 6 |
+
"eval_samples_per_second": 3.757,
|
| 7 |
+
"eval_steps_per_second": 0.472
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,460 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.820627802690583,
|
| 3 |
+
"best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-520",
|
| 4 |
+
"epoch": 3.2442307692307693,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 520,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.019230769230769232,
|
| 13 |
+
"grad_norm": 8.085018157958984,
|
| 14 |
+
"learning_rate": 9.615384615384616e-06,
|
| 15 |
+
"loss": 3.4301,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.038461538461538464,
|
| 20 |
+
"grad_norm": 8.687586784362793,
|
| 21 |
+
"learning_rate": 1.923076923076923e-05,
|
| 22 |
+
"loss": 3.3945,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.057692307692307696,
|
| 27 |
+
"grad_norm": 9.751335144042969,
|
| 28 |
+
"learning_rate": 2.8846153846153845e-05,
|
| 29 |
+
"loss": 3.426,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.07692307692307693,
|
| 34 |
+
"grad_norm": 8.339034080505371,
|
| 35 |
+
"learning_rate": 3.846153846153846e-05,
|
| 36 |
+
"loss": 3.4135,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.09615384615384616,
|
| 41 |
+
"grad_norm": 7.60875129699707,
|
| 42 |
+
"learning_rate": 4.8076923076923084e-05,
|
| 43 |
+
"loss": 3.295,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.11538461538461539,
|
| 48 |
+
"grad_norm": 8.210917472839355,
|
| 49 |
+
"learning_rate": 4.9145299145299147e-05,
|
| 50 |
+
"loss": 3.2379,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.1346153846153846,
|
| 55 |
+
"grad_norm": 8.948013305664062,
|
| 56 |
+
"learning_rate": 4.8076923076923084e-05,
|
| 57 |
+
"loss": 3.181,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.15384615384615385,
|
| 62 |
+
"grad_norm": 9.540766716003418,
|
| 63 |
+
"learning_rate": 4.700854700854701e-05,
|
| 64 |
+
"loss": 3.0669,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.17307692307692307,
|
| 69 |
+
"grad_norm": 10.327016830444336,
|
| 70 |
+
"learning_rate": 4.594017094017094e-05,
|
| 71 |
+
"loss": 2.9399,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.19230769230769232,
|
| 76 |
+
"grad_norm": 10.140628814697266,
|
| 77 |
+
"learning_rate": 4.4871794871794874e-05,
|
| 78 |
+
"loss": 2.7217,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.21153846153846154,
|
| 83 |
+
"grad_norm": 10.16087818145752,
|
| 84 |
+
"learning_rate": 4.3803418803418805e-05,
|
| 85 |
+
"loss": 2.5126,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.23076923076923078,
|
| 90 |
+
"grad_norm": 9.189348220825195,
|
| 91 |
+
"learning_rate": 4.2735042735042735e-05,
|
| 92 |
+
"loss": 2.4065,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.25,
|
| 97 |
+
"grad_norm": 11.86257553100586,
|
| 98 |
+
"learning_rate": 4.166666666666667e-05,
|
| 99 |
+
"loss": 2.1315,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.2519230769230769,
|
| 104 |
+
"eval_accuracy": 0.38565022421524664,
|
| 105 |
+
"eval_loss": 2.2071354389190674,
|
| 106 |
+
"eval_runtime": 59.4454,
|
| 107 |
+
"eval_samples_per_second": 3.751,
|
| 108 |
+
"eval_steps_per_second": 0.471,
|
| 109 |
+
"step": 131
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 1.0173076923076922,
|
| 113 |
+
"grad_norm": 10.583504676818848,
|
| 114 |
+
"learning_rate": 4.05982905982906e-05,
|
| 115 |
+
"loss": 1.9809,
|
| 116 |
+
"step": 140
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 1.0365384615384616,
|
| 120 |
+
"grad_norm": 12.537158966064453,
|
| 121 |
+
"learning_rate": 3.952991452991453e-05,
|
| 122 |
+
"loss": 1.8472,
|
| 123 |
+
"step": 150
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 1.0557692307692308,
|
| 127 |
+
"grad_norm": 13.388022422790527,
|
| 128 |
+
"learning_rate": 3.846153846153846e-05,
|
| 129 |
+
"loss": 1.6871,
|
| 130 |
+
"step": 160
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 1.075,
|
| 134 |
+
"grad_norm": 20.201147079467773,
|
| 135 |
+
"learning_rate": 3.739316239316239e-05,
|
| 136 |
+
"loss": 1.7488,
|
| 137 |
+
"step": 170
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.0942307692307693,
|
| 141 |
+
"grad_norm": 14.368993759155273,
|
| 142 |
+
"learning_rate": 3.6324786324786323e-05,
|
| 143 |
+
"loss": 1.6513,
|
| 144 |
+
"step": 180
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.1134615384615385,
|
| 148 |
+
"grad_norm": 11.838800430297852,
|
| 149 |
+
"learning_rate": 3.525641025641026e-05,
|
| 150 |
+
"loss": 1.5155,
|
| 151 |
+
"step": 190
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 1.1326923076923077,
|
| 155 |
+
"grad_norm": 12.411248207092285,
|
| 156 |
+
"learning_rate": 3.418803418803419e-05,
|
| 157 |
+
"loss": 1.6197,
|
| 158 |
+
"step": 200
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.1519230769230768,
|
| 162 |
+
"grad_norm": 9.437255859375,
|
| 163 |
+
"learning_rate": 3.311965811965812e-05,
|
| 164 |
+
"loss": 1.356,
|
| 165 |
+
"step": 210
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.1711538461538462,
|
| 169 |
+
"grad_norm": 9.415695190429688,
|
| 170 |
+
"learning_rate": 3.205128205128206e-05,
|
| 171 |
+
"loss": 1.2385,
|
| 172 |
+
"step": 220
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 1.1903846153846154,
|
| 176 |
+
"grad_norm": 14.387253761291504,
|
| 177 |
+
"learning_rate": 3.098290598290599e-05,
|
| 178 |
+
"loss": 1.1676,
|
| 179 |
+
"step": 230
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.2096153846153845,
|
| 183 |
+
"grad_norm": 18.609485626220703,
|
| 184 |
+
"learning_rate": 2.9914529914529915e-05,
|
| 185 |
+
"loss": 1.0883,
|
| 186 |
+
"step": 240
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 1.228846153846154,
|
| 190 |
+
"grad_norm": 8.951836585998535,
|
| 191 |
+
"learning_rate": 2.8846153846153845e-05,
|
| 192 |
+
"loss": 1.0067,
|
| 193 |
+
"step": 250
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 1.248076923076923,
|
| 197 |
+
"grad_norm": 9.234614372253418,
|
| 198 |
+
"learning_rate": 2.777777777777778e-05,
|
| 199 |
+
"loss": 1.1581,
|
| 200 |
+
"step": 260
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.251923076923077,
|
| 204 |
+
"eval_accuracy": 0.6547085201793722,
|
| 205 |
+
"eval_loss": 1.158636212348938,
|
| 206 |
+
"eval_runtime": 59.7594,
|
| 207 |
+
"eval_samples_per_second": 3.732,
|
| 208 |
+
"eval_steps_per_second": 0.469,
|
| 209 |
+
"step": 262
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 2.0153846153846153,
|
| 213 |
+
"grad_norm": 9.139333724975586,
|
| 214 |
+
"learning_rate": 2.670940170940171e-05,
|
| 215 |
+
"loss": 0.8048,
|
| 216 |
+
"step": 270
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 2.0346153846153845,
|
| 220 |
+
"grad_norm": 19.066368103027344,
|
| 221 |
+
"learning_rate": 2.564102564102564e-05,
|
| 222 |
+
"loss": 1.0916,
|
| 223 |
+
"step": 280
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 2.0538461538461537,
|
| 227 |
+
"grad_norm": 5.820519924163818,
|
| 228 |
+
"learning_rate": 2.4572649572649573e-05,
|
| 229 |
+
"loss": 0.8838,
|
| 230 |
+
"step": 290
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.0730769230769233,
|
| 234 |
+
"grad_norm": 9.551618576049805,
|
| 235 |
+
"learning_rate": 2.3504273504273504e-05,
|
| 236 |
+
"loss": 0.7372,
|
| 237 |
+
"step": 300
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 2.0923076923076924,
|
| 241 |
+
"grad_norm": 5.678286075592041,
|
| 242 |
+
"learning_rate": 2.2435897435897437e-05,
|
| 243 |
+
"loss": 0.7344,
|
| 244 |
+
"step": 310
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 2.1115384615384616,
|
| 248 |
+
"grad_norm": 7.4141411781311035,
|
| 249 |
+
"learning_rate": 2.1367521367521368e-05,
|
| 250 |
+
"loss": 0.5431,
|
| 251 |
+
"step": 320
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 2.1307692307692307,
|
| 255 |
+
"grad_norm": 12.246278762817383,
|
| 256 |
+
"learning_rate": 2.02991452991453e-05,
|
| 257 |
+
"loss": 0.7999,
|
| 258 |
+
"step": 330
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 2.15,
|
| 262 |
+
"grad_norm": 9.39754581451416,
|
| 263 |
+
"learning_rate": 1.923076923076923e-05,
|
| 264 |
+
"loss": 0.7329,
|
| 265 |
+
"step": 340
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 2.169230769230769,
|
| 269 |
+
"grad_norm": 11.727778434753418,
|
| 270 |
+
"learning_rate": 1.8162393162393162e-05,
|
| 271 |
+
"loss": 0.5766,
|
| 272 |
+
"step": 350
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.1884615384615387,
|
| 276 |
+
"grad_norm": 5.652421474456787,
|
| 277 |
+
"learning_rate": 1.7094017094017095e-05,
|
| 278 |
+
"loss": 0.6079,
|
| 279 |
+
"step": 360
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 2.207692307692308,
|
| 283 |
+
"grad_norm": 18.59898567199707,
|
| 284 |
+
"learning_rate": 1.602564102564103e-05,
|
| 285 |
+
"loss": 0.5848,
|
| 286 |
+
"step": 370
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 2.226923076923077,
|
| 290 |
+
"grad_norm": 9.42827033996582,
|
| 291 |
+
"learning_rate": 1.4957264957264958e-05,
|
| 292 |
+
"loss": 0.4852,
|
| 293 |
+
"step": 380
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 2.246153846153846,
|
| 297 |
+
"grad_norm": 10.546754837036133,
|
| 298 |
+
"learning_rate": 1.388888888888889e-05,
|
| 299 |
+
"loss": 0.6854,
|
| 300 |
+
"step": 390
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 2.251923076923077,
|
| 304 |
+
"eval_accuracy": 0.7623318385650224,
|
| 305 |
+
"eval_loss": 0.816370964050293,
|
| 306 |
+
"eval_runtime": 60.0828,
|
| 307 |
+
"eval_samples_per_second": 3.712,
|
| 308 |
+
"eval_steps_per_second": 0.466,
|
| 309 |
+
"step": 393
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 3.0134615384615384,
|
| 313 |
+
"grad_norm": 7.366909980773926,
|
| 314 |
+
"learning_rate": 1.282051282051282e-05,
|
| 315 |
+
"loss": 0.4249,
|
| 316 |
+
"step": 400
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 3.0326923076923076,
|
| 320 |
+
"grad_norm": 19.081708908081055,
|
| 321 |
+
"learning_rate": 1.1752136752136752e-05,
|
| 322 |
+
"loss": 0.4456,
|
| 323 |
+
"step": 410
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 3.0519230769230767,
|
| 327 |
+
"grad_norm": 6.918426990509033,
|
| 328 |
+
"learning_rate": 1.0683760683760684e-05,
|
| 329 |
+
"loss": 0.4338,
|
| 330 |
+
"step": 420
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 3.0711538461538463,
|
| 334 |
+
"grad_norm": 20.953861236572266,
|
| 335 |
+
"learning_rate": 9.615384615384616e-06,
|
| 336 |
+
"loss": 0.4404,
|
| 337 |
+
"step": 430
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 3.0903846153846155,
|
| 341 |
+
"grad_norm": 7.814371585845947,
|
| 342 |
+
"learning_rate": 8.547008547008548e-06,
|
| 343 |
+
"loss": 0.415,
|
| 344 |
+
"step": 440
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 3.1096153846153847,
|
| 348 |
+
"grad_norm": 5.119196891784668,
|
| 349 |
+
"learning_rate": 7.478632478632479e-06,
|
| 350 |
+
"loss": 0.387,
|
| 351 |
+
"step": 450
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 3.128846153846154,
|
| 355 |
+
"grad_norm": 12.346413612365723,
|
| 356 |
+
"learning_rate": 6.41025641025641e-06,
|
| 357 |
+
"loss": 0.3826,
|
| 358 |
+
"step": 460
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 3.148076923076923,
|
| 362 |
+
"grad_norm": 13.700543403625488,
|
| 363 |
+
"learning_rate": 5.341880341880342e-06,
|
| 364 |
+
"loss": 0.2908,
|
| 365 |
+
"step": 470
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 3.167307692307692,
|
| 369 |
+
"grad_norm": 11.321188926696777,
|
| 370 |
+
"learning_rate": 4.273504273504274e-06,
|
| 371 |
+
"loss": 0.4446,
|
| 372 |
+
"step": 480
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 3.1865384615384613,
|
| 376 |
+
"grad_norm": 11.503890037536621,
|
| 377 |
+
"learning_rate": 3.205128205128205e-06,
|
| 378 |
+
"loss": 0.3246,
|
| 379 |
+
"step": 490
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 3.205769230769231,
|
| 383 |
+
"grad_norm": 9.901238441467285,
|
| 384 |
+
"learning_rate": 2.136752136752137e-06,
|
| 385 |
+
"loss": 0.3634,
|
| 386 |
+
"step": 500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 3.225,
|
| 390 |
+
"grad_norm": 4.689971446990967,
|
| 391 |
+
"learning_rate": 1.0683760683760685e-06,
|
| 392 |
+
"loss": 0.3183,
|
| 393 |
+
"step": 510
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 3.2442307692307693,
|
| 397 |
+
"grad_norm": 17.20209503173828,
|
| 398 |
+
"learning_rate": 0.0,
|
| 399 |
+
"loss": 0.3349,
|
| 400 |
+
"step": 520
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 3.2442307692307693,
|
| 404 |
+
"eval_accuracy": 0.820627802690583,
|
| 405 |
+
"eval_loss": 0.6745598912239075,
|
| 406 |
+
"eval_runtime": 60.6402,
|
| 407 |
+
"eval_samples_per_second": 3.677,
|
| 408 |
+
"eval_steps_per_second": 0.462,
|
| 409 |
+
"step": 520
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 3.2442307692307693,
|
| 413 |
+
"step": 520,
|
| 414 |
+
"total_flos": 5.181183958187704e+18,
|
| 415 |
+
"train_loss": 1.3941553766910846,
|
| 416 |
+
"train_runtime": 2772.4544,
|
| 417 |
+
"train_samples_per_second": 1.5,
|
| 418 |
+
"train_steps_per_second": 0.188
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 3.2442307692307693,
|
| 422 |
+
"eval_accuracy": 0.6320582877959927,
|
| 423 |
+
"eval_loss": 1.9572809934616089,
|
| 424 |
+
"eval_runtime": 147.8534,
|
| 425 |
+
"eval_samples_per_second": 3.713,
|
| 426 |
+
"eval_steps_per_second": 0.467,
|
| 427 |
+
"step": 520
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 3.2442307692307693,
|
| 431 |
+
"eval_accuracy": 0.6320582877959927,
|
| 432 |
+
"eval_loss": 1.9572811126708984,
|
| 433 |
+
"eval_runtime": 146.1218,
|
| 434 |
+
"eval_samples_per_second": 3.757,
|
| 435 |
+
"eval_steps_per_second": 0.472,
|
| 436 |
+
"step": 520
|
| 437 |
+
}
|
| 438 |
+
],
|
| 439 |
+
"logging_steps": 10,
|
| 440 |
+
"max_steps": 520,
|
| 441 |
+
"num_input_tokens_seen": 0,
|
| 442 |
+
"num_train_epochs": 9223372036854775807,
|
| 443 |
+
"save_steps": 500,
|
| 444 |
+
"stateful_callbacks": {
|
| 445 |
+
"TrainerControl": {
|
| 446 |
+
"args": {
|
| 447 |
+
"should_epoch_stop": false,
|
| 448 |
+
"should_evaluate": false,
|
| 449 |
+
"should_log": false,
|
| 450 |
+
"should_save": true,
|
| 451 |
+
"should_training_stop": true
|
| 452 |
+
},
|
| 453 |
+
"attributes": {}
|
| 454 |
+
}
|
| 455 |
+
},
|
| 456 |
+
"total_flos": 5.181183958187704e+18,
|
| 457 |
+
"train_batch_size": 8,
|
| 458 |
+
"trial_name": null,
|
| 459 |
+
"trial_params": null
|
| 460 |
+
}
|