Upload folder using huggingface_hub
Browse files- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +449 -0
- training_args.bin +3 -0
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e95e33471399ccfc7bed85d400c59f70fd5463e4db899c05e6657a938717603f
|
| 3 |
+
size 2492275004
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ded95083a69212f90b065f310cf63b3552b500af09f54416ddbc1aba99d20a0
|
| 3 |
+
size 14308
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:69dae4013c514d4dd5af226e2985d14d77eacee0977117596e76139a814ff8d0
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,449 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.28394800424575806,
|
| 3 |
+
"best_model_checkpoint": "/kaggle/working/wav2vec2-speech-emotion/checkpoint-1100",
|
| 4 |
+
"epoch": 12.64367816091954,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 1100,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.28735632183908044,
|
| 13 |
+
"grad_norm": 81616.25,
|
| 14 |
+
"learning_rate": 2.5e-06,
|
| 15 |
+
"loss": 1.9364,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.5747126436781609,
|
| 20 |
+
"grad_norm": 113624.453125,
|
| 21 |
+
"learning_rate": 5e-06,
|
| 22 |
+
"loss": 1.9286,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.8620689655172413,
|
| 27 |
+
"grad_norm": 74490.7890625,
|
| 28 |
+
"learning_rate": 7.500000000000001e-06,
|
| 29 |
+
"loss": 1.9194,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 1.1494252873563218,
|
| 34 |
+
"grad_norm": 94447.8125,
|
| 35 |
+
"learning_rate": 1e-05,
|
| 36 |
+
"loss": 1.9086,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 1.1494252873563218,
|
| 41 |
+
"eval_accuracy": 0.2261904761904762,
|
| 42 |
+
"eval_loss": 1.9561063051223755,
|
| 43 |
+
"eval_runtime": 10.2339,
|
| 44 |
+
"eval_samples_per_second": 16.416,
|
| 45 |
+
"eval_steps_per_second": 1.075,
|
| 46 |
+
"step": 100
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 1.4367816091954024,
|
| 50 |
+
"grad_norm": 60535.6171875,
|
| 51 |
+
"learning_rate": 1.25e-05,
|
| 52 |
+
"loss": 1.8885,
|
| 53 |
+
"step": 125
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 1.7241379310344827,
|
| 57 |
+
"grad_norm": 118006.9765625,
|
| 58 |
+
"learning_rate": 1.5000000000000002e-05,
|
| 59 |
+
"loss": 1.8847,
|
| 60 |
+
"step": 150
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 2.0114942528735633,
|
| 64 |
+
"grad_norm": 109447.6953125,
|
| 65 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 66 |
+
"loss": 1.8595,
|
| 67 |
+
"step": 175
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 2.2988505747126435,
|
| 71 |
+
"grad_norm": 131271.265625,
|
| 72 |
+
"learning_rate": 2e-05,
|
| 73 |
+
"loss": 1.7796,
|
| 74 |
+
"step": 200
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.2988505747126435,
|
| 78 |
+
"eval_accuracy": 0.42857142857142855,
|
| 79 |
+
"eval_loss": 1.737054705619812,
|
| 80 |
+
"eval_runtime": 10.3261,
|
| 81 |
+
"eval_samples_per_second": 16.269,
|
| 82 |
+
"eval_steps_per_second": 1.065,
|
| 83 |
+
"step": 200
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 2.586206896551724,
|
| 87 |
+
"grad_norm": 222429.421875,
|
| 88 |
+
"learning_rate": 1.9974751105436266e-05,
|
| 89 |
+
"loss": 1.6504,
|
| 90 |
+
"step": 225
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 2.873563218390805,
|
| 94 |
+
"grad_norm": 482113.40625,
|
| 95 |
+
"learning_rate": 1.98991319230804e-05,
|
| 96 |
+
"loss": 1.4559,
|
| 97 |
+
"step": 250
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 3.160919540229885,
|
| 101 |
+
"grad_norm": 375893.84375,
|
| 102 |
+
"learning_rate": 1.9773524313084857e-05,
|
| 103 |
+
"loss": 1.2633,
|
| 104 |
+
"step": 275
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 3.4482758620689653,
|
| 108 |
+
"grad_norm": 542423.5625,
|
| 109 |
+
"learning_rate": 1.959856256610988e-05,
|
| 110 |
+
"loss": 1.1442,
|
| 111 |
+
"step": 300
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 3.4482758620689653,
|
| 115 |
+
"eval_accuracy": 0.6726190476190477,
|
| 116 |
+
"eval_loss": 1.0434563159942627,
|
| 117 |
+
"eval_runtime": 10.2866,
|
| 118 |
+
"eval_samples_per_second": 16.332,
|
| 119 |
+
"eval_steps_per_second": 1.069,
|
| 120 |
+
"step": 300
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.735632183908046,
|
| 124 |
+
"grad_norm": 321237.03125,
|
| 125 |
+
"learning_rate": 1.937513020029588e-05,
|
| 126 |
+
"loss": 1.0287,
|
| 127 |
+
"step": 325
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.022988505747127,
|
| 131 |
+
"grad_norm": 494109.96875,
|
| 132 |
+
"learning_rate": 1.9104355499692166e-05,
|
| 133 |
+
"loss": 0.929,
|
| 134 |
+
"step": 350
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 4.310344827586207,
|
| 138 |
+
"grad_norm": 399620.6875,
|
| 139 |
+
"learning_rate": 1.8787605816671956e-05,
|
| 140 |
+
"loss": 0.809,
|
| 141 |
+
"step": 375
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 4.597701149425287,
|
| 145 |
+
"grad_norm": 320774.09375,
|
| 146 |
+
"learning_rate": 1.8426480667105178e-05,
|
| 147 |
+
"loss": 0.7284,
|
| 148 |
+
"step": 400
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 4.597701149425287,
|
| 152 |
+
"eval_accuracy": 0.7857142857142857,
|
| 153 |
+
"eval_loss": 0.8009253144264221,
|
| 154 |
+
"eval_runtime": 10.0989,
|
| 155 |
+
"eval_samples_per_second": 16.635,
|
| 156 |
+
"eval_steps_per_second": 1.089,
|
| 157 |
+
"step": 400
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 4.885057471264368,
|
| 161 |
+
"grad_norm": 316548.5625,
|
| 162 |
+
"learning_rate": 1.8022803653156983e-05,
|
| 163 |
+
"loss": 0.7361,
|
| 164 |
+
"step": 425
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 5.172413793103448,
|
| 168 |
+
"grad_norm": 592323.0,
|
| 169 |
+
"learning_rate": 1.757861325449997e-05,
|
| 170 |
+
"loss": 0.6226,
|
| 171 |
+
"step": 450
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 5.459770114942529,
|
| 175 |
+
"grad_norm": 686948.0625,
|
| 176 |
+
"learning_rate": 1.7096152534442515e-05,
|
| 177 |
+
"loss": 0.5987,
|
| 178 |
+
"step": 475
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 5.747126436781609,
|
| 182 |
+
"grad_norm": 314310.65625,
|
| 183 |
+
"learning_rate": 1.6577857812954994e-05,
|
| 184 |
+
"loss": 0.5259,
|
| 185 |
+
"step": 500
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 5.747126436781609,
|
| 189 |
+
"eval_accuracy": 0.8511904761904762,
|
| 190 |
+
"eval_loss": 0.5674353837966919,
|
| 191 |
+
"eval_runtime": 10.2098,
|
| 192 |
+
"eval_samples_per_second": 16.455,
|
| 193 |
+
"eval_steps_per_second": 1.077,
|
| 194 |
+
"step": 500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 6.0344827586206895,
|
| 198 |
+
"grad_norm": 747515.625,
|
| 199 |
+
"learning_rate": 1.6026346363792565e-05,
|
| 200 |
+
"loss": 0.534,
|
| 201 |
+
"step": 525
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 6.32183908045977,
|
| 205 |
+
"grad_norm": 308987.9375,
|
| 206 |
+
"learning_rate": 1.5444403197841345e-05,
|
| 207 |
+
"loss": 0.448,
|
| 208 |
+
"step": 550
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 6.609195402298851,
|
| 212 |
+
"grad_norm": 1300700.75,
|
| 213 |
+
"learning_rate": 1.4834966999429179e-05,
|
| 214 |
+
"loss": 0.439,
|
| 215 |
+
"step": 575
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 6.896551724137931,
|
| 219 |
+
"grad_norm": 359445.15625,
|
| 220 |
+
"learning_rate": 1.4201115286619464e-05,
|
| 221 |
+
"loss": 0.3734,
|
| 222 |
+
"step": 600
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 6.896551724137931,
|
| 226 |
+
"eval_accuracy": 0.8273809523809523,
|
| 227 |
+
"eval_loss": 0.6119872331619263,
|
| 228 |
+
"eval_runtime": 10.1904,
|
| 229 |
+
"eval_samples_per_second": 16.486,
|
| 230 |
+
"eval_steps_per_second": 1.079,
|
| 231 |
+
"step": 600
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 7.183908045977011,
|
| 235 |
+
"grad_norm": 156436.734375,
|
| 236 |
+
"learning_rate": 1.3546048870425356e-05,
|
| 237 |
+
"loss": 0.331,
|
| 238 |
+
"step": 625
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 7.471264367816092,
|
| 242 |
+
"grad_norm": 920202.625,
|
| 243 |
+
"learning_rate": 1.2873075691421808e-05,
|
| 244 |
+
"loss": 0.3074,
|
| 245 |
+
"step": 650
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 7.758620689655173,
|
| 249 |
+
"grad_norm": 501704.25,
|
| 250 |
+
"learning_rate": 1.2185594115376991e-05,
|
| 251 |
+
"loss": 0.2982,
|
| 252 |
+
"step": 675
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 8.045977011494253,
|
| 256 |
+
"grad_norm": 210118.0625,
|
| 257 |
+
"learning_rate": 1.1487075772256517e-05,
|
| 258 |
+
"loss": 0.2702,
|
| 259 |
+
"step": 700
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 8.045977011494253,
|
| 263 |
+
"eval_accuracy": 0.9464285714285714,
|
| 264 |
+
"eval_loss": 0.2987803518772125,
|
| 265 |
+
"eval_runtime": 10.1874,
|
| 266 |
+
"eval_samples_per_second": 16.491,
|
| 267 |
+
"eval_steps_per_second": 1.08,
|
| 268 |
+
"step": 700
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 8.333333333333334,
|
| 272 |
+
"grad_norm": 529805.875,
|
| 273 |
+
"learning_rate": 1.0781048025259648e-05,
|
| 274 |
+
"loss": 0.2609,
|
| 275 |
+
"step": 725
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 8.620689655172415,
|
| 279 |
+
"grad_norm": 280091.625,
|
| 280 |
+
"learning_rate": 1.0071076158414977e-05,
|
| 281 |
+
"loss": 0.2612,
|
| 282 |
+
"step": 750
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 8.908045977011493,
|
| 286 |
+
"grad_norm": 457920.8125,
|
| 287 |
+
"learning_rate": 9.360745372684346e-06,
|
| 288 |
+
"loss": 0.2353,
|
| 289 |
+
"step": 775
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 9.195402298850574,
|
| 293 |
+
"grad_norm": 791098.25,
|
| 294 |
+
"learning_rate": 8.653642681490608e-06,
|
| 295 |
+
"loss": 0.2171,
|
| 296 |
+
"step": 800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 9.195402298850574,
|
| 300 |
+
"eval_accuracy": 0.8928571428571429,
|
| 301 |
+
"eval_loss": 0.34595435857772827,
|
| 302 |
+
"eval_runtime": 10.2959,
|
| 303 |
+
"eval_samples_per_second": 16.317,
|
| 304 |
+
"eval_steps_per_second": 1.068,
|
| 305 |
+
"step": 800
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 9.482758620689655,
|
| 309 |
+
"grad_norm": 490233.0,
|
| 310 |
+
"learning_rate": 7.953338797092902e-06,
|
| 311 |
+
"loss": 0.1884,
|
| 312 |
+
"step": 825
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 9.770114942528735,
|
| 316 |
+
"grad_norm": 605037.8125,
|
| 317 |
+
"learning_rate": 7.263370099279173e-06,
|
| 318 |
+
"loss": 0.1573,
|
| 319 |
+
"step": 850
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 10.057471264367816,
|
| 323 |
+
"grad_norm": 504736.34375,
|
| 324 |
+
"learning_rate": 6.587220777430097e-06,
|
| 325 |
+
"loss": 0.1424,
|
| 326 |
+
"step": 875
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 10.344827586206897,
|
| 330 |
+
"grad_norm": 559532.25,
|
| 331 |
+
"learning_rate": 5.928305236133016e-06,
|
| 332 |
+
"loss": 0.1634,
|
| 333 |
+
"step": 900
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 10.344827586206897,
|
| 337 |
+
"eval_accuracy": 0.9345238095238095,
|
| 338 |
+
"eval_loss": 0.29700732231140137,
|
| 339 |
+
"eval_runtime": 10.0813,
|
| 340 |
+
"eval_samples_per_second": 16.665,
|
| 341 |
+
"eval_steps_per_second": 1.091,
|
| 342 |
+
"step": 900
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 10.632183908045977,
|
| 346 |
+
"grad_norm": 426639.75,
|
| 347 |
+
"learning_rate": 5.2899508531936526e-06,
|
| 348 |
+
"loss": 0.1735,
|
| 349 |
+
"step": 925
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 10.919540229885058,
|
| 353 |
+
"grad_norm": 421293.25,
|
| 354 |
+
"learning_rate": 4.675381177113837e-06,
|
| 355 |
+
"loss": 0.2091,
|
| 356 |
+
"step": 950
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 11.206896551724139,
|
| 360 |
+
"grad_norm": 1287860.25,
|
| 361 |
+
"learning_rate": 4.087699648884248e-06,
|
| 362 |
+
"loss": 0.1832,
|
| 363 |
+
"step": 975
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 11.494252873563218,
|
| 367 |
+
"grad_norm": 596496.875,
|
| 368 |
+
"learning_rate": 3.529873930293546e-06,
|
| 369 |
+
"loss": 0.1483,
|
| 370 |
+
"step": 1000
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 11.494252873563218,
|
| 374 |
+
"eval_accuracy": 0.9226190476190477,
|
| 375 |
+
"eval_loss": 0.3032456934452057,
|
| 376 |
+
"eval_runtime": 10.027,
|
| 377 |
+
"eval_samples_per_second": 16.755,
|
| 378 |
+
"eval_steps_per_second": 1.097,
|
| 379 |
+
"step": 1000
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 11.781609195402298,
|
| 383 |
+
"grad_norm": 223663.953125,
|
| 384 |
+
"learning_rate": 3.004720917892464e-06,
|
| 385 |
+
"loss": 0.1588,
|
| 386 |
+
"step": 1025
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 12.068965517241379,
|
| 390 |
+
"grad_norm": 55755.37109375,
|
| 391 |
+
"learning_rate": 2.514892518288988e-06,
|
| 392 |
+
"loss": 0.1392,
|
| 393 |
+
"step": 1050
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 12.35632183908046,
|
| 397 |
+
"grad_norm": 260650.140625,
|
| 398 |
+
"learning_rate": 2.0628622566063063e-06,
|
| 399 |
+
"loss": 0.1493,
|
| 400 |
+
"step": 1075
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 12.64367816091954,
|
| 404 |
+
"grad_norm": 60438.0390625,
|
| 405 |
+
"learning_rate": 1.6509127857277784e-06,
|
| 406 |
+
"loss": 0.1077,
|
| 407 |
+
"step": 1100
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 12.64367816091954,
|
| 411 |
+
"eval_accuracy": 0.9166666666666666,
|
| 412 |
+
"eval_loss": 0.28394800424575806,
|
| 413 |
+
"eval_runtime": 10.0833,
|
| 414 |
+
"eval_samples_per_second": 16.661,
|
| 415 |
+
"eval_steps_per_second": 1.091,
|
| 416 |
+
"step": 1100
|
| 417 |
+
}
|
| 418 |
+
],
|
| 419 |
+
"logging_steps": 25,
|
| 420 |
+
"max_steps": 1305,
|
| 421 |
+
"num_input_tokens_seen": 0,
|
| 422 |
+
"num_train_epochs": 15,
|
| 423 |
+
"save_steps": 100,
|
| 424 |
+
"stateful_callbacks": {
|
| 425 |
+
"EarlyStoppingCallback": {
|
| 426 |
+
"args": {
|
| 427 |
+
"early_stopping_patience": 3,
|
| 428 |
+
"early_stopping_threshold": 0.0
|
| 429 |
+
},
|
| 430 |
+
"attributes": {
|
| 431 |
+
"early_stopping_patience_counter": 0
|
| 432 |
+
}
|
| 433 |
+
},
|
| 434 |
+
"TrainerControl": {
|
| 435 |
+
"args": {
|
| 436 |
+
"should_epoch_stop": false,
|
| 437 |
+
"should_evaluate": false,
|
| 438 |
+
"should_log": false,
|
| 439 |
+
"should_save": true,
|
| 440 |
+
"should_training_stop": false
|
| 441 |
+
},
|
| 442 |
+
"attributes": {}
|
| 443 |
+
}
|
| 444 |
+
},
|
| 445 |
+
"total_flos": 4.682951691910107e+18,
|
| 446 |
+
"train_batch_size": 16,
|
| 447 |
+
"trial_name": null,
|
| 448 |
+
"trial_params": null
|
| 449 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:083d3480b5664c2448ad1338e2bc8728268febcfbec1253ec2c060e4c6628bf1
|
| 3 |
+
size 5432
|