Upload folder using huggingface_hub
Browse files- head.pt +3 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +319 -0
- training_args.bin +3 -0
head.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c0858f64786ece9aee3b49091464a82fb78981c292dd87775d1d004a5ece5795
|
| 3 |
+
size 42374
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3d21d1a3f2d59ff771ba0c63db81e3f625fb3dc1c4c0b3371cce64bb28ea4940
|
| 3 |
+
size 15177767650
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77e72f0b05410bbd4ca8ae7b4b2a128906dfcadef413f799e494094d36ab5f6c
|
| 3 |
+
size 336006040
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:70baeded00dccc74bd1983a5270f31dfa41517005de6aadbf3fc2e13929870d8
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:505e140348cf428a705b42cb8aafd791b8d0b0ac55c70e82b985098cfa08a46a
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,319 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 18750,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.08,
|
| 13 |
+
"grad_norm": 18.305925369262695,
|
| 14 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 15 |
+
"loss": 1.635,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.16,
|
| 20 |
+
"grad_norm": 15.116157531738281,
|
| 21 |
+
"learning_rate": 2.6666666666666667e-05,
|
| 22 |
+
"loss": 0.5409,
|
| 23 |
+
"step": 1000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.24,
|
| 27 |
+
"grad_norm": 13.341463088989258,
|
| 28 |
+
"learning_rate": 4e-05,
|
| 29 |
+
"loss": 0.4258,
|
| 30 |
+
"step": 1500
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.32,
|
| 34 |
+
"grad_norm": 11.549241065979004,
|
| 35 |
+
"learning_rate": 4.9993231029486544e-05,
|
| 36 |
+
"loss": 0.3779,
|
| 37 |
+
"step": 2000
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.4,
|
| 41 |
+
"grad_norm": 13.320897102355957,
|
| 42 |
+
"learning_rate": 4.983095894354858e-05,
|
| 43 |
+
"loss": 0.3566,
|
| 44 |
+
"step": 2500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.48,
|
| 48 |
+
"grad_norm": 15.671639442443848,
|
| 49 |
+
"learning_rate": 4.9453690018345144e-05,
|
| 50 |
+
"loss": 0.3249,
|
| 51 |
+
"step": 3000
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.56,
|
| 55 |
+
"grad_norm": 10.806138038635254,
|
| 56 |
+
"learning_rate": 4.88646908061933e-05,
|
| 57 |
+
"loss": 0.2979,
|
| 58 |
+
"step": 3500
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.64,
|
| 62 |
+
"grad_norm": 0.03432423993945122,
|
| 63 |
+
"learning_rate": 4.806906110888606e-05,
|
| 64 |
+
"loss": 0.2806,
|
| 65 |
+
"step": 4000
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.72,
|
| 69 |
+
"grad_norm": 11.576502799987793,
|
| 70 |
+
"learning_rate": 4.707368982147318e-05,
|
| 71 |
+
"loss": 0.257,
|
| 72 |
+
"step": 4500
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.8,
|
| 76 |
+
"grad_norm": 0.5009346008300781,
|
| 77 |
+
"learning_rate": 4.588719528532342e-05,
|
| 78 |
+
"loss": 0.2509,
|
| 79 |
+
"step": 5000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.88,
|
| 83 |
+
"grad_norm": 10.398884773254395,
|
| 84 |
+
"learning_rate": 4.4519850666916484e-05,
|
| 85 |
+
"loss": 0.2642,
|
| 86 |
+
"step": 5500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.96,
|
| 90 |
+
"grad_norm": 0.5586361885070801,
|
| 91 |
+
"learning_rate": 4.2983495008466276e-05,
|
| 92 |
+
"loss": 0.2452,
|
| 93 |
+
"step": 6000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 1.0,
|
| 97 |
+
"eval_accuracy": 0.9526,
|
| 98 |
+
"eval_loss": 0.2175797075033188,
|
| 99 |
+
"eval_runtime": 288.1127,
|
| 100 |
+
"eval_samples_per_second": 17.354,
|
| 101 |
+
"eval_steps_per_second": 2.169,
|
| 102 |
+
"step": 6250
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.04,
|
| 106 |
+
"grad_norm": 15.244111061096191,
|
| 107 |
+
"learning_rate": 4.129143072053638e-05,
|
| 108 |
+
"loss": 0.1726,
|
| 109 |
+
"step": 6500
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 1.12,
|
| 113 |
+
"grad_norm": 23.2961368560791,
|
| 114 |
+
"learning_rate": 3.945830840419966e-05,
|
| 115 |
+
"loss": 0.1662,
|
| 116 |
+
"step": 7000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 1.2,
|
| 120 |
+
"grad_norm": 48.89887619018555,
|
| 121 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 122 |
+
"loss": 0.1588,
|
| 123 |
+
"step": 7500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 1.28,
|
| 127 |
+
"grad_norm": 1.5729694496258162e-05,
|
| 128 |
+
"learning_rate": 3.543346136204545e-05,
|
| 129 |
+
"loss": 0.1415,
|
| 130 |
+
"step": 8000
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 1.3599999999999999,
|
| 134 |
+
"grad_norm": 0.0002951535861939192,
|
| 135 |
+
"learning_rate": 3.327658544712395e-05,
|
| 136 |
+
"loss": 0.1514,
|
| 137 |
+
"step": 8500
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.44,
|
| 141 |
+
"grad_norm": 15.348397254943848,
|
| 142 |
+
"learning_rate": 3.104804738999169e-05,
|
| 143 |
+
"loss": 0.1365,
|
| 144 |
+
"step": 9000
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.52,
|
| 148 |
+
"grad_norm": 0.00016808633517939597,
|
| 149 |
+
"learning_rate": 2.876714280623708e-05,
|
| 150 |
+
"loss": 0.1339,
|
| 151 |
+
"step": 9500
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 1.6,
|
| 155 |
+
"grad_norm": 30.164745330810547,
|
| 156 |
+
"learning_rate": 2.6453620722761896e-05,
|
| 157 |
+
"loss": 0.138,
|
| 158 |
+
"step": 10000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.6800000000000002,
|
| 162 |
+
"grad_norm": 6.264598050620407e-05,
|
| 163 |
+
"learning_rate": 2.4127512582437485e-05,
|
| 164 |
+
"loss": 0.1395,
|
| 165 |
+
"step": 10500
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.76,
|
| 169 |
+
"grad_norm": 0.0032692451495677233,
|
| 170 |
+
"learning_rate": 2.1808958803485136e-05,
|
| 171 |
+
"loss": 0.1151,
|
| 172 |
+
"step": 11000
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 1.8399999999999999,
|
| 176 |
+
"grad_norm": 0.0009920683223754168,
|
| 177 |
+
"learning_rate": 1.9518034395302414e-05,
|
| 178 |
+
"loss": 0.1159,
|
| 179 |
+
"step": 11500
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.92,
|
| 183 |
+
"grad_norm": 0.0016635229112580419,
|
| 184 |
+
"learning_rate": 1.7274575140626318e-05,
|
| 185 |
+
"loss": 0.1204,
|
| 186 |
+
"step": 12000
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.0,
|
| 190 |
+
"grad_norm": 0.007010475266724825,
|
| 191 |
+
"learning_rate": 1.509800584902108e-05,
|
| 192 |
+
"loss": 0.1096,
|
| 193 |
+
"step": 12500
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 2.0,
|
| 197 |
+
"eval_accuracy": 0.9636,
|
| 198 |
+
"eval_loss": 0.16214776039123535,
|
| 199 |
+
"eval_runtime": 287.6078,
|
| 200 |
+
"eval_samples_per_second": 17.385,
|
| 201 |
+
"eval_steps_per_second": 2.173,
|
| 202 |
+
"step": 12500
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 2.08,
|
| 206 |
+
"grad_norm": 0.001354572712443769,
|
| 207 |
+
"learning_rate": 1.3007172168743854e-05,
|
| 208 |
+
"loss": 0.0312,
|
| 209 |
+
"step": 13000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 2.16,
|
| 213 |
+
"grad_norm": 8.306991730933078e-06,
|
| 214 |
+
"learning_rate": 1.1020177413231334e-05,
|
| 215 |
+
"loss": 0.0281,
|
| 216 |
+
"step": 13500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 2.24,
|
| 220 |
+
"grad_norm": 2.6390673156129196e-05,
|
| 221 |
+
"learning_rate": 9.154225815032242e-06,
|
| 222 |
+
"loss": 0.0361,
|
| 223 |
+
"step": 14000
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 2.32,
|
| 227 |
+
"grad_norm": 0.00028656359063461423,
|
| 228 |
+
"learning_rate": 7.4254735643584564e-06,
|
| 229 |
+
"loss": 0.06,
|
| 230 |
+
"step": 14500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.4,
|
| 234 |
+
"grad_norm": 3.382603972568177e-05,
|
| 235 |
+
"learning_rate": 5.848888922025553e-06,
|
| 236 |
+
"loss": 0.0392,
|
| 237 |
+
"step": 15000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 2.48,
|
| 241 |
+
"grad_norm": 0.001894937246106565,
|
| 242 |
+
"learning_rate": 4.438122617983443e-06,
|
| 243 |
+
"loss": 0.0332,
|
| 244 |
+
"step": 15500
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 2.56,
|
| 248 |
+
"grad_norm": 2.6853955205297098e-05,
|
| 249 |
+
"learning_rate": 3.205389657580943e-06,
|
| 250 |
+
"loss": 0.0362,
|
| 251 |
+
"step": 16000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 2.64,
|
| 255 |
+
"grad_norm": 0.7828443646430969,
|
| 256 |
+
"learning_rate": 2.1613635589349756e-06,
|
| 257 |
+
"loss": 0.0324,
|
| 258 |
+
"step": 16500
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 2.7199999999999998,
|
| 262 |
+
"grad_norm": 0.1359904706478119,
|
| 263 |
+
"learning_rate": 1.31508393714177e-06,
|
| 264 |
+
"loss": 0.0474,
|
| 265 |
+
"step": 17000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 2.8,
|
| 269 |
+
"grad_norm": 0.0003148973046336323,
|
| 270 |
+
"learning_rate": 6.738782355044049e-07,
|
| 271 |
+
"loss": 0.0387,
|
| 272 |
+
"step": 17500
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.88,
|
| 276 |
+
"grad_norm": 0.00042670563561841846,
|
| 277 |
+
"learning_rate": 2.4329828146074095e-07,
|
| 278 |
+
"loss": 0.028,
|
| 279 |
+
"step": 18000
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 2.96,
|
| 283 |
+
"grad_norm": 9.56340809352696e-05,
|
| 284 |
+
"learning_rate": 2.7072216536885853e-08,
|
| 285 |
+
"loss": 0.0349,
|
| 286 |
+
"step": 18500
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 3.0,
|
| 290 |
+
"eval_accuracy": 0.9716,
|
| 291 |
+
"eval_loss": 0.18466220796108246,
|
| 292 |
+
"eval_runtime": 286.2615,
|
| 293 |
+
"eval_samples_per_second": 17.467,
|
| 294 |
+
"eval_steps_per_second": 2.183,
|
| 295 |
+
"step": 18750
|
| 296 |
+
}
|
| 297 |
+
],
|
| 298 |
+
"logging_steps": 500,
|
| 299 |
+
"max_steps": 18750,
|
| 300 |
+
"num_input_tokens_seen": 0,
|
| 301 |
+
"num_train_epochs": 3,
|
| 302 |
+
"save_steps": 500,
|
| 303 |
+
"stateful_callbacks": {
|
| 304 |
+
"TrainerControl": {
|
| 305 |
+
"args": {
|
| 306 |
+
"should_epoch_stop": false,
|
| 307 |
+
"should_evaluate": false,
|
| 308 |
+
"should_log": false,
|
| 309 |
+
"should_save": true,
|
| 310 |
+
"should_training_stop": true
|
| 311 |
+
},
|
| 312 |
+
"attributes": {}
|
| 313 |
+
}
|
| 314 |
+
},
|
| 315 |
+
"total_flos": 0.0,
|
| 316 |
+
"train_batch_size": 8,
|
| 317 |
+
"trial_name": null,
|
| 318 |
+
"trial_params": null
|
| 319 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e655c9f4ec242123684bdefde757c27947012689a113dea3c6edf3c217226334
|
| 3 |
+
size 5368
|