Upload 7 files
Browse files- config.json +27 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +352 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "distilroberta-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3072,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"max_position_embeddings": 514,
|
| 17 |
+
"model_type": "roberta",
|
| 18 |
+
"num_attention_heads": 12,
|
| 19 |
+
"num_hidden_layers": 6,
|
| 20 |
+
"pad_token_id": 1,
|
| 21 |
+
"position_embedding_type": "absolute",
|
| 22 |
+
"torch_dtype": "float32",
|
| 23 |
+
"transformers_version": "4.29.2",
|
| 24 |
+
"type_vocab_size": 1,
|
| 25 |
+
"use_cache": true,
|
| 26 |
+
"vocab_size": 50265
|
| 27 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb4c006a4b3df152981071a1aefb8e42368f7404f3a4214b509f7ffa4ab9a0f5
|
| 3 |
+
size 657423685
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:48ee0133b6343e276cb0a9ba4755a66a80f4661ee7feeb9e5f368856bbeed33d
|
| 3 |
+
size 328721273
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b13d368549414e65beca691381a1f3b2157e1dbda6ba22df0a6ddc770f7bdbcd
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81c193d574ff1d1901d2e1f4e9a3aed191b4df1d90a0499d9b7cb27531f4fe73
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,352 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.927797833935019,
|
| 5 |
+
"global_step": 22000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.23,
|
| 12 |
+
"learning_rate": 1.9548736462093864e-05,
|
| 13 |
+
"loss": 1.999,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.45,
|
| 18 |
+
"learning_rate": 1.9097472924187726e-05,
|
| 19 |
+
"loss": 1.6311,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.68,
|
| 24 |
+
"learning_rate": 1.8646209386281592e-05,
|
| 25 |
+
"loss": 1.4725,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.9,
|
| 30 |
+
"learning_rate": 1.8194945848375454e-05,
|
| 31 |
+
"loss": 1.4118,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.0,
|
| 36 |
+
"eval_loss": 1.260017991065979,
|
| 37 |
+
"eval_runtime": 13.7559,
|
| 38 |
+
"eval_samples_per_second": 645.178,
|
| 39 |
+
"eval_steps_per_second": 80.693,
|
| 40 |
+
"step": 2216
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 1.13,
|
| 44 |
+
"learning_rate": 1.7743682310469313e-05,
|
| 45 |
+
"loss": 1.3416,
|
| 46 |
+
"step": 2500
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 1.35,
|
| 50 |
+
"learning_rate": 1.729241877256318e-05,
|
| 51 |
+
"loss": 1.2715,
|
| 52 |
+
"step": 3000
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.58,
|
| 56 |
+
"learning_rate": 1.684115523465704e-05,
|
| 57 |
+
"loss": 1.215,
|
| 58 |
+
"step": 3500
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 1.81,
|
| 62 |
+
"learning_rate": 1.6389891696750904e-05,
|
| 63 |
+
"loss": 1.1616,
|
| 64 |
+
"step": 4000
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 2.0,
|
| 68 |
+
"eval_loss": 1.0639451742172241,
|
| 69 |
+
"eval_runtime": 13.601,
|
| 70 |
+
"eval_samples_per_second": 652.527,
|
| 71 |
+
"eval_steps_per_second": 81.612,
|
| 72 |
+
"step": 4432
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 2.03,
|
| 76 |
+
"learning_rate": 1.5938628158844766e-05,
|
| 77 |
+
"loss": 1.1381,
|
| 78 |
+
"step": 4500
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 2.26,
|
| 82 |
+
"learning_rate": 1.548736462093863e-05,
|
| 83 |
+
"loss": 1.1219,
|
| 84 |
+
"step": 5000
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 2.48,
|
| 88 |
+
"learning_rate": 1.5036101083032493e-05,
|
| 89 |
+
"loss": 1.0613,
|
| 90 |
+
"step": 5500
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 2.71,
|
| 94 |
+
"learning_rate": 1.4584837545126355e-05,
|
| 95 |
+
"loss": 1.078,
|
| 96 |
+
"step": 6000
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 2.93,
|
| 100 |
+
"learning_rate": 1.413357400722022e-05,
|
| 101 |
+
"loss": 1.0446,
|
| 102 |
+
"step": 6500
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 3.0,
|
| 106 |
+
"eval_loss": 0.9587522149085999,
|
| 107 |
+
"eval_runtime": 13.6249,
|
| 108 |
+
"eval_samples_per_second": 651.379,
|
| 109 |
+
"eval_steps_per_second": 81.468,
|
| 110 |
+
"step": 6648
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 3.16,
|
| 114 |
+
"learning_rate": 1.368231046931408e-05,
|
| 115 |
+
"loss": 1.0442,
|
| 116 |
+
"step": 7000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 3.38,
|
| 120 |
+
"learning_rate": 1.3231046931407942e-05,
|
| 121 |
+
"loss": 1.0028,
|
| 122 |
+
"step": 7500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 3.61,
|
| 126 |
+
"learning_rate": 1.2779783393501806e-05,
|
| 127 |
+
"loss": 0.9996,
|
| 128 |
+
"step": 8000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 3.84,
|
| 132 |
+
"learning_rate": 1.2328519855595669e-05,
|
| 133 |
+
"loss": 0.974,
|
| 134 |
+
"step": 8500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 4.0,
|
| 138 |
+
"eval_loss": 0.8713343739509583,
|
| 139 |
+
"eval_runtime": 13.5623,
|
| 140 |
+
"eval_samples_per_second": 654.389,
|
| 141 |
+
"eval_steps_per_second": 81.845,
|
| 142 |
+
"step": 8864
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 4.06,
|
| 146 |
+
"learning_rate": 1.1877256317689531e-05,
|
| 147 |
+
"loss": 0.9667,
|
| 148 |
+
"step": 9000
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 4.29,
|
| 152 |
+
"learning_rate": 1.1425992779783395e-05,
|
| 153 |
+
"loss": 0.9457,
|
| 154 |
+
"step": 9500
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 4.51,
|
| 158 |
+
"learning_rate": 1.0974729241877257e-05,
|
| 159 |
+
"loss": 0.9311,
|
| 160 |
+
"step": 10000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 4.74,
|
| 164 |
+
"learning_rate": 1.0523465703971121e-05,
|
| 165 |
+
"loss": 0.9321,
|
| 166 |
+
"step": 10500
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 4.96,
|
| 170 |
+
"learning_rate": 1.0072202166064984e-05,
|
| 171 |
+
"loss": 0.9109,
|
| 172 |
+
"step": 11000
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 5.0,
|
| 176 |
+
"eval_loss": 0.8490720987319946,
|
| 177 |
+
"eval_runtime": 13.6964,
|
| 178 |
+
"eval_samples_per_second": 647.978,
|
| 179 |
+
"eval_steps_per_second": 81.043,
|
| 180 |
+
"step": 11080
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 5.19,
|
| 184 |
+
"learning_rate": 9.620938628158846e-06,
|
| 185 |
+
"loss": 0.8969,
|
| 186 |
+
"step": 11500
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 5.42,
|
| 190 |
+
"learning_rate": 9.169675090252709e-06,
|
| 191 |
+
"loss": 0.8924,
|
| 192 |
+
"step": 12000
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 5.64,
|
| 196 |
+
"learning_rate": 8.718411552346571e-06,
|
| 197 |
+
"loss": 0.8812,
|
| 198 |
+
"step": 12500
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 5.87,
|
| 202 |
+
"learning_rate": 8.267148014440433e-06,
|
| 203 |
+
"loss": 0.8651,
|
| 204 |
+
"step": 13000
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 6.0,
|
| 208 |
+
"eval_loss": 0.7949955463409424,
|
| 209 |
+
"eval_runtime": 13.8508,
|
| 210 |
+
"eval_samples_per_second": 640.758,
|
| 211 |
+
"eval_steps_per_second": 80.14,
|
| 212 |
+
"step": 13296
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 6.09,
|
| 216 |
+
"learning_rate": 7.815884476534297e-06,
|
| 217 |
+
"loss": 0.8299,
|
| 218 |
+
"step": 13500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 6.32,
|
| 222 |
+
"learning_rate": 7.36462093862816e-06,
|
| 223 |
+
"loss": 0.8698,
|
| 224 |
+
"step": 14000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 6.54,
|
| 228 |
+
"learning_rate": 6.913357400722023e-06,
|
| 229 |
+
"loss": 0.8441,
|
| 230 |
+
"step": 14500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 6.77,
|
| 234 |
+
"learning_rate": 6.4620938628158845e-06,
|
| 235 |
+
"loss": 0.8028,
|
| 236 |
+
"step": 15000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 6.99,
|
| 240 |
+
"learning_rate": 6.010830324909748e-06,
|
| 241 |
+
"loss": 0.8449,
|
| 242 |
+
"step": 15500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 7.0,
|
| 246 |
+
"eval_loss": 0.7730704545974731,
|
| 247 |
+
"eval_runtime": 13.7429,
|
| 248 |
+
"eval_samples_per_second": 645.79,
|
| 249 |
+
"eval_steps_per_second": 80.769,
|
| 250 |
+
"step": 15512
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 7.22,
|
| 254 |
+
"learning_rate": 5.559566787003611e-06,
|
| 255 |
+
"loss": 0.8188,
|
| 256 |
+
"step": 16000
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 7.45,
|
| 260 |
+
"learning_rate": 5.108303249097474e-06,
|
| 261 |
+
"loss": 0.8166,
|
| 262 |
+
"step": 16500
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 7.67,
|
| 266 |
+
"learning_rate": 4.6570397111913365e-06,
|
| 267 |
+
"loss": 0.8159,
|
| 268 |
+
"step": 17000
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 7.9,
|
| 272 |
+
"learning_rate": 4.205776173285199e-06,
|
| 273 |
+
"loss": 0.8304,
|
| 274 |
+
"step": 17500
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 8.0,
|
| 278 |
+
"eval_loss": 0.7481271028518677,
|
| 279 |
+
"eval_runtime": 7.0241,
|
| 280 |
+
"eval_samples_per_second": 1263.503,
|
| 281 |
+
"eval_steps_per_second": 158.027,
|
| 282 |
+
"step": 17728
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 8.12,
|
| 286 |
+
"learning_rate": 3.7545126353790616e-06,
|
| 287 |
+
"loss": 0.815,
|
| 288 |
+
"step": 18000
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 8.35,
|
| 292 |
+
"learning_rate": 3.303249097472924e-06,
|
| 293 |
+
"loss": 0.7999,
|
| 294 |
+
"step": 18500
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 8.57,
|
| 298 |
+
"learning_rate": 2.8519855595667872e-06,
|
| 299 |
+
"loss": 0.7977,
|
| 300 |
+
"step": 19000
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 8.8,
|
| 304 |
+
"learning_rate": 2.40072202166065e-06,
|
| 305 |
+
"loss": 0.7856,
|
| 306 |
+
"step": 19500
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 9.0,
|
| 310 |
+
"eval_loss": 0.7300745248794556,
|
| 311 |
+
"eval_runtime": 7.4235,
|
| 312 |
+
"eval_samples_per_second": 1195.525,
|
| 313 |
+
"eval_steps_per_second": 149.525,
|
| 314 |
+
"step": 19944
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 9.03,
|
| 318 |
+
"learning_rate": 1.949458483754513e-06,
|
| 319 |
+
"loss": 0.7805,
|
| 320 |
+
"step": 20000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 9.25,
|
| 324 |
+
"learning_rate": 1.4981949458483756e-06,
|
| 325 |
+
"loss": 0.7801,
|
| 326 |
+
"step": 20500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 9.48,
|
| 330 |
+
"learning_rate": 1.0469314079422384e-06,
|
| 331 |
+
"loss": 0.7528,
|
| 332 |
+
"step": 21000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 9.7,
|
| 336 |
+
"learning_rate": 5.956678700361012e-07,
|
| 337 |
+
"loss": 0.7698,
|
| 338 |
+
"step": 21500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 9.93,
|
| 342 |
+
"learning_rate": 1.444043321299639e-07,
|
| 343 |
+
"loss": 0.7997,
|
| 344 |
+
"step": 22000
|
| 345 |
+
}
|
| 346 |
+
],
|
| 347 |
+
"max_steps": 22160,
|
| 348 |
+
"num_train_epochs": 10,
|
| 349 |
+
"total_flos": 1139660369332650.0,
|
| 350 |
+
"trial_name": null,
|
| 351 |
+
"trial_params": null
|
| 352 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3520cc2928dde03a46d7756689306bd4592d01b86cb98f8df4eb0ce4c90016b1
|
| 3 |
+
size 3899
|