Jingxuan Tang
commited on
Commit
·
3d9a836
1
Parent(s):
c4e98a0
Upload 3 files
Browse files- adapter_config.json +20 -0
- adapter_model.bin +3 -0
- trainer_state.json +291 -0
adapter_config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "indiejoseph/cantonese-llama-2-7b-oasst-v1",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 32,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 16,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"down_proj"
|
| 18 |
+
],
|
| 19 |
+
"task_type": "CAUSAL_LM"
|
| 20 |
+
}
|
adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0fa37fb870780f26aa1e9127f6dcfc5b66b835d1f9dc63cfa9c06ccc18d58d01
|
| 3 |
+
size 15476477
|
trainer_state.json
ADDED
|
@@ -0,0 +1,291 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8720703125,
|
| 3 |
+
"best_model_checkpoint": "./alma-7b-parallel-ft-lora-canto/checkpoint-589",
|
| 4 |
+
"epoch": 1.99673735725938,
|
| 5 |
+
"global_step": 612,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.1,
|
| 12 |
+
"learning_rate": 0.0006831300510639731,
|
| 13 |
+
"loss": 1.8545,
|
| 14 |
+
"step": 31
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.1,
|
| 18 |
+
"eval_loss": 1.275390625,
|
| 19 |
+
"eval_runtime": 161.389,
|
| 20 |
+
"eval_samples_per_second": 26.049,
|
| 21 |
+
"eval_steps_per_second": 1.63,
|
| 22 |
+
"step": 31
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"epoch": 0.2,
|
| 26 |
+
"learning_rate": 0.0003944053188733077,
|
| 27 |
+
"loss": 1.2078,
|
| 28 |
+
"step": 62
|
| 29 |
+
},
|
| 30 |
+
{
|
| 31 |
+
"epoch": 0.2,
|
| 32 |
+
"eval_loss": 1.1513671875,
|
| 33 |
+
"eval_runtime": 161.4937,
|
| 34 |
+
"eval_samples_per_second": 26.032,
|
| 35 |
+
"eval_steps_per_second": 1.629,
|
| 36 |
+
"step": 62
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 0.3,
|
| 40 |
+
"learning_rate": 0.00030348848933344196,
|
| 41 |
+
"loss": 1.1077,
|
| 42 |
+
"step": 93
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 0.3,
|
| 46 |
+
"eval_loss": 1.0869140625,
|
| 47 |
+
"eval_runtime": 161.573,
|
| 48 |
+
"eval_samples_per_second": 26.019,
|
| 49 |
+
"eval_steps_per_second": 1.628,
|
| 50 |
+
"step": 93
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.4,
|
| 54 |
+
"learning_rate": 0.00025577443333465435,
|
| 55 |
+
"loss": 1.067,
|
| 56 |
+
"step": 124
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.4,
|
| 60 |
+
"eval_loss": 1.0439453125,
|
| 61 |
+
"eval_runtime": 161.4591,
|
| 62 |
+
"eval_samples_per_second": 26.038,
|
| 63 |
+
"eval_steps_per_second": 1.629,
|
| 64 |
+
"step": 124
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.51,
|
| 68 |
+
"learning_rate": 0.0002252213082307254,
|
| 69 |
+
"loss": 1.0179,
|
| 70 |
+
"step": 155
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 0.51,
|
| 74 |
+
"eval_loss": 1.015625,
|
| 75 |
+
"eval_runtime": 161.6031,
|
| 76 |
+
"eval_samples_per_second": 26.014,
|
| 77 |
+
"eval_steps_per_second": 1.627,
|
| 78 |
+
"step": 155
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.61,
|
| 82 |
+
"learning_rate": 0.00020351933162035313,
|
| 83 |
+
"loss": 0.9852,
|
| 84 |
+
"step": 186
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.61,
|
| 88 |
+
"eval_loss": 0.98876953125,
|
| 89 |
+
"eval_runtime": 161.723,
|
| 90 |
+
"eval_samples_per_second": 25.995,
|
| 91 |
+
"eval_steps_per_second": 1.626,
|
| 92 |
+
"step": 186
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.71,
|
| 96 |
+
"learning_rate": 0.0001870828693386971,
|
| 97 |
+
"loss": 0.9682,
|
| 98 |
+
"step": 217
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.71,
|
| 102 |
+
"eval_loss": 0.97314453125,
|
| 103 |
+
"eval_runtime": 161.5091,
|
| 104 |
+
"eval_samples_per_second": 26.029,
|
| 105 |
+
"eval_steps_per_second": 1.628,
|
| 106 |
+
"step": 217
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 0.81,
|
| 110 |
+
"learning_rate": 0.00017407765595569785,
|
| 111 |
+
"loss": 0.9565,
|
| 112 |
+
"step": 248
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.81,
|
| 116 |
+
"eval_loss": 0.95751953125,
|
| 117 |
+
"eval_runtime": 161.4684,
|
| 118 |
+
"eval_samples_per_second": 26.036,
|
| 119 |
+
"eval_steps_per_second": 1.629,
|
| 120 |
+
"step": 248
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.91,
|
| 124 |
+
"learning_rate": 0.00016345506187300654,
|
| 125 |
+
"loss": 0.9322,
|
| 126 |
+
"step": 279
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 0.91,
|
| 130 |
+
"eval_loss": 0.94091796875,
|
| 131 |
+
"eval_runtime": 161.6846,
|
| 132 |
+
"eval_samples_per_second": 26.001,
|
| 133 |
+
"eval_steps_per_second": 1.627,
|
| 134 |
+
"step": 279
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 1.01,
|
| 138 |
+
"learning_rate": 0.0001545664419689318,
|
| 139 |
+
"loss": 0.9079,
|
| 140 |
+
"step": 310
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.01,
|
| 144 |
+
"eval_loss": 0.93212890625,
|
| 145 |
+
"eval_runtime": 159.6077,
|
| 146 |
+
"eval_samples_per_second": 26.34,
|
| 147 |
+
"eval_steps_per_second": 1.648,
|
| 148 |
+
"step": 310
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 1.11,
|
| 152 |
+
"learning_rate": 0.00014698618394803282,
|
| 153 |
+
"loss": 0.8272,
|
| 154 |
+
"step": 341
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 1.11,
|
| 158 |
+
"eval_loss": 0.92333984375,
|
| 159 |
+
"eval_runtime": 159.6433,
|
| 160 |
+
"eval_samples_per_second": 26.334,
|
| 161 |
+
"eval_steps_per_second": 1.647,
|
| 162 |
+
"step": 341
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 1.21,
|
| 166 |
+
"learning_rate": 0.0001404218994998819,
|
| 167 |
+
"loss": 0.7966,
|
| 168 |
+
"step": 372
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.21,
|
| 172 |
+
"eval_loss": 0.92041015625,
|
| 173 |
+
"eval_runtime": 159.6168,
|
| 174 |
+
"eval_samples_per_second": 26.338,
|
| 175 |
+
"eval_steps_per_second": 1.648,
|
| 176 |
+
"step": 372
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.31,
|
| 180 |
+
"learning_rate": 0.00013466519604525415,
|
| 181 |
+
"loss": 0.7947,
|
| 182 |
+
"step": 403
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.31,
|
| 186 |
+
"eval_loss": 0.9150390625,
|
| 187 |
+
"eval_runtime": 159.659,
|
| 188 |
+
"eval_samples_per_second": 26.331,
|
| 189 |
+
"eval_steps_per_second": 1.647,
|
| 190 |
+
"step": 403
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 1.42,
|
| 194 |
+
"learning_rate": 0.00012956299912940142,
|
| 195 |
+
"loss": 0.8089,
|
| 196 |
+
"step": 434
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 1.42,
|
| 200 |
+
"eval_loss": 0.9052734375,
|
| 201 |
+
"eval_runtime": 159.599,
|
| 202 |
+
"eval_samples_per_second": 26.341,
|
| 203 |
+
"eval_steps_per_second": 1.648,
|
| 204 |
+
"step": 434
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 1.52,
|
| 208 |
+
"learning_rate": 0.000125,
|
| 209 |
+
"loss": 0.8113,
|
| 210 |
+
"step": 465
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 1.52,
|
| 214 |
+
"eval_loss": 0.89990234375,
|
| 215 |
+
"eval_runtime": 159.5797,
|
| 216 |
+
"eval_samples_per_second": 26.344,
|
| 217 |
+
"eval_steps_per_second": 1.648,
|
| 218 |
+
"step": 465
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.62,
|
| 222 |
+
"learning_rate": 0.00012088746297956931,
|
| 223 |
+
"loss": 0.7849,
|
| 224 |
+
"step": 496
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.62,
|
| 228 |
+
"eval_loss": 0.89111328125,
|
| 229 |
+
"eval_runtime": 159.5898,
|
| 230 |
+
"eval_samples_per_second": 26.343,
|
| 231 |
+
"eval_steps_per_second": 1.648,
|
| 232 |
+
"step": 496
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 1.72,
|
| 236 |
+
"learning_rate": 0.00011715583722580123,
|
| 237 |
+
"loss": 0.7796,
|
| 238 |
+
"step": 527
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 1.72,
|
| 242 |
+
"eval_loss": 0.8828125,
|
| 243 |
+
"eval_runtime": 159.6567,
|
| 244 |
+
"eval_samples_per_second": 26.331,
|
| 245 |
+
"eval_steps_per_second": 1.647,
|
| 246 |
+
"step": 527
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 1.82,
|
| 250 |
+
"learning_rate": 0.00011374973339937476,
|
| 251 |
+
"loss": 0.7768,
|
| 252 |
+
"step": 558
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.82,
|
| 256 |
+
"eval_loss": 0.87548828125,
|
| 257 |
+
"eval_runtime": 159.5528,
|
| 258 |
+
"eval_samples_per_second": 26.349,
|
| 259 |
+
"eval_steps_per_second": 1.648,
|
| 260 |
+
"step": 558
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.92,
|
| 264 |
+
"learning_rate": 0.00011062441971717747,
|
| 265 |
+
"loss": 0.7784,
|
| 266 |
+
"step": 589
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.92,
|
| 270 |
+
"eval_loss": 0.8720703125,
|
| 271 |
+
"eval_runtime": 159.542,
|
| 272 |
+
"eval_samples_per_second": 26.35,
|
| 273 |
+
"eval_steps_per_second": 1.648,
|
| 274 |
+
"step": 589
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 2.0,
|
| 278 |
+
"step": 612,
|
| 279 |
+
"total_flos": 7.958718731863982e+17,
|
| 280 |
+
"train_loss": 0.9485874550015319,
|
| 281 |
+
"train_runtime": 6936.4877,
|
| 282 |
+
"train_samples_per_second": 5.654,
|
| 283 |
+
"train_steps_per_second": 0.088
|
| 284 |
+
}
|
| 285 |
+
],
|
| 286 |
+
"max_steps": 612,
|
| 287 |
+
"num_train_epochs": 2,
|
| 288 |
+
"total_flos": 7.958718731863982e+17,
|
| 289 |
+
"trial_name": null,
|
| 290 |
+
"trial_params": null
|
| 291 |
+
}
|