prxy5606 commited on
Commit
6dc2dff
·
verified ·
1 Parent(s): a86ff4f

Training in progress, step 200, checkpoint

Browse files
last-checkpoint/adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3af809830a2054e87872d814ab9ebe0451180179d4f82a820c4b8793c8279eb8
3
  size 289452128
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f53e569794f1f7fbf7a48b3b4e30bc76369d41af8e2278feaf3dfd6a0bcf1a9
3
  size 289452128
last-checkpoint/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f9856b28b070cd867215d4a88aa7b957e01dcd587c3a92dea7a91644a8d1a54e
3
  size 147359892
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ef26e1b502b05f78ae60b1df8a0d2685f34eaad67815497271a6307c1ffc7c9
3
  size 147359892
last-checkpoint/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:abd1a482e420668d1277f3ad8c2b2f1de528eff5a6f680bdfaad32301916698d
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c43e29062fd6982f73f8c831e10bcb009e4fd332231d37403d379348ee9eeaf
3
  size 14244
last-checkpoint/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d8ce05761f46e7cf72fb17a02e3a0ca15c9d25ce3babf590eeb40568923b8bac
3
  size 1064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d2d754412c61116546142914503e7369d0cc35d3c380a07e5218f595d76b6d96
3
  size 1064
last-checkpoint/trainer_state.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
- "best_metric": 0.7633453607559204,
3
- "best_model_checkpoint": "miner_id_24/checkpoint-150",
4
- "epoch": 0.15166835187057634,
5
  "eval_steps": 50,
6
- "global_step": 150,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
@@ -1089,6 +1089,364 @@
1089
  "eval_samples_per_second": 33.03,
1090
  "eval_steps_per_second": 8.267,
1091
  "step": 150
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1092
  }
1093
  ],
1094
  "logging_steps": 1,
@@ -1112,12 +1470,12 @@
1112
  "should_evaluate": false,
1113
  "should_log": false,
1114
  "should_save": true,
1115
- "should_training_stop": false
1116
  },
1117
  "attributes": {}
1118
  }
1119
  },
1120
- "total_flos": 5.021468359183565e+16,
1121
  "train_batch_size": 8,
1122
  "trial_name": null,
1123
  "trial_params": null
 
1
  {
2
+ "best_metric": 0.7592693567276001,
3
+ "best_model_checkpoint": "miner_id_24/checkpoint-200",
4
+ "epoch": 0.20222446916076844,
5
  "eval_steps": 50,
6
+ "global_step": 200,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
 
1089
  "eval_samples_per_second": 33.03,
1090
  "eval_steps_per_second": 8.267,
1091
  "step": 150
1092
+ },
1093
+ {
1094
+ "epoch": 0.15267947421638017,
1095
+ "grad_norm": 0.1272272765636444,
1096
+ "learning_rate": 1.553232954407171e-05,
1097
+ "loss": 0.6778,
1098
+ "step": 151
1099
+ },
1100
+ {
1101
+ "epoch": 0.15369059656218403,
1102
+ "grad_norm": 0.1399427205324173,
1103
+ "learning_rate": 1.4938160786375572e-05,
1104
+ "loss": 0.7174,
1105
+ "step": 152
1106
+ },
1107
+ {
1108
+ "epoch": 0.15470171890798787,
1109
+ "grad_norm": 0.16902732849121094,
1110
+ "learning_rate": 1.435357758543015e-05,
1111
+ "loss": 0.7299,
1112
+ "step": 153
1113
+ },
1114
+ {
1115
+ "epoch": 0.1557128412537917,
1116
+ "grad_norm": 0.1675402969121933,
1117
+ "learning_rate": 1.3778739760445552e-05,
1118
+ "loss": 0.7037,
1119
+ "step": 154
1120
+ },
1121
+ {
1122
+ "epoch": 0.15672396359959556,
1123
+ "grad_norm": 0.21070344746112823,
1124
+ "learning_rate": 1.3213804466343421e-05,
1125
+ "loss": 0.7654,
1126
+ "step": 155
1127
+ },
1128
+ {
1129
+ "epoch": 0.1577350859453994,
1130
+ "grad_norm": 0.21207165718078613,
1131
+ "learning_rate": 1.2658926150792322e-05,
1132
+ "loss": 0.8009,
1133
+ "step": 156
1134
+ },
1135
+ {
1136
+ "epoch": 0.15874620829120323,
1137
+ "grad_norm": 0.2189692109823227,
1138
+ "learning_rate": 1.2114256511983274e-05,
1139
+ "loss": 0.7592,
1140
+ "step": 157
1141
+ },
1142
+ {
1143
+ "epoch": 0.1597573306370071,
1144
+ "grad_norm": 0.2378927320241928,
1145
+ "learning_rate": 1.157994445715706e-05,
1146
+ "loss": 0.7884,
1147
+ "step": 158
1148
+ },
1149
+ {
1150
+ "epoch": 0.16076845298281092,
1151
+ "grad_norm": 0.24428603053092957,
1152
+ "learning_rate": 1.1056136061894384e-05,
1153
+ "loss": 0.814,
1154
+ "step": 159
1155
+ },
1156
+ {
1157
+ "epoch": 0.16177957532861476,
1158
+ "grad_norm": 0.23859427869319916,
1159
+ "learning_rate": 1.0542974530180327e-05,
1160
+ "loss": 0.7384,
1161
+ "step": 160
1162
+ },
1163
+ {
1164
+ "epoch": 0.16279069767441862,
1165
+ "grad_norm": 0.257879376411438,
1166
+ "learning_rate": 1.0040600155253765e-05,
1167
+ "loss": 0.7736,
1168
+ "step": 161
1169
+ },
1170
+ {
1171
+ "epoch": 0.16380182002022245,
1172
+ "grad_norm": 0.2455569952726364,
1173
+ "learning_rate": 9.549150281252633e-06,
1174
+ "loss": 0.7593,
1175
+ "step": 162
1176
+ },
1177
+ {
1178
+ "epoch": 0.16481294236602628,
1179
+ "grad_norm": 0.28459832072257996,
1180
+ "learning_rate": 9.068759265665384e-06,
1181
+ "loss": 0.8493,
1182
+ "step": 163
1183
+ },
1184
+ {
1185
+ "epoch": 0.16582406471183014,
1186
+ "grad_norm": 0.2694055140018463,
1187
+ "learning_rate": 8.599558442598998e-06,
1188
+ "loss": 0.8126,
1189
+ "step": 164
1190
+ },
1191
+ {
1192
+ "epoch": 0.16683518705763398,
1193
+ "grad_norm": 0.2462318241596222,
1194
+ "learning_rate": 8.141676086873572e-06,
1195
+ "loss": 0.7374,
1196
+ "step": 165
1197
+ },
1198
+ {
1199
+ "epoch": 0.1678463094034378,
1200
+ "grad_norm": 0.2713393568992615,
1201
+ "learning_rate": 7.695237378953223e-06,
1202
+ "loss": 0.7638,
1203
+ "step": 166
1204
+ },
1205
+ {
1206
+ "epoch": 0.16885743174924167,
1207
+ "grad_norm": 0.25184810161590576,
1208
+ "learning_rate": 7.260364370723044e-06,
1209
+ "loss": 0.7354,
1210
+ "step": 167
1211
+ },
1212
+ {
1213
+ "epoch": 0.1698685540950455,
1214
+ "grad_norm": 0.2712859809398651,
1215
+ "learning_rate": 6.837175952121306e-06,
1216
+ "loss": 0.7911,
1217
+ "step": 168
1218
+ },
1219
+ {
1220
+ "epoch": 0.17087967644084934,
1221
+ "grad_norm": 0.2823898494243622,
1222
+ "learning_rate": 6.425787818636131e-06,
1223
+ "loss": 0.7985,
1224
+ "step": 169
1225
+ },
1226
+ {
1227
+ "epoch": 0.1718907987866532,
1228
+ "grad_norm": 0.25658711791038513,
1229
+ "learning_rate": 6.026312439675552e-06,
1230
+ "loss": 0.7691,
1231
+ "step": 170
1232
+ },
1233
+ {
1234
+ "epoch": 0.17290192113245703,
1235
+ "grad_norm": 0.2877839207649231,
1236
+ "learning_rate": 5.6388590278194096e-06,
1237
+ "loss": 0.761,
1238
+ "step": 171
1239
+ },
1240
+ {
1241
+ "epoch": 0.17391304347826086,
1242
+ "grad_norm": 0.28454214334487915,
1243
+ "learning_rate": 5.263533508961827e-06,
1244
+ "loss": 0.7568,
1245
+ "step": 172
1246
+ },
1247
+ {
1248
+ "epoch": 0.17492416582406473,
1249
+ "grad_norm": 0.26517605781555176,
1250
+ "learning_rate": 4.900438493352055e-06,
1251
+ "loss": 0.6914,
1252
+ "step": 173
1253
+ },
1254
+ {
1255
+ "epoch": 0.17593528816986856,
1256
+ "grad_norm": 0.2830790579319,
1257
+ "learning_rate": 4.549673247541875e-06,
1258
+ "loss": 0.7735,
1259
+ "step": 174
1260
+ },
1261
+ {
1262
+ "epoch": 0.1769464105156724,
1263
+ "grad_norm": 0.306985467672348,
1264
+ "learning_rate": 4.2113336672471245e-06,
1265
+ "loss": 0.7257,
1266
+ "step": 175
1267
+ },
1268
+ {
1269
+ "epoch": 0.17795753286147623,
1270
+ "grad_norm": 0.29978933930397034,
1271
+ "learning_rate": 3.885512251130763e-06,
1272
+ "loss": 0.7711,
1273
+ "step": 176
1274
+ },
1275
+ {
1276
+ "epoch": 0.1789686552072801,
1277
+ "grad_norm": 0.289646178483963,
1278
+ "learning_rate": 3.5722980755146517e-06,
1279
+ "loss": 0.8057,
1280
+ "step": 177
1281
+ },
1282
+ {
1283
+ "epoch": 0.17997977755308392,
1284
+ "grad_norm": 0.2966819405555725,
1285
+ "learning_rate": 3.271776770026963e-06,
1286
+ "loss": 0.7282,
1287
+ "step": 178
1288
+ },
1289
+ {
1290
+ "epoch": 0.18099089989888775,
1291
+ "grad_norm": 0.31478750705718994,
1292
+ "learning_rate": 2.9840304941919415e-06,
1293
+ "loss": 0.7856,
1294
+ "step": 179
1295
+ },
1296
+ {
1297
+ "epoch": 0.1820020222446916,
1298
+ "grad_norm": 0.29623422026634216,
1299
+ "learning_rate": 2.7091379149682685e-06,
1300
+ "loss": 0.6985,
1301
+ "step": 180
1302
+ },
1303
+ {
1304
+ "epoch": 0.18301314459049545,
1305
+ "grad_norm": 0.32452911138534546,
1306
+ "learning_rate": 2.4471741852423237e-06,
1307
+ "loss": 0.7571,
1308
+ "step": 181
1309
+ },
1310
+ {
1311
+ "epoch": 0.18402426693629928,
1312
+ "grad_norm": 0.2937887907028198,
1313
+ "learning_rate": 2.1982109232821178e-06,
1314
+ "loss": 0.6749,
1315
+ "step": 182
1316
+ },
1317
+ {
1318
+ "epoch": 0.18503538928210314,
1319
+ "grad_norm": 0.322902113199234,
1320
+ "learning_rate": 1.962316193157593e-06,
1321
+ "loss": 0.756,
1322
+ "step": 183
1323
+ },
1324
+ {
1325
+ "epoch": 0.18604651162790697,
1326
+ "grad_norm": 0.3696511685848236,
1327
+ "learning_rate": 1.7395544861325718e-06,
1328
+ "loss": 0.849,
1329
+ "step": 184
1330
+ },
1331
+ {
1332
+ "epoch": 0.1870576339737108,
1333
+ "grad_norm": 0.33034321665763855,
1334
+ "learning_rate": 1.5299867030334814e-06,
1335
+ "loss": 0.7859,
1336
+ "step": 185
1337
+ },
1338
+ {
1339
+ "epoch": 0.18806875631951467,
1340
+ "grad_norm": 0.34689730405807495,
1341
+ "learning_rate": 1.333670137599713e-06,
1342
+ "loss": 0.8283,
1343
+ "step": 186
1344
+ },
1345
+ {
1346
+ "epoch": 0.1890798786653185,
1347
+ "grad_norm": 0.3476114571094513,
1348
+ "learning_rate": 1.1506584608200367e-06,
1349
+ "loss": 0.7518,
1350
+ "step": 187
1351
+ },
1352
+ {
1353
+ "epoch": 0.19009100101112233,
1354
+ "grad_norm": 0.37273553013801575,
1355
+ "learning_rate": 9.810017062595322e-07,
1356
+ "loss": 0.816,
1357
+ "step": 188
1358
+ },
1359
+ {
1360
+ "epoch": 0.1911021233569262,
1361
+ "grad_norm": 0.35195815563201904,
1362
+ "learning_rate": 8.247462563808817e-07,
1363
+ "loss": 0.7478,
1364
+ "step": 189
1365
+ },
1366
+ {
1367
+ "epoch": 0.19211324570273003,
1368
+ "grad_norm": 0.3585607707500458,
1369
+ "learning_rate": 6.819348298638839e-07,
1370
+ "loss": 0.7828,
1371
+ "step": 190
1372
+ },
1373
+ {
1374
+ "epoch": 0.19312436804853386,
1375
+ "grad_norm": 0.3714843988418579,
1376
+ "learning_rate": 5.526064699265753e-07,
1377
+ "loss": 0.7635,
1378
+ "step": 191
1379
+ },
1380
+ {
1381
+ "epoch": 0.19413549039433772,
1382
+ "grad_norm": 0.37644708156585693,
1383
+ "learning_rate": 4.367965336512403e-07,
1384
+ "loss": 0.8012,
1385
+ "step": 192
1386
+ },
1387
+ {
1388
+ "epoch": 0.19514661274014156,
1389
+ "grad_norm": 0.4035707712173462,
1390
+ "learning_rate": 3.3453668231809286e-07,
1391
+ "loss": 0.8388,
1392
+ "step": 193
1393
+ },
1394
+ {
1395
+ "epoch": 0.1961577350859454,
1396
+ "grad_norm": 0.3666282892227173,
1397
+ "learning_rate": 2.458548727494292e-07,
1398
+ "loss": 0.6399,
1399
+ "step": 194
1400
+ },
1401
+ {
1402
+ "epoch": 0.19716885743174925,
1403
+ "grad_norm": 0.37958642840385437,
1404
+ "learning_rate": 1.7077534966650766e-07,
1405
+ "loss": 0.6872,
1406
+ "step": 195
1407
+ },
1408
+ {
1409
+ "epoch": 0.19817997977755308,
1410
+ "grad_norm": 0.4451029300689697,
1411
+ "learning_rate": 1.0931863906127327e-07,
1412
+ "loss": 0.8135,
1413
+ "step": 196
1414
+ },
1415
+ {
1416
+ "epoch": 0.19919110212335692,
1417
+ "grad_norm": 0.492899090051651,
1418
+ "learning_rate": 6.150154258476315e-08,
1419
+ "loss": 0.8836,
1420
+ "step": 197
1421
+ },
1422
+ {
1423
+ "epoch": 0.20020222446916078,
1424
+ "grad_norm": 0.47472137212753296,
1425
+ "learning_rate": 2.7337132953697554e-08,
1426
+ "loss": 0.72,
1427
+ "step": 198
1428
+ },
1429
+ {
1430
+ "epoch": 0.2012133468149646,
1431
+ "grad_norm": 0.46337440609931946,
1432
+ "learning_rate": 6.834750376549792e-09,
1433
+ "loss": 0.7419,
1434
+ "step": 199
1435
+ },
1436
+ {
1437
+ "epoch": 0.20222446916076844,
1438
+ "grad_norm": 0.6667613387107849,
1439
+ "learning_rate": 0.0,
1440
+ "loss": 0.7994,
1441
+ "step": 200
1442
+ },
1443
+ {
1444
+ "epoch": 0.20222446916076844,
1445
+ "eval_loss": 0.7592693567276001,
1446
+ "eval_runtime": 50.4908,
1447
+ "eval_samples_per_second": 32.996,
1448
+ "eval_steps_per_second": 8.259,
1449
+ "step": 200
1450
  }
1451
  ],
1452
  "logging_steps": 1,
 
1470
  "should_evaluate": false,
1471
  "should_log": false,
1472
  "should_save": true,
1473
+ "should_training_stop": true
1474
  },
1475
  "attributes": {}
1476
  }
1477
  },
1478
+ "total_flos": 6.70907881268183e+16,
1479
  "train_batch_size": 8,
1480
  "trial_name": null,
1481
  "trial_params": null