prxy5607 commited on
Commit
0b55a94
·
verified ·
1 Parent(s): 5b342e9

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 sha256:34b59f344b369bedb22556e759d6b47fc1ded6905e7bdd5ce64eb22318abec10
3
  size 100690184
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22e0fdbc3e2cc7be4e0af52bf14803d5dbf4f3e1d193bfe9ae1fd5f707a09bca
3
  size 100690184
last-checkpoint/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6b9bd760dcf0d6307314038b65dd65114f6bb15250bf1e99c5a759bd128186f0
3
  size 51344890
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db29a3423acb6bfb07f15a14d50007656935bc1901c552e3d1486f248482564d
3
  size 51344890
last-checkpoint/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6d906c11372bf6b66f2ff5751d2d1bd68080b6989750efb059c1b7abe720285f
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7849462d330266c20154d67e1c5ca42b18138b5054d0f57ddaace1e37853abba
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": 3.6782333850860596,
3
- "best_model_checkpoint": "miner_id_24/checkpoint-150",
4
- "epoch": 2.2222222222222223,
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": 85.196,
1090
  "eval_steps_per_second": 21.673,
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": 1.1148884881440768e+16,
1121
  "train_batch_size": 8,
1122
  "trial_name": null,
1123
  "trial_params": null
 
1
  {
2
+ "best_metric": 3.651360034942627,
3
+ "best_model_checkpoint": "miner_id_24/checkpoint-200",
4
+ "epoch": 2.962962962962963,
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": 85.196,
1090
  "eval_steps_per_second": 21.673,
1091
  "step": 150
1092
+ },
1093
+ {
1094
+ "epoch": 2.237037037037037,
1095
+ "grad_norm": 203.73736572265625,
1096
+ "learning_rate": 1.553232954407171e-05,
1097
+ "loss": 14.9,
1098
+ "step": 151
1099
+ },
1100
+ {
1101
+ "epoch": 2.251851851851852,
1102
+ "grad_norm": 153.0937957763672,
1103
+ "learning_rate": 1.4938160786375572e-05,
1104
+ "loss": 14.0142,
1105
+ "step": 152
1106
+ },
1107
+ {
1108
+ "epoch": 2.2666666666666666,
1109
+ "grad_norm": 117.68968963623047,
1110
+ "learning_rate": 1.435357758543015e-05,
1111
+ "loss": 14.3636,
1112
+ "step": 153
1113
+ },
1114
+ {
1115
+ "epoch": 2.2814814814814817,
1116
+ "grad_norm": 137.20506286621094,
1117
+ "learning_rate": 1.3778739760445552e-05,
1118
+ "loss": 13.7021,
1119
+ "step": 154
1120
+ },
1121
+ {
1122
+ "epoch": 2.2962962962962963,
1123
+ "grad_norm": 161.4315185546875,
1124
+ "learning_rate": 1.3213804466343421e-05,
1125
+ "loss": 14.8398,
1126
+ "step": 155
1127
+ },
1128
+ {
1129
+ "epoch": 2.311111111111111,
1130
+ "grad_norm": 162.87452697753906,
1131
+ "learning_rate": 1.2658926150792322e-05,
1132
+ "loss": 14.2438,
1133
+ "step": 156
1134
+ },
1135
+ {
1136
+ "epoch": 2.325925925925926,
1137
+ "grad_norm": 255.86105346679688,
1138
+ "learning_rate": 1.2114256511983274e-05,
1139
+ "loss": 14.2575,
1140
+ "step": 157
1141
+ },
1142
+ {
1143
+ "epoch": 2.3407407407407406,
1144
+ "grad_norm": 171.7278594970703,
1145
+ "learning_rate": 1.157994445715706e-05,
1146
+ "loss": 14.3891,
1147
+ "step": 158
1148
+ },
1149
+ {
1150
+ "epoch": 2.3555555555555556,
1151
+ "grad_norm": 138.58694458007812,
1152
+ "learning_rate": 1.1056136061894384e-05,
1153
+ "loss": 14.4252,
1154
+ "step": 159
1155
+ },
1156
+ {
1157
+ "epoch": 2.3703703703703702,
1158
+ "grad_norm": 131.3074188232422,
1159
+ "learning_rate": 1.0542974530180327e-05,
1160
+ "loss": 13.6398,
1161
+ "step": 160
1162
+ },
1163
+ {
1164
+ "epoch": 2.3851851851851853,
1165
+ "grad_norm": 132.39199829101562,
1166
+ "learning_rate": 1.0040600155253765e-05,
1167
+ "loss": 13.8621,
1168
+ "step": 161
1169
+ },
1170
+ {
1171
+ "epoch": 2.4,
1172
+ "grad_norm": 208.60899353027344,
1173
+ "learning_rate": 9.549150281252633e-06,
1174
+ "loss": 13.7883,
1175
+ "step": 162
1176
+ },
1177
+ {
1178
+ "epoch": 2.414814814814815,
1179
+ "grad_norm": 151.96185302734375,
1180
+ "learning_rate": 9.068759265665384e-06,
1181
+ "loss": 14.6574,
1182
+ "step": 163
1183
+ },
1184
+ {
1185
+ "epoch": 2.4296296296296296,
1186
+ "grad_norm": 149.67416381835938,
1187
+ "learning_rate": 8.599558442598998e-06,
1188
+ "loss": 13.8391,
1189
+ "step": 164
1190
+ },
1191
+ {
1192
+ "epoch": 2.4444444444444446,
1193
+ "grad_norm": 179.20135498046875,
1194
+ "learning_rate": 8.141676086873572e-06,
1195
+ "loss": 14.5998,
1196
+ "step": 165
1197
+ },
1198
+ {
1199
+ "epoch": 2.4592592592592593,
1200
+ "grad_norm": 240.76341247558594,
1201
+ "learning_rate": 7.695237378953223e-06,
1202
+ "loss": 14.8283,
1203
+ "step": 166
1204
+ },
1205
+ {
1206
+ "epoch": 2.474074074074074,
1207
+ "grad_norm": 177.29092407226562,
1208
+ "learning_rate": 7.260364370723044e-06,
1209
+ "loss": 13.991,
1210
+ "step": 167
1211
+ },
1212
+ {
1213
+ "epoch": 2.488888888888889,
1214
+ "grad_norm": 112.82994842529297,
1215
+ "learning_rate": 6.837175952121306e-06,
1216
+ "loss": 13.6775,
1217
+ "step": 168
1218
+ },
1219
+ {
1220
+ "epoch": 2.5037037037037035,
1221
+ "grad_norm": 134.5693817138672,
1222
+ "learning_rate": 6.425787818636131e-06,
1223
+ "loss": 14.0775,
1224
+ "step": 169
1225
+ },
1226
+ {
1227
+ "epoch": 2.5185185185185186,
1228
+ "grad_norm": 116.91983795166016,
1229
+ "learning_rate": 6.026312439675552e-06,
1230
+ "loss": 13.7677,
1231
+ "step": 170
1232
+ },
1233
+ {
1234
+ "epoch": 2.533333333333333,
1235
+ "grad_norm": 164.41635131835938,
1236
+ "learning_rate": 5.6388590278194096e-06,
1237
+ "loss": 14.3629,
1238
+ "step": 171
1239
+ },
1240
+ {
1241
+ "epoch": 2.5481481481481483,
1242
+ "grad_norm": 159.20077514648438,
1243
+ "learning_rate": 5.263533508961827e-06,
1244
+ "loss": 13.9835,
1245
+ "step": 172
1246
+ },
1247
+ {
1248
+ "epoch": 2.562962962962963,
1249
+ "grad_norm": 119.64713287353516,
1250
+ "learning_rate": 4.900438493352055e-06,
1251
+ "loss": 14.1561,
1252
+ "step": 173
1253
+ },
1254
+ {
1255
+ "epoch": 2.5777777777777775,
1256
+ "grad_norm": 166.39501953125,
1257
+ "learning_rate": 4.549673247541875e-06,
1258
+ "loss": 13.9771,
1259
+ "step": 174
1260
+ },
1261
+ {
1262
+ "epoch": 2.5925925925925926,
1263
+ "grad_norm": 158.79327392578125,
1264
+ "learning_rate": 4.2113336672471245e-06,
1265
+ "loss": 13.909,
1266
+ "step": 175
1267
+ },
1268
+ {
1269
+ "epoch": 2.6074074074074076,
1270
+ "grad_norm": 170.07273864746094,
1271
+ "learning_rate": 3.885512251130763e-06,
1272
+ "loss": 14.2716,
1273
+ "step": 176
1274
+ },
1275
+ {
1276
+ "epoch": 2.6222222222222222,
1277
+ "grad_norm": 175.40963745117188,
1278
+ "learning_rate": 3.5722980755146517e-06,
1279
+ "loss": 13.8635,
1280
+ "step": 177
1281
+ },
1282
+ {
1283
+ "epoch": 2.637037037037037,
1284
+ "grad_norm": 165.8638458251953,
1285
+ "learning_rate": 3.271776770026963e-06,
1286
+ "loss": 15.226,
1287
+ "step": 178
1288
+ },
1289
+ {
1290
+ "epoch": 2.651851851851852,
1291
+ "grad_norm": 156.8959197998047,
1292
+ "learning_rate": 2.9840304941919415e-06,
1293
+ "loss": 13.9328,
1294
+ "step": 179
1295
+ },
1296
+ {
1297
+ "epoch": 2.6666666666666665,
1298
+ "grad_norm": 137.80484008789062,
1299
+ "learning_rate": 2.7091379149682685e-06,
1300
+ "loss": 13.5431,
1301
+ "step": 180
1302
+ },
1303
+ {
1304
+ "epoch": 2.6814814814814816,
1305
+ "grad_norm": 157.63360595703125,
1306
+ "learning_rate": 2.4471741852423237e-06,
1307
+ "loss": 14.3916,
1308
+ "step": 181
1309
+ },
1310
+ {
1311
+ "epoch": 2.696296296296296,
1312
+ "grad_norm": 170.91567993164062,
1313
+ "learning_rate": 2.1982109232821178e-06,
1314
+ "loss": 13.752,
1315
+ "step": 182
1316
+ },
1317
+ {
1318
+ "epoch": 2.7111111111111112,
1319
+ "grad_norm": 174.46258544921875,
1320
+ "learning_rate": 1.962316193157593e-06,
1321
+ "loss": 13.6756,
1322
+ "step": 183
1323
+ },
1324
+ {
1325
+ "epoch": 2.725925925925926,
1326
+ "grad_norm": 123.57269287109375,
1327
+ "learning_rate": 1.7395544861325718e-06,
1328
+ "loss": 13.3758,
1329
+ "step": 184
1330
+ },
1331
+ {
1332
+ "epoch": 2.7407407407407405,
1333
+ "grad_norm": 113.82854461669922,
1334
+ "learning_rate": 1.5299867030334814e-06,
1335
+ "loss": 14.2526,
1336
+ "step": 185
1337
+ },
1338
+ {
1339
+ "epoch": 2.7555555555555555,
1340
+ "grad_norm": 147.1977996826172,
1341
+ "learning_rate": 1.333670137599713e-06,
1342
+ "loss": 14.2643,
1343
+ "step": 186
1344
+ },
1345
+ {
1346
+ "epoch": 2.7703703703703706,
1347
+ "grad_norm": 134.13990783691406,
1348
+ "learning_rate": 1.1506584608200367e-06,
1349
+ "loss": 13.8424,
1350
+ "step": 187
1351
+ },
1352
+ {
1353
+ "epoch": 2.785185185185185,
1354
+ "grad_norm": 136.7030029296875,
1355
+ "learning_rate": 9.810017062595322e-07,
1356
+ "loss": 14.0423,
1357
+ "step": 188
1358
+ },
1359
+ {
1360
+ "epoch": 2.8,
1361
+ "grad_norm": 158.4160614013672,
1362
+ "learning_rate": 8.247462563808817e-07,
1363
+ "loss": 13.6131,
1364
+ "step": 189
1365
+ },
1366
+ {
1367
+ "epoch": 2.814814814814815,
1368
+ "grad_norm": 151.19692993164062,
1369
+ "learning_rate": 6.819348298638839e-07,
1370
+ "loss": 14.2082,
1371
+ "step": 190
1372
+ },
1373
+ {
1374
+ "epoch": 2.8296296296296295,
1375
+ "grad_norm": 201.72494506835938,
1376
+ "learning_rate": 5.526064699265753e-07,
1377
+ "loss": 13.4682,
1378
+ "step": 191
1379
+ },
1380
+ {
1381
+ "epoch": 2.8444444444444446,
1382
+ "grad_norm": 144.17710876464844,
1383
+ "learning_rate": 4.367965336512403e-07,
1384
+ "loss": 14.1444,
1385
+ "step": 192
1386
+ },
1387
+ {
1388
+ "epoch": 2.859259259259259,
1389
+ "grad_norm": 149.2476806640625,
1390
+ "learning_rate": 3.3453668231809286e-07,
1391
+ "loss": 14.1861,
1392
+ "step": 193
1393
+ },
1394
+ {
1395
+ "epoch": 2.8740740740740742,
1396
+ "grad_norm": 154.860107421875,
1397
+ "learning_rate": 2.458548727494292e-07,
1398
+ "loss": 14.5229,
1399
+ "step": 194
1400
+ },
1401
+ {
1402
+ "epoch": 2.888888888888889,
1403
+ "grad_norm": 136.7914276123047,
1404
+ "learning_rate": 1.7077534966650766e-07,
1405
+ "loss": 14.2759,
1406
+ "step": 195
1407
+ },
1408
+ {
1409
+ "epoch": 2.9037037037037035,
1410
+ "grad_norm": 183.71226501464844,
1411
+ "learning_rate": 1.0931863906127327e-07,
1412
+ "loss": 13.7139,
1413
+ "step": 196
1414
+ },
1415
+ {
1416
+ "epoch": 2.9185185185185185,
1417
+ "grad_norm": 169.7905731201172,
1418
+ "learning_rate": 6.150154258476315e-08,
1419
+ "loss": 13.6161,
1420
+ "step": 197
1421
+ },
1422
+ {
1423
+ "epoch": 2.9333333333333336,
1424
+ "grad_norm": 161.94277954101562,
1425
+ "learning_rate": 2.7337132953697554e-08,
1426
+ "loss": 14.7131,
1427
+ "step": 198
1428
+ },
1429
+ {
1430
+ "epoch": 2.948148148148148,
1431
+ "grad_norm": 203.21755981445312,
1432
+ "learning_rate": 6.834750376549792e-09,
1433
+ "loss": 13.5579,
1434
+ "step": 199
1435
+ },
1436
+ {
1437
+ "epoch": 2.962962962962963,
1438
+ "grad_norm": 139.2266387939453,
1439
+ "learning_rate": 0.0,
1440
+ "loss": 14.1591,
1441
+ "step": 200
1442
+ },
1443
+ {
1444
+ "epoch": 2.962962962962963,
1445
+ "eval_loss": 3.651360034942627,
1446
+ "eval_runtime": 1.3259,
1447
+ "eval_samples_per_second": 85.981,
1448
+ "eval_steps_per_second": 21.872,
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": 1.4874493864378368e+16,
1479
  "train_batch_size": 8,
1480
  "trial_name": null,
1481
  "trial_params": null