YeRyeongLee
commited on
Commit
•
bee5fad
1
Parent(s):
f8514c1
End of training
Browse files- all_results.json +3 -3
- train_results.json +3 -3
- train_results.txt +3 -3
- trainer_state.json +115 -5
all_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 10.0,
|
3 |
"train_loss": 0.6484309974533823,
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 100102,
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 10.0,
|
3 |
"train_loss": 0.6484309974533823,
|
4 |
+
"train_runtime": 11384.9561,
|
5 |
"train_samples": 100102,
|
6 |
+
"train_samples_per_second": 87.925,
|
7 |
+
"train_steps_per_second": 0.343
|
8 |
}
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 10.0,
|
3 |
"train_loss": 0.6484309974533823,
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 100102,
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 10.0,
|
3 |
"train_loss": 0.6484309974533823,
|
4 |
+
"train_runtime": 11384.9561,
|
5 |
"train_samples": 100102,
|
6 |
+
"train_samples_per_second": 87.925,
|
7 |
+
"train_steps_per_second": 0.343
|
8 |
}
|
train_results.txt
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
epoch = 10.0
|
2 |
train_loss = 0.6484309974533823
|
3 |
-
train_runtime =
|
4 |
train_samples = 100102
|
5 |
-
train_samples_per_second =
|
6 |
-
train_steps_per_second = 0.
|
|
|
1 |
epoch = 10.0
|
2 |
train_loss = 0.6484309974533823
|
3 |
+
train_runtime = 11384.9561
|
4 |
train_samples = 100102
|
5 |
+
train_samples_per_second = 87.925
|
6 |
+
train_steps_per_second = 0.343
|
trainer_state.json
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
{
|
2 |
-
"best_metric":
|
3 |
-
"best_model_checkpoint":
|
4 |
"epoch": 9.999680409076383,
|
5 |
"global_step": 3910,
|
6 |
"is_hyper_param_search": false,
|
@@ -241,6 +241,17 @@
|
|
241 |
"loss": 1.0543,
|
242 |
"step": 390
|
243 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
244 |
{
|
245 |
"epoch": 1.02,
|
246 |
"learning_rate": 1.9948849104859337e-05,
|
@@ -475,6 +486,17 @@
|
|
475 |
"loss": 0.8573,
|
476 |
"step": 780
|
477 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
478 |
{
|
479 |
"epoch": 2.02,
|
480 |
"learning_rate": 1.773231031543052e-05,
|
@@ -709,6 +731,17 @@
|
|
709 |
"loss": 0.7217,
|
710 |
"step": 1170
|
711 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
712 |
{
|
713 |
"epoch": 3.02,
|
714 |
"learning_rate": 1.5515771526001706e-05,
|
@@ -943,6 +976,17 @@
|
|
943 |
"loss": 0.6305,
|
944 |
"step": 1560
|
945 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
946 |
{
|
947 |
"epoch": 4.02,
|
948 |
"learning_rate": 1.329923273657289e-05,
|
@@ -1177,6 +1221,17 @@
|
|
1177 |
"loss": 0.5294,
|
1178 |
"step": 1950
|
1179 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1180 |
{
|
1181 |
"epoch": 5.01,
|
1182 |
"learning_rate": 1.1082693947144077e-05,
|
@@ -1411,6 +1466,17 @@
|
|
1411 |
"loss": 0.4689,
|
1412 |
"step": 2340
|
1413 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1414 |
{
|
1415 |
"epoch": 6.01,
|
1416 |
"learning_rate": 8.86615515771526e-06,
|
@@ -1645,6 +1711,17 @@
|
|
1645 |
"loss": 0.433,
|
1646 |
"step": 2730
|
1647 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1648 |
{
|
1649 |
"epoch": 7.01,
|
1650 |
"learning_rate": 6.649616368286445e-06,
|
@@ -1879,6 +1956,17 @@
|
|
1879 |
"loss": 0.419,
|
1880 |
"step": 3120
|
1881 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1882 |
{
|
1883 |
"epoch": 8.01,
|
1884 |
"learning_rate": 4.43307757885763e-06,
|
@@ -2113,6 +2201,17 @@
|
|
2113 |
"loss": 0.3766,
|
2114 |
"step": 3510
|
2115 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2116 |
{
|
2117 |
"epoch": 9.0,
|
2118 |
"learning_rate": 2.216538789428815e-06,
|
@@ -2353,14 +2452,25 @@
|
|
2353 |
"loss": 0.3395,
|
2354 |
"step": 3910
|
2355 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2356 |
{
|
2357 |
"epoch": 10.0,
|
2358 |
"step": 3910,
|
2359 |
"total_flos": 1.0186138340339622e+17,
|
2360 |
"train_loss": 0.6484309974533823,
|
2361 |
-
"train_runtime":
|
2362 |
-
"train_samples_per_second":
|
2363 |
-
"train_steps_per_second": 0.
|
2364 |
}
|
2365 |
],
|
2366 |
"max_steps": 3910,
|
|
|
1 |
{
|
2 |
+
"best_metric": 0.7961117029190063,
|
3 |
+
"best_model_checkpoint": "./models/emot_and_voice/checkpoint-1173",
|
4 |
"epoch": 9.999680409076383,
|
5 |
"global_step": 3910,
|
6 |
"is_hyper_param_search": false,
|
|
|
241 |
"loss": 1.0543,
|
242 |
"step": 390
|
243 |
},
|
244 |
+
{
|
245 |
+
"epoch": 1.0,
|
246 |
+
"eval_accuracy": 0.6530612244897959,
|
247 |
+
"eval_auprc": 49.690567272400955,
|
248 |
+
"eval_loss": 0.9922736287117004,
|
249 |
+
"eval_micro f1 score": 65.3061224489796,
|
250 |
+
"eval_runtime": 41.8031,
|
251 |
+
"eval_samples_per_second": 266.081,
|
252 |
+
"eval_steps_per_second": 8.325,
|
253 |
+
"step": 391
|
254 |
+
},
|
255 |
{
|
256 |
"epoch": 1.02,
|
257 |
"learning_rate": 1.9948849104859337e-05,
|
|
|
486 |
"loss": 0.8573,
|
487 |
"step": 780
|
488 |
},
|
489 |
+
{
|
490 |
+
"epoch": 2.0,
|
491 |
+
"eval_accuracy": 0.699901105816776,
|
492 |
+
"eval_auprc": 64.40713435079698,
|
493 |
+
"eval_loss": 0.822926938533783,
|
494 |
+
"eval_micro f1 score": 69.9901105816776,
|
495 |
+
"eval_runtime": 41.7698,
|
496 |
+
"eval_samples_per_second": 266.293,
|
497 |
+
"eval_steps_per_second": 8.331,
|
498 |
+
"step": 782
|
499 |
+
},
|
500 |
{
|
501 |
"epoch": 2.02,
|
502 |
"learning_rate": 1.773231031543052e-05,
|
|
|
731 |
"loss": 0.7217,
|
732 |
"step": 1170
|
733 |
},
|
734 |
+
{
|
735 |
+
"epoch": 3.0,
|
736 |
+
"eval_accuracy": 0.7105996583655488,
|
737 |
+
"eval_auprc": 67.46402201308636,
|
738 |
+
"eval_loss": 0.7961117029190063,
|
739 |
+
"eval_micro f1 score": 71.05996583655488,
|
740 |
+
"eval_runtime": 41.7385,
|
741 |
+
"eval_samples_per_second": 266.493,
|
742 |
+
"eval_steps_per_second": 8.338,
|
743 |
+
"step": 1173
|
744 |
+
},
|
745 |
{
|
746 |
"epoch": 3.02,
|
747 |
"learning_rate": 1.5515771526001706e-05,
|
|
|
976 |
"loss": 0.6305,
|
977 |
"step": 1560
|
978 |
},
|
979 |
+
{
|
980 |
+
"epoch": 4.0,
|
981 |
+
"eval_accuracy": 0.7112289849860649,
|
982 |
+
"eval_auprc": 68.51906233767703,
|
983 |
+
"eval_loss": 0.8162872791290283,
|
984 |
+
"eval_micro f1 score": 71.12289849860647,
|
985 |
+
"eval_runtime": 41.7027,
|
986 |
+
"eval_samples_per_second": 266.721,
|
987 |
+
"eval_steps_per_second": 8.345,
|
988 |
+
"step": 1564
|
989 |
+
},
|
990 |
{
|
991 |
"epoch": 4.02,
|
992 |
"learning_rate": 1.329923273657289e-05,
|
|
|
1221 |
"loss": 0.5294,
|
1222 |
"step": 1950
|
1223 |
},
|
1224 |
+
{
|
1225 |
+
"epoch": 5.0,
|
1226 |
+
"eval_accuracy": 0.7101501393508945,
|
1227 |
+
"eval_auprc": 68.73344594150042,
|
1228 |
+
"eval_loss": 0.8205118775367737,
|
1229 |
+
"eval_micro f1 score": 71.01501393508946,
|
1230 |
+
"eval_runtime": 41.8651,
|
1231 |
+
"eval_samples_per_second": 265.687,
|
1232 |
+
"eval_steps_per_second": 8.312,
|
1233 |
+
"step": 1955
|
1234 |
+
},
|
1235 |
{
|
1236 |
"epoch": 5.01,
|
1237 |
"learning_rate": 1.1082693947144077e-05,
|
|
|
1466 |
"loss": 0.4689,
|
1467 |
"step": 2340
|
1468 |
},
|
1469 |
+
{
|
1470 |
+
"epoch": 6.0,
|
1471 |
+
"eval_accuracy": 0.7116785040007192,
|
1472 |
+
"eval_auprc": 68.77509414907618,
|
1473 |
+
"eval_loss": 0.8716110587120056,
|
1474 |
+
"eval_micro f1 score": 71.16785040007191,
|
1475 |
+
"eval_runtime": 41.7514,
|
1476 |
+
"eval_samples_per_second": 266.41,
|
1477 |
+
"eval_steps_per_second": 8.335,
|
1478 |
+
"step": 2346
|
1479 |
+
},
|
1480 |
{
|
1481 |
"epoch": 6.01,
|
1482 |
"learning_rate": 8.86615515771526e-06,
|
|
|
1711 |
"loss": 0.433,
|
1712 |
"step": 2730
|
1713 |
},
|
1714 |
+
{
|
1715 |
+
"epoch": 7.0,
|
1716 |
+
"eval_accuracy": 0.7098804279421019,
|
1717 |
+
"eval_auprc": 68.36526805926971,
|
1718 |
+
"eval_loss": 0.9086081385612488,
|
1719 |
+
"eval_micro f1 score": 70.9880427942102,
|
1720 |
+
"eval_runtime": 41.7909,
|
1721 |
+
"eval_samples_per_second": 266.158,
|
1722 |
+
"eval_steps_per_second": 8.327,
|
1723 |
+
"step": 2737
|
1724 |
+
},
|
1725 |
{
|
1726 |
"epoch": 7.01,
|
1727 |
"learning_rate": 6.649616368286445e-06,
|
|
|
1956 |
"loss": 0.419,
|
1957 |
"step": 3120
|
1958 |
},
|
1959 |
+
{
|
1960 |
+
"epoch": 8.0,
|
1961 |
+
"eval_accuracy": 0.7067337948395217,
|
1962 |
+
"eval_auprc": 68.46062183107752,
|
1963 |
+
"eval_loss": 0.9290213584899902,
|
1964 |
+
"eval_micro f1 score": 70.67337948395217,
|
1965 |
+
"eval_runtime": 41.7253,
|
1966 |
+
"eval_samples_per_second": 266.577,
|
1967 |
+
"eval_steps_per_second": 8.34,
|
1968 |
+
"step": 3128
|
1969 |
+
},
|
1970 |
{
|
1971 |
"epoch": 8.01,
|
1972 |
"learning_rate": 4.43307757885763e-06,
|
|
|
2201 |
"loss": 0.3766,
|
2202 |
"step": 3510
|
2203 |
},
|
2204 |
+
{
|
2205 |
+
"epoch": 9.0,
|
2206 |
+
"eval_accuracy": 0.7064640834307291,
|
2207 |
+
"eval_auprc": 68.51324967918615,
|
2208 |
+
"eval_loss": 0.9618651866912842,
|
2209 |
+
"eval_micro f1 score": 70.6464083430729,
|
2210 |
+
"eval_runtime": 41.7225,
|
2211 |
+
"eval_samples_per_second": 266.595,
|
2212 |
+
"eval_steps_per_second": 8.341,
|
2213 |
+
"step": 3519
|
2214 |
+
},
|
2215 |
{
|
2216 |
"epoch": 9.0,
|
2217 |
"learning_rate": 2.216538789428815e-06,
|
|
|
2452 |
"loss": 0.3395,
|
2453 |
"step": 3910
|
2454 |
},
|
2455 |
+
{
|
2456 |
+
"epoch": 10.0,
|
2457 |
+
"eval_accuracy": 0.707183313854176,
|
2458 |
+
"eval_auprc": 68.45618753896134,
|
2459 |
+
"eval_loss": 0.971839189529419,
|
2460 |
+
"eval_micro f1 score": 70.71833138541761,
|
2461 |
+
"eval_runtime": 41.6752,
|
2462 |
+
"eval_samples_per_second": 266.897,
|
2463 |
+
"eval_steps_per_second": 8.35,
|
2464 |
+
"step": 3910
|
2465 |
+
},
|
2466 |
{
|
2467 |
"epoch": 10.0,
|
2468 |
"step": 3910,
|
2469 |
"total_flos": 1.0186138340339622e+17,
|
2470 |
"train_loss": 0.6484309974533823,
|
2471 |
+
"train_runtime": 11384.9561,
|
2472 |
+
"train_samples_per_second": 87.925,
|
2473 |
+
"train_steps_per_second": 0.343
|
2474 |
}
|
2475 |
],
|
2476 |
"max_steps": 3910,
|