GlycerinLOL
commited on
Commit
•
19f8eba
1
Parent(s):
a7630b1
End of training
Browse files- all_results.json +5 -5
- train_results.json +5 -5
- trainer_state.json +204 -22
all_results.json
CHANGED
@@ -1,7 +1,7 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"train_loss":
|
4 |
-
"train_runtime":
|
5 |
-
"train_samples_per_second":
|
6 |
-
"train_steps_per_second": 0.
|
7 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 24.0,
|
3 |
+
"train_loss": 0.22391605226564926,
|
4 |
+
"train_runtime": 17118.5241,
|
5 |
+
"train_samples_per_second": 70.1,
|
6 |
+
"train_steps_per_second": 0.73
|
7 |
}
|
train_results.json
CHANGED
@@ -1,7 +1,7 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"train_loss":
|
4 |
-
"train_runtime":
|
5 |
-
"train_samples_per_second":
|
6 |
-
"train_steps_per_second": 0.
|
7 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 24.0,
|
3 |
+
"train_loss": 0.22391605226564926,
|
4 |
+
"train_runtime": 17118.5241,
|
5 |
+
"train_samples_per_second": 70.1,
|
6 |
+
"train_steps_per_second": 0.73
|
7 |
}
|
trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch":
|
5 |
"eval_steps": 500,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -346,36 +346,218 @@
|
|
346 |
},
|
347 |
{
|
348 |
"epoch": 16.0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
349 |
"eval_f1": 0.9034,
|
350 |
-
"eval_gen_len": 19.
|
351 |
-
"eval_loss": 1.
|
352 |
"eval_precision": 0.9159,
|
353 |
"eval_recall": 0.8916,
|
354 |
-
"eval_rouge1": 0.
|
355 |
-
"eval_rouge2": 0.
|
356 |
-
"eval_rougeL": 0.
|
357 |
-
"eval_rougeLsum": 0.
|
358 |
-
"eval_runtime":
|
359 |
-
"eval_samples_per_second": 8.
|
360 |
-
"eval_steps_per_second": 0.
|
361 |
-
"step":
|
362 |
},
|
363 |
{
|
364 |
-
"epoch":
|
365 |
-
"
|
366 |
-
"
|
367 |
-
"
|
368 |
-
|
369 |
-
|
370 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
371 |
}
|
372 |
],
|
373 |
"logging_steps": 500,
|
374 |
-
"max_steps":
|
375 |
"num_input_tokens_seen": 0,
|
376 |
-
"num_train_epochs":
|
377 |
"save_steps": 500,
|
378 |
-
"total_flos":
|
379 |
"train_batch_size": 24,
|
380 |
"trial_name": null,
|
381 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 24.0,
|
5 |
"eval_steps": 500,
|
6 |
+
"global_step": 12504,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
346 |
},
|
347 |
{
|
348 |
"epoch": 16.0,
|
349 |
+
"eval_f1": 0.9025,
|
350 |
+
"eval_gen_len": 19.942545454545453,
|
351 |
+
"eval_loss": 1.5439822673797607,
|
352 |
+
"eval_precision": 0.9151,
|
353 |
+
"eval_recall": 0.8905,
|
354 |
+
"eval_rouge1": 0.4427,
|
355 |
+
"eval_rouge2": 0.225,
|
356 |
+
"eval_rougeL": 0.382,
|
357 |
+
"eval_rougeLsum": 0.382,
|
358 |
+
"eval_runtime": 314.8749,
|
359 |
+
"eval_samples_per_second": 8.734,
|
360 |
+
"eval_steps_per_second": 0.546,
|
361 |
+
"step": 8336
|
362 |
+
},
|
363 |
+
{
|
364 |
+
"epoch": 16.31,
|
365 |
+
"learning_rate": 6.404350607805503e-06,
|
366 |
+
"loss": 0.8806,
|
367 |
+
"step": 8500
|
368 |
+
},
|
369 |
+
{
|
370 |
+
"epoch": 17.0,
|
371 |
+
"eval_f1": 0.9036,
|
372 |
+
"eval_gen_len": 19.88509090909091,
|
373 |
+
"eval_loss": 1.5509530305862427,
|
374 |
+
"eval_precision": 0.9159,
|
375 |
+
"eval_recall": 0.8919,
|
376 |
+
"eval_rouge1": 0.4495,
|
377 |
+
"eval_rouge2": 0.2279,
|
378 |
+
"eval_rougeL": 0.3868,
|
379 |
+
"eval_rougeLsum": 0.3869,
|
380 |
+
"eval_runtime": 312.7951,
|
381 |
+
"eval_samples_per_second": 8.792,
|
382 |
+
"eval_steps_per_second": 0.55,
|
383 |
+
"step": 8857
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 17.27,
|
387 |
+
"learning_rate": 5.6046065259117085e-06,
|
388 |
+
"loss": 0.8683,
|
389 |
+
"step": 9000
|
390 |
+
},
|
391 |
+
{
|
392 |
+
"epoch": 18.0,
|
393 |
+
"eval_f1": 0.9038,
|
394 |
+
"eval_gen_len": 19.88290909090909,
|
395 |
+
"eval_loss": 1.56792151927948,
|
396 |
+
"eval_precision": 0.9161,
|
397 |
+
"eval_recall": 0.8921,
|
398 |
+
"eval_rouge1": 0.4473,
|
399 |
+
"eval_rouge2": 0.2282,
|
400 |
+
"eval_rougeL": 0.3856,
|
401 |
+
"eval_rougeLsum": 0.3857,
|
402 |
+
"eval_runtime": 314.8371,
|
403 |
+
"eval_samples_per_second": 8.735,
|
404 |
+
"eval_steps_per_second": 0.546,
|
405 |
+
"step": 9378
|
406 |
+
},
|
407 |
+
{
|
408 |
+
"epoch": 18.23,
|
409 |
+
"learning_rate": 4.804862444017915e-06,
|
410 |
+
"loss": 0.8413,
|
411 |
+
"step": 9500
|
412 |
+
},
|
413 |
+
{
|
414 |
+
"epoch": 19.0,
|
415 |
+
"eval_f1": 0.9035,
|
416 |
+
"eval_gen_len": 19.913454545454545,
|
417 |
+
"eval_loss": 1.574545979499817,
|
418 |
+
"eval_precision": 0.9159,
|
419 |
+
"eval_recall": 0.8918,
|
420 |
+
"eval_rouge1": 0.4492,
|
421 |
+
"eval_rouge2": 0.2282,
|
422 |
+
"eval_rougeL": 0.3861,
|
423 |
+
"eval_rougeLsum": 0.3864,
|
424 |
+
"eval_runtime": 311.5846,
|
425 |
+
"eval_samples_per_second": 8.826,
|
426 |
+
"eval_steps_per_second": 0.552,
|
427 |
+
"step": 9899
|
428 |
+
},
|
429 |
+
{
|
430 |
+
"epoch": 19.19,
|
431 |
+
"learning_rate": 4.005118362124121e-06,
|
432 |
+
"loss": 0.8257,
|
433 |
+
"step": 10000
|
434 |
+
},
|
435 |
+
{
|
436 |
+
"epoch": 20.0,
|
437 |
+
"eval_f1": 0.9031,
|
438 |
+
"eval_gen_len": 19.899636363636365,
|
439 |
+
"eval_loss": 1.583512544631958,
|
440 |
+
"eval_precision": 0.9153,
|
441 |
+
"eval_recall": 0.8915,
|
442 |
+
"eval_rouge1": 0.4471,
|
443 |
+
"eval_rouge2": 0.2266,
|
444 |
+
"eval_rougeL": 0.3852,
|
445 |
+
"eval_rougeLsum": 0.3853,
|
446 |
+
"eval_runtime": 311.7771,
|
447 |
+
"eval_samples_per_second": 8.82,
|
448 |
+
"eval_steps_per_second": 0.552,
|
449 |
+
"step": 10420
|
450 |
+
},
|
451 |
+
{
|
452 |
+
"epoch": 20.15,
|
453 |
+
"learning_rate": 3.2053742802303266e-06,
|
454 |
+
"loss": 0.8097,
|
455 |
+
"step": 10500
|
456 |
+
},
|
457 |
+
{
|
458 |
+
"epoch": 21.0,
|
459 |
+
"eval_f1": 0.9034,
|
460 |
+
"eval_gen_len": 19.907272727272726,
|
461 |
+
"eval_loss": 1.59569251537323,
|
462 |
+
"eval_precision": 0.9156,
|
463 |
+
"eval_recall": 0.8919,
|
464 |
+
"eval_rouge1": 0.4472,
|
465 |
+
"eval_rouge2": 0.2271,
|
466 |
+
"eval_rougeL": 0.3856,
|
467 |
+
"eval_rougeLsum": 0.3856,
|
468 |
+
"eval_runtime": 309.5923,
|
469 |
+
"eval_samples_per_second": 8.883,
|
470 |
+
"eval_steps_per_second": 0.556,
|
471 |
+
"step": 10941
|
472 |
+
},
|
473 |
+
{
|
474 |
+
"epoch": 21.11,
|
475 |
+
"learning_rate": 2.4056301983365325e-06,
|
476 |
+
"loss": 0.7926,
|
477 |
+
"step": 11000
|
478 |
+
},
|
479 |
+
{
|
480 |
+
"epoch": 22.0,
|
481 |
"eval_f1": 0.9034,
|
482 |
+
"eval_gen_len": 19.892,
|
483 |
+
"eval_loss": 1.595582127571106,
|
484 |
"eval_precision": 0.9159,
|
485 |
"eval_recall": 0.8916,
|
486 |
+
"eval_rouge1": 0.4479,
|
487 |
+
"eval_rouge2": 0.2282,
|
488 |
+
"eval_rougeL": 0.3855,
|
489 |
+
"eval_rougeLsum": 0.3857,
|
490 |
+
"eval_runtime": 311.5772,
|
491 |
+
"eval_samples_per_second": 8.826,
|
492 |
+
"eval_steps_per_second": 0.552,
|
493 |
+
"step": 11462
|
494 |
},
|
495 |
{
|
496 |
+
"epoch": 22.07,
|
497 |
+
"learning_rate": 1.6058861164427384e-06,
|
498 |
+
"loss": 0.7841,
|
499 |
+
"step": 11500
|
500 |
+
},
|
501 |
+
{
|
502 |
+
"epoch": 23.0,
|
503 |
+
"eval_f1": 0.9028,
|
504 |
+
"eval_gen_len": 19.912,
|
505 |
+
"eval_loss": 1.5990447998046875,
|
506 |
+
"eval_precision": 0.9155,
|
507 |
+
"eval_recall": 0.8908,
|
508 |
+
"eval_rouge1": 0.4444,
|
509 |
+
"eval_rouge2": 0.2261,
|
510 |
+
"eval_rougeL": 0.3833,
|
511 |
+
"eval_rougeLsum": 0.3834,
|
512 |
+
"eval_runtime": 311.6057,
|
513 |
+
"eval_samples_per_second": 8.825,
|
514 |
+
"eval_steps_per_second": 0.552,
|
515 |
+
"step": 11983
|
516 |
+
},
|
517 |
+
{
|
518 |
+
"epoch": 23.03,
|
519 |
+
"learning_rate": 8.061420345489445e-07,
|
520 |
+
"loss": 0.7734,
|
521 |
+
"step": 12000
|
522 |
+
},
|
523 |
+
{
|
524 |
+
"epoch": 23.99,
|
525 |
+
"learning_rate": 6.397952655150352e-09,
|
526 |
+
"loss": 0.7669,
|
527 |
+
"step": 12500
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 24.0,
|
531 |
+
"eval_f1": 0.9034,
|
532 |
+
"eval_gen_len": 19.90290909090909,
|
533 |
+
"eval_loss": 1.6053136587142944,
|
534 |
+
"eval_precision": 0.9159,
|
535 |
+
"eval_recall": 0.8916,
|
536 |
+
"eval_rouge1": 0.4481,
|
537 |
+
"eval_rouge2": 0.2283,
|
538 |
+
"eval_rougeL": 0.3861,
|
539 |
+
"eval_rougeLsum": 0.3863,
|
540 |
+
"eval_runtime": 314.5795,
|
541 |
+
"eval_samples_per_second": 8.742,
|
542 |
+
"eval_steps_per_second": 0.547,
|
543 |
+
"step": 12504
|
544 |
+
},
|
545 |
+
{
|
546 |
+
"epoch": 24.0,
|
547 |
+
"step": 12504,
|
548 |
+
"total_flos": 2.569106349028344e+18,
|
549 |
+
"train_loss": 0.22391605226564926,
|
550 |
+
"train_runtime": 17118.5241,
|
551 |
+
"train_samples_per_second": 70.1,
|
552 |
+
"train_steps_per_second": 0.73
|
553 |
}
|
554 |
],
|
555 |
"logging_steps": 500,
|
556 |
+
"max_steps": 12504,
|
557 |
"num_input_tokens_seen": 0,
|
558 |
+
"num_train_epochs": 24,
|
559 |
"save_steps": 500,
|
560 |
+
"total_flos": 2.569106349028344e+18,
|
561 |
"train_batch_size": 24,
|
562 |
"trial_name": null,
|
563 |
"trial_params": null
|