AmberYifan commited on
Commit
be49355
1 Parent(s): 5894365

Training in progress, epoch 2, checkpoint

Browse files
last-checkpoint/global_step626/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6eaa45e9530f3d94283c55295f5df838550f3797754204bd6a943f3ba08f621c
3
  size 14496051368
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:72b28c87a0b569b3cc768e0a0029466a38d889ac3e8c323278451d1ca6270ff4
3
  size 14496051368
last-checkpoint/global_step626/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2601edf9aed3d175e477d9b3d8da64d09732b3c34c11dc4ce1d00f83b390b413
3
  size 14496051368
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea636794ca08cfc2e535604b34ee5822571583353c0e3a21b74e02a803fd9d2b
3
  size 14496051368
last-checkpoint/global_step626/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:fc8817173341cb0b47e4e8950086453c26262bc97554145c71baba5385504170
3
  size 14496051368
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f8395e47f5b1bd2c21bb34071c37647503760e7705fdd9fe69fe0b2cbeb592b1
3
  size 14496051368
last-checkpoint/global_step626/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d2e51c2ee812a2f36c46ba620c91d50fb73a6299cbd982523160da9bc1b2fd5a
3
  size 14496051368
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d5c92268c285f9c22feee30be7e771a26606e68b1e17ea0ed5aded9fc990889
3
  size 14496051368
last-checkpoint/latest CHANGED
@@ -1 +1 @@
1
- global_step313
 
1
+ global_step626
last-checkpoint/model-00001-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:32280386f8debc0ebe8ce635cd888ce70c9804ac22e7ec96cb0f751f6db5f30c
3
  size 4949453792
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3e3d6b8494c2b4bdbb752b4fff7c6da0258880b954076d5ef3e53ddbbe43d0a0
3
  size 4949453792
last-checkpoint/model-00002-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:02a68b4da40fc76a1cca231f1ff42e03c6fd243cf71f83203b26d45fe66e8233
3
  size 4999819336
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:563014950b0781d39fa4b245a7ac0e534496966d7f78e5a55a436236e1546492
3
  size 4999819336
last-checkpoint/model-00003-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:cec47342687cbe127b9cc2bcb4d712c96f982569137a77533c7430af0089db42
3
  size 4546807800
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b41dbd4fd684a2e89f7ed562d64fdcc3a134081743f54f515e9ddd58937bdbb8
3
  size 4546807800
last-checkpoint/rng_state_0.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f1ce770862b76ceecb3453a5d0b9ef65311eab0f6a60f52997726a89d718de08
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2307c03867cef25b5028feb9a23f80e784b9af9a615de13ddca560a6a90fb593
3
  size 15024
last-checkpoint/rng_state_1.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2bcae6573442578b752be3e988d4ccb38056b45c31b5c02eb579ec6cebcfa62e
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50218cfaecdd818354e567b7167c13899e3b42297e7d8f58bd7e732cfa547800
3
  size 15024
last-checkpoint/rng_state_2.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:398ff45072fee2975e88b9e078f915103d75d94bc08753303cf855b915973623
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9814a66b49861f5495b06dae3be12ddf7185b88e2cae1fb808ca9efd99d5807f
3
  size 15024
last-checkpoint/rng_state_3.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7f4b210bc832d4c648e3bee8c21dca26e5a8b365d6ec90c638062005a052e57b
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a7be93040a65e0a29975f6c70b94418e1fdf88423a50c58aa572141d3c92fbfc
3
  size 15024
last-checkpoint/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c019f2bd3222f739db75f458caa9c28765649010b6fa70739851be6f2b00c68b
3
  size 1064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c3fd160e69e5766b513d913785299b2ff6d181cae9fd47f741adc100cb17d22e
3
  size 1064
last-checkpoint/trainer_state.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
- "epoch": 1.0,
5
  "eval_steps": 500,
6
- "global_step": 313,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
@@ -503,6 +503,487 @@
503
  "eval_samples_per_second": 16.156,
504
  "eval_steps_per_second": 0.565,
505
  "step": 313
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
506
  }
507
  ],
508
  "logging_steps": 10,
 
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
+ "epoch": 2.0,
5
  "eval_steps": 500,
6
+ "global_step": 626,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
 
503
  "eval_samples_per_second": 16.156,
504
  "eval_steps_per_second": 0.565,
505
  "step": 313
506
+ },
507
+ {
508
+ "epoch": 1.0223642172523961,
509
+ "grad_norm": 0.001081863238066753,
510
+ "learning_rate": 3.662721893491124e-07,
511
+ "logits/chosen": -3.09375,
512
+ "logits/rejected": -2.875,
513
+ "logps/chosen": -178.0,
514
+ "logps/rejected": -340.0,
515
+ "loss": 0.0,
516
+ "rewards/accuracies": 1.0,
517
+ "rewards/chosen": -0.228515625,
518
+ "rewards/margins": 21.375,
519
+ "rewards/rejected": -21.5,
520
+ "step": 320
521
+ },
522
+ {
523
+ "epoch": 1.0543130990415335,
524
+ "grad_norm": 0.0014656429504141843,
525
+ "learning_rate": 3.603550295857988e-07,
526
+ "logits/chosen": -3.171875,
527
+ "logits/rejected": -2.90625,
528
+ "logps/chosen": -164.0,
529
+ "logps/rejected": -344.0,
530
+ "loss": 0.0,
531
+ "rewards/accuracies": 1.0,
532
+ "rewards/chosen": -0.0927734375,
533
+ "rewards/margins": 22.0,
534
+ "rewards/rejected": -22.0,
535
+ "step": 330
536
+ },
537
+ {
538
+ "epoch": 1.0862619808306708,
539
+ "grad_norm": 9.752625657144038e-05,
540
+ "learning_rate": 3.544378698224852e-07,
541
+ "logits/chosen": -3.171875,
542
+ "logits/rejected": -2.90625,
543
+ "logps/chosen": -153.0,
544
+ "logps/rejected": -338.0,
545
+ "loss": 0.0,
546
+ "rewards/accuracies": 1.0,
547
+ "rewards/chosen": -0.0031585693359375,
548
+ "rewards/margins": 20.875,
549
+ "rewards/rejected": -20.875,
550
+ "step": 340
551
+ },
552
+ {
553
+ "epoch": 1.1182108626198084,
554
+ "grad_norm": 5.907578925715793e-05,
555
+ "learning_rate": 3.485207100591716e-07,
556
+ "logits/chosen": -3.140625,
557
+ "logits/rejected": -2.90625,
558
+ "logps/chosen": -158.0,
559
+ "logps/rejected": -354.0,
560
+ "loss": 0.0,
561
+ "rewards/accuracies": 1.0,
562
+ "rewards/chosen": -0.1630859375,
563
+ "rewards/margins": 22.75,
564
+ "rewards/rejected": -22.875,
565
+ "step": 350
566
+ },
567
+ {
568
+ "epoch": 1.1501597444089458,
569
+ "grad_norm": 0.008399378571967797,
570
+ "learning_rate": 3.42603550295858e-07,
571
+ "logits/chosen": -3.140625,
572
+ "logits/rejected": -2.90625,
573
+ "logps/chosen": -175.0,
574
+ "logps/rejected": -348.0,
575
+ "loss": 0.0,
576
+ "rewards/accuracies": 1.0,
577
+ "rewards/chosen": -0.0084228515625,
578
+ "rewards/margins": 22.125,
579
+ "rewards/rejected": -22.125,
580
+ "step": 360
581
+ },
582
+ {
583
+ "epoch": 1.182108626198083,
584
+ "grad_norm": 0.00096142832181881,
585
+ "learning_rate": 3.366863905325444e-07,
586
+ "logits/chosen": -3.109375,
587
+ "logits/rejected": -2.875,
588
+ "logps/chosen": -146.0,
589
+ "logps/rejected": -356.0,
590
+ "loss": 0.0,
591
+ "rewards/accuracies": 1.0,
592
+ "rewards/chosen": -0.034912109375,
593
+ "rewards/margins": 22.875,
594
+ "rewards/rejected": -22.875,
595
+ "step": 370
596
+ },
597
+ {
598
+ "epoch": 1.2140575079872205,
599
+ "grad_norm": 1.117707679199626e-05,
600
+ "learning_rate": 3.3076923076923075e-07,
601
+ "logits/chosen": -3.140625,
602
+ "logits/rejected": -3.0,
603
+ "logps/chosen": -158.0,
604
+ "logps/rejected": -352.0,
605
+ "loss": 0.0,
606
+ "rewards/accuracies": 1.0,
607
+ "rewards/chosen": -0.10986328125,
608
+ "rewards/margins": 22.75,
609
+ "rewards/rejected": -22.75,
610
+ "step": 380
611
+ },
612
+ {
613
+ "epoch": 1.2460063897763578,
614
+ "grad_norm": 3.391687654228242e-05,
615
+ "learning_rate": 3.2485207100591716e-07,
616
+ "logits/chosen": -3.078125,
617
+ "logits/rejected": -2.84375,
618
+ "logps/chosen": -134.0,
619
+ "logps/rejected": -352.0,
620
+ "loss": 0.0,
621
+ "rewards/accuracies": 1.0,
622
+ "rewards/chosen": -0.0036773681640625,
623
+ "rewards/margins": 23.0,
624
+ "rewards/rejected": -23.0,
625
+ "step": 390
626
+ },
627
+ {
628
+ "epoch": 1.2779552715654952,
629
+ "grad_norm": 0.00015741926047163637,
630
+ "learning_rate": 3.189349112426035e-07,
631
+ "logits/chosen": -3.0,
632
+ "logits/rejected": -2.890625,
633
+ "logps/chosen": -141.0,
634
+ "logps/rejected": -358.0,
635
+ "loss": 0.0,
636
+ "rewards/accuracies": 1.0,
637
+ "rewards/chosen": -0.318359375,
638
+ "rewards/margins": 23.5,
639
+ "rewards/rejected": -23.75,
640
+ "step": 400
641
+ },
642
+ {
643
+ "epoch": 1.3099041533546325,
644
+ "grad_norm": 0.001381822064334206,
645
+ "learning_rate": 3.130177514792899e-07,
646
+ "logits/chosen": -3.09375,
647
+ "logits/rejected": -2.765625,
648
+ "logps/chosen": -150.0,
649
+ "logps/rejected": -360.0,
650
+ "loss": 0.0,
651
+ "rewards/accuracies": 1.0,
652
+ "rewards/chosen": -0.07568359375,
653
+ "rewards/margins": 24.125,
654
+ "rewards/rejected": -24.25,
655
+ "step": 410
656
+ },
657
+ {
658
+ "epoch": 1.34185303514377,
659
+ "grad_norm": 0.0008640673608836212,
660
+ "learning_rate": 3.071005917159763e-07,
661
+ "logits/chosen": -3.078125,
662
+ "logits/rejected": -2.9375,
663
+ "logps/chosen": -169.0,
664
+ "logps/rejected": -356.0,
665
+ "loss": 0.0,
666
+ "rewards/accuracies": 1.0,
667
+ "rewards/chosen": -0.244140625,
668
+ "rewards/margins": 23.25,
669
+ "rewards/rejected": -23.5,
670
+ "step": 420
671
+ },
672
+ {
673
+ "epoch": 1.3738019169329074,
674
+ "grad_norm": 0.00010813946327226968,
675
+ "learning_rate": 3.011834319526627e-07,
676
+ "logits/chosen": -3.125,
677
+ "logits/rejected": -2.859375,
678
+ "logps/chosen": -157.0,
679
+ "logps/rejected": -364.0,
680
+ "loss": 0.0,
681
+ "rewards/accuracies": 1.0,
682
+ "rewards/chosen": -0.2255859375,
683
+ "rewards/margins": 23.875,
684
+ "rewards/rejected": -24.125,
685
+ "step": 430
686
+ },
687
+ {
688
+ "epoch": 1.4057507987220448,
689
+ "grad_norm": 0.0001540987124577923,
690
+ "learning_rate": 2.952662721893491e-07,
691
+ "logits/chosen": -3.078125,
692
+ "logits/rejected": -2.859375,
693
+ "logps/chosen": -164.0,
694
+ "logps/rejected": -360.0,
695
+ "loss": 0.0,
696
+ "rewards/accuracies": 1.0,
697
+ "rewards/chosen": -0.13671875,
698
+ "rewards/margins": 23.375,
699
+ "rewards/rejected": -23.5,
700
+ "step": 440
701
+ },
702
+ {
703
+ "epoch": 1.4376996805111821,
704
+ "grad_norm": 0.0003904163069516865,
705
+ "learning_rate": 2.893491124260355e-07,
706
+ "logits/chosen": -3.0625,
707
+ "logits/rejected": -2.84375,
708
+ "logps/chosen": -166.0,
709
+ "logps/rejected": -364.0,
710
+ "loss": 0.0,
711
+ "rewards/accuracies": 1.0,
712
+ "rewards/chosen": -0.0277099609375,
713
+ "rewards/margins": 24.625,
714
+ "rewards/rejected": -24.75,
715
+ "step": 450
716
+ },
717
+ {
718
+ "epoch": 1.4696485623003195,
719
+ "grad_norm": 0.00010719442458275845,
720
+ "learning_rate": 2.834319526627219e-07,
721
+ "logits/chosen": -3.03125,
722
+ "logits/rejected": -2.84375,
723
+ "logps/chosen": -156.0,
724
+ "logps/rejected": -348.0,
725
+ "loss": 0.0,
726
+ "rewards/accuracies": 1.0,
727
+ "rewards/chosen": -0.431640625,
728
+ "rewards/margins": 22.25,
729
+ "rewards/rejected": -22.75,
730
+ "step": 460
731
+ },
732
+ {
733
+ "epoch": 1.5015974440894568,
734
+ "grad_norm": 4.819727273913693e-05,
735
+ "learning_rate": 2.7751479289940824e-07,
736
+ "logits/chosen": -3.03125,
737
+ "logits/rejected": -2.890625,
738
+ "logps/chosen": -164.0,
739
+ "logps/rejected": -362.0,
740
+ "loss": 0.0,
741
+ "rewards/accuracies": 1.0,
742
+ "rewards/chosen": -0.33203125,
743
+ "rewards/margins": 24.0,
744
+ "rewards/rejected": -24.375,
745
+ "step": 470
746
+ },
747
+ {
748
+ "epoch": 1.5335463258785942,
749
+ "grad_norm": 8.148529465105985e-05,
750
+ "learning_rate": 2.715976331360947e-07,
751
+ "logits/chosen": -3.03125,
752
+ "logits/rejected": -2.875,
753
+ "logps/chosen": -159.0,
754
+ "logps/rejected": -356.0,
755
+ "loss": 0.0,
756
+ "rewards/accuracies": 1.0,
757
+ "rewards/chosen": -0.318359375,
758
+ "rewards/margins": 22.875,
759
+ "rewards/rejected": -23.25,
760
+ "step": 480
761
+ },
762
+ {
763
+ "epoch": 1.5654952076677318,
764
+ "grad_norm": 0.00023912166796823171,
765
+ "learning_rate": 2.6568047337278106e-07,
766
+ "logits/chosen": -3.0625,
767
+ "logits/rejected": -2.84375,
768
+ "logps/chosen": -167.0,
769
+ "logps/rejected": -364.0,
770
+ "loss": 0.0,
771
+ "rewards/accuracies": 1.0,
772
+ "rewards/chosen": -0.33203125,
773
+ "rewards/margins": 23.875,
774
+ "rewards/rejected": -24.25,
775
+ "step": 490
776
+ },
777
+ {
778
+ "epoch": 1.5974440894568689,
779
+ "grad_norm": 0.00038137779946660004,
780
+ "learning_rate": 2.5976331360946746e-07,
781
+ "logits/chosen": -3.03125,
782
+ "logits/rejected": -2.765625,
783
+ "logps/chosen": -161.0,
784
+ "logps/rejected": -358.0,
785
+ "loss": 0.0,
786
+ "rewards/accuracies": 1.0,
787
+ "rewards/chosen": -0.09912109375,
788
+ "rewards/margins": 23.375,
789
+ "rewards/rejected": -23.5,
790
+ "step": 500
791
+ },
792
+ {
793
+ "epoch": 1.6293929712460065,
794
+ "grad_norm": 0.00020618761735157146,
795
+ "learning_rate": 2.538461538461538e-07,
796
+ "logits/chosen": -2.96875,
797
+ "logits/rejected": -2.75,
798
+ "logps/chosen": -164.0,
799
+ "logps/rejected": -374.0,
800
+ "loss": 0.0,
801
+ "rewards/accuracies": 1.0,
802
+ "rewards/chosen": -0.26953125,
803
+ "rewards/margins": 25.125,
804
+ "rewards/rejected": -25.375,
805
+ "step": 510
806
+ },
807
+ {
808
+ "epoch": 1.6613418530351438,
809
+ "grad_norm": 0.00173536920939462,
810
+ "learning_rate": 2.479289940828402e-07,
811
+ "logits/chosen": -2.984375,
812
+ "logits/rejected": -2.828125,
813
+ "logps/chosen": -143.0,
814
+ "logps/rejected": -378.0,
815
+ "loss": 0.0,
816
+ "rewards/accuracies": 1.0,
817
+ "rewards/chosen": -0.2431640625,
818
+ "rewards/margins": 25.125,
819
+ "rewards/rejected": -25.375,
820
+ "step": 520
821
+ },
822
+ {
823
+ "epoch": 1.6932907348242812,
824
+ "grad_norm": 0.00036883452678222923,
825
+ "learning_rate": 2.420118343195266e-07,
826
+ "logits/chosen": -2.96875,
827
+ "logits/rejected": -2.8125,
828
+ "logps/chosen": -179.0,
829
+ "logps/rejected": -374.0,
830
+ "loss": 0.0,
831
+ "rewards/accuracies": 1.0,
832
+ "rewards/chosen": -0.73828125,
833
+ "rewards/margins": 23.75,
834
+ "rewards/rejected": -24.5,
835
+ "step": 530
836
+ },
837
+ {
838
+ "epoch": 1.7252396166134185,
839
+ "grad_norm": 0.00046710302845877187,
840
+ "learning_rate": 2.36094674556213e-07,
841
+ "logits/chosen": -2.953125,
842
+ "logits/rejected": -2.78125,
843
+ "logps/chosen": -183.0,
844
+ "logps/rejected": -364.0,
845
+ "loss": 0.0,
846
+ "rewards/accuracies": 1.0,
847
+ "rewards/chosen": -0.54296875,
848
+ "rewards/margins": 24.25,
849
+ "rewards/rejected": -24.75,
850
+ "step": 540
851
+ },
852
+ {
853
+ "epoch": 1.7571884984025559,
854
+ "grad_norm": 0.0012311707501514388,
855
+ "learning_rate": 2.301775147928994e-07,
856
+ "logits/chosen": -2.984375,
857
+ "logits/rejected": -2.765625,
858
+ "logps/chosen": -184.0,
859
+ "logps/rejected": -370.0,
860
+ "loss": 0.0,
861
+ "rewards/accuracies": 1.0,
862
+ "rewards/chosen": -0.349609375,
863
+ "rewards/margins": 24.5,
864
+ "rewards/rejected": -24.875,
865
+ "step": 550
866
+ },
867
+ {
868
+ "epoch": 1.7891373801916934,
869
+ "grad_norm": 6.751851643657405e-05,
870
+ "learning_rate": 2.242603550295858e-07,
871
+ "logits/chosen": -2.9375,
872
+ "logits/rejected": -2.8125,
873
+ "logps/chosen": -155.0,
874
+ "logps/rejected": -360.0,
875
+ "loss": 0.0,
876
+ "rewards/accuracies": 1.0,
877
+ "rewards/chosen": -0.326171875,
878
+ "rewards/margins": 23.75,
879
+ "rewards/rejected": -24.0,
880
+ "step": 560
881
+ },
882
+ {
883
+ "epoch": 1.8210862619808306,
884
+ "grad_norm": 0.00011924326319637662,
885
+ "learning_rate": 2.1834319526627217e-07,
886
+ "logits/chosen": -3.0,
887
+ "logits/rejected": -2.8125,
888
+ "logps/chosen": -164.0,
889
+ "logps/rejected": -358.0,
890
+ "loss": 0.0,
891
+ "rewards/accuracies": 1.0,
892
+ "rewards/chosen": -0.40625,
893
+ "rewards/margins": 23.25,
894
+ "rewards/rejected": -23.625,
895
+ "step": 570
896
+ },
897
+ {
898
+ "epoch": 1.8530351437699681,
899
+ "grad_norm": 1.591754779283049e-05,
900
+ "learning_rate": 2.1242603550295858e-07,
901
+ "logits/chosen": -3.078125,
902
+ "logits/rejected": -2.8125,
903
+ "logps/chosen": -162.0,
904
+ "logps/rejected": -364.0,
905
+ "loss": 0.0,
906
+ "rewards/accuracies": 1.0,
907
+ "rewards/chosen": -0.369140625,
908
+ "rewards/margins": 24.0,
909
+ "rewards/rejected": -24.375,
910
+ "step": 580
911
+ },
912
+ {
913
+ "epoch": 1.8849840255591053,
914
+ "grad_norm": 0.001273109125943514,
915
+ "learning_rate": 2.0650887573964496e-07,
916
+ "logits/chosen": -3.03125,
917
+ "logits/rejected": -2.75,
918
+ "logps/chosen": -163.0,
919
+ "logps/rejected": -366.0,
920
+ "loss": 0.0,
921
+ "rewards/accuracies": 1.0,
922
+ "rewards/chosen": -0.3828125,
923
+ "rewards/margins": 23.25,
924
+ "rewards/rejected": -23.625,
925
+ "step": 590
926
+ },
927
+ {
928
+ "epoch": 1.9169329073482428,
929
+ "grad_norm": 5.600029962176166e-05,
930
+ "learning_rate": 2.0059171597633133e-07,
931
+ "logits/chosen": -2.953125,
932
+ "logits/rejected": -2.78125,
933
+ "logps/chosen": -158.0,
934
+ "logps/rejected": -368.0,
935
+ "loss": 0.0,
936
+ "rewards/accuracies": 1.0,
937
+ "rewards/chosen": -0.283203125,
938
+ "rewards/margins": 24.125,
939
+ "rewards/rejected": -24.5,
940
+ "step": 600
941
+ },
942
+ {
943
+ "epoch": 1.9488817891373802,
944
+ "grad_norm": 2.143142202843072e-05,
945
+ "learning_rate": 1.9467455621301774e-07,
946
+ "logits/chosen": -2.9375,
947
+ "logits/rejected": -2.734375,
948
+ "logps/chosen": -165.0,
949
+ "logps/rejected": -382.0,
950
+ "loss": 0.0,
951
+ "rewards/accuracies": 1.0,
952
+ "rewards/chosen": -0.68359375,
953
+ "rewards/margins": 25.125,
954
+ "rewards/rejected": -25.75,
955
+ "step": 610
956
+ },
957
+ {
958
+ "epoch": 1.9808306709265175,
959
+ "grad_norm": 0.00015877860321082202,
960
+ "learning_rate": 1.8875739644970412e-07,
961
+ "logits/chosen": -2.921875,
962
+ "logits/rejected": -2.734375,
963
+ "logps/chosen": -153.0,
964
+ "logps/rejected": -380.0,
965
+ "loss": 0.0,
966
+ "rewards/accuracies": 1.0,
967
+ "rewards/chosen": -0.6015625,
968
+ "rewards/margins": 24.625,
969
+ "rewards/rejected": -25.25,
970
+ "step": 620
971
+ },
972
+ {
973
+ "epoch": 2.0,
974
+ "eval_logits/chosen": -3.0,
975
+ "eval_logits/rejected": -2.75,
976
+ "eval_logps/chosen": -155.0,
977
+ "eval_logps/rejected": -376.0,
978
+ "eval_loss": 1.6905501070141327e-06,
979
+ "eval_rewards/accuracies": 1.0,
980
+ "eval_rewards/chosen": -0.443359375,
981
+ "eval_rewards/margins": 25.25,
982
+ "eval_rewards/rejected": -25.625,
983
+ "eval_runtime": 12.4078,
984
+ "eval_samples_per_second": 16.119,
985
+ "eval_steps_per_second": 0.564,
986
+ "step": 626
987
  }
988
  ],
989
  "logging_steps": 10,