Training in progress, epoch 2, checkpoint
Browse files- last-checkpoint/global_step626/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step626/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step626/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step626/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step626/zero_pp_rank_0_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step626/zero_pp_rank_1_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step626/zero_pp_rank_2_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step626/zero_pp_rank_3_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/latest +1 -1
- last-checkpoint/model-00001-of-00003.safetensors +1 -1
- last-checkpoint/model-00002-of-00003.safetensors +1 -1
- last-checkpoint/model-00003-of-00003.safetensors +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +483 -2
last-checkpoint/global_step626/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:27d585ba1697963257544700f540088a3ddadcd486dcf5a0ea6b2fb75f76beec
|
| 3 |
+
size 13476836524
|
last-checkpoint/global_step626/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0521bcb74c18ce0b90945bad53d5b74da1f448635973466e357fcafd351fa2f2
|
| 3 |
+
size 13476836524
|
last-checkpoint/global_step626/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b9a724112c06ae67f5a02507130d435d754f58df3bd7e7a5fd1b094fc9c49e1
|
| 3 |
+
size 13476836524
|
last-checkpoint/global_step626/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:af0d6be1d02c7d519d3c268c2a4d05d69b6cb8c6f62b23b94682bb310164151e
|
| 3 |
+
size 13476836524
|
last-checkpoint/global_step626/zero_pp_rank_0_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0f04249e1b8fe1c2cdc217cf32f7eea80a19cb28668041a7568bda12381df332
|
| 3 |
+
size 150693
|
last-checkpoint/global_step626/zero_pp_rank_1_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2a8e7204e992a56a415d70b4b33a5427de747e33a8410a8d2e57b2d3760f41a3
|
| 3 |
+
size 150693
|
last-checkpoint/global_step626/zero_pp_rank_2_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b82bc0de7fa407ebf2e63d8203a63e23579c6341814abb2367c0d8f0fb95b0c2
|
| 3 |
+
size 150693
|
last-checkpoint/global_step626/zero_pp_rank_3_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f568cb4b0ff22d0f18efb7a633cdedc227f3a80be6b0c1fcc6c09a4985216c2c
|
| 3 |
+
size 150693
|
last-checkpoint/latest
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
|
|
|
|
| 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:
|
| 3 |
size 4938985352
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:762cae3805ca348bace68eb88216168639e55ee5b02ddf91eab2dd85a4701bfc
|
| 3 |
size 4938985352
|
last-checkpoint/model-00002-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4947390880
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b6693d0730f6e5d8214e7d5cd3a70e167636ab0244b92239035a21395008abbc
|
| 3 |
size 4947390880
|
last-checkpoint/model-00003-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3590488816
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:97524e01d50d496c8aee01dc63f0f72b0bd6f0924d4cdcf2e9fd8b9a8aebd114
|
| 3 |
size 3590488816
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 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:
|
| 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:
|
| 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:
|
| 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:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7725b44c69725321786f1e58dccd7ea4d3ae5794ea47bd49c0d4a139aec266d
|
| 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":
|
| 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,
|
|
@@ -503,6 +503,487 @@
|
|
| 503 |
"eval_samples_per_second": 16.149,
|
| 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": 16.149,
|
| 504 |
"eval_steps_per_second": 0.565,
|
| 505 |
"step": 313
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.0223642172523961,
|
| 509 |
+
"grad_norm": 15.294291429304797,
|
| 510 |
+
"learning_rate": 3.662721893491124e-07,
|
| 511 |
+
"logits/chosen": -1.515625,
|
| 512 |
+
"logits/rejected": -1.6015625,
|
| 513 |
+
"logps/chosen": -242.0,
|
| 514 |
+
"logps/rejected": -288.0,
|
| 515 |
+
"loss": 0.3827,
|
| 516 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 517 |
+
"rewards/chosen": 0.53515625,
|
| 518 |
+
"rewards/margins": 0.9765625,
|
| 519 |
+
"rewards/rejected": -0.443359375,
|
| 520 |
+
"step": 320
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.0543130990415335,
|
| 524 |
+
"grad_norm": 17.823605067222292,
|
| 525 |
+
"learning_rate": 3.603550295857988e-07,
|
| 526 |
+
"logits/chosen": -1.546875,
|
| 527 |
+
"logits/rejected": -1.59375,
|
| 528 |
+
"logps/chosen": -237.0,
|
| 529 |
+
"logps/rejected": -296.0,
|
| 530 |
+
"loss": 0.3641,
|
| 531 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 532 |
+
"rewards/chosen": 0.609375,
|
| 533 |
+
"rewards/margins": 1.171875,
|
| 534 |
+
"rewards/rejected": -0.5546875,
|
| 535 |
+
"step": 330
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.0862619808306708,
|
| 539 |
+
"grad_norm": 17.184762171933475,
|
| 540 |
+
"learning_rate": 3.544378698224852e-07,
|
| 541 |
+
"logits/chosen": -1.625,
|
| 542 |
+
"logits/rejected": -1.6484375,
|
| 543 |
+
"logps/chosen": -246.0,
|
| 544 |
+
"logps/rejected": -294.0,
|
| 545 |
+
"loss": 0.3731,
|
| 546 |
+
"rewards/accuracies": 0.8125,
|
| 547 |
+
"rewards/chosen": 0.6171875,
|
| 548 |
+
"rewards/margins": 1.1328125,
|
| 549 |
+
"rewards/rejected": -0.515625,
|
| 550 |
+
"step": 340
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 1.1182108626198084,
|
| 554 |
+
"grad_norm": 17.160818312302805,
|
| 555 |
+
"learning_rate": 3.485207100591716e-07,
|
| 556 |
+
"logits/chosen": -1.609375,
|
| 557 |
+
"logits/rejected": -1.6171875,
|
| 558 |
+
"logps/chosen": -227.0,
|
| 559 |
+
"logps/rejected": -300.0,
|
| 560 |
+
"loss": 0.3685,
|
| 561 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 562 |
+
"rewards/chosen": 0.7265625,
|
| 563 |
+
"rewards/margins": 1.4296875,
|
| 564 |
+
"rewards/rejected": -0.703125,
|
| 565 |
+
"step": 350
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 1.1501597444089458,
|
| 569 |
+
"grad_norm": 17.582344513710463,
|
| 570 |
+
"learning_rate": 3.42603550295858e-07,
|
| 571 |
+
"logits/chosen": -1.59375,
|
| 572 |
+
"logits/rejected": -1.6796875,
|
| 573 |
+
"logps/chosen": -245.0,
|
| 574 |
+
"logps/rejected": -290.0,
|
| 575 |
+
"loss": 0.3573,
|
| 576 |
+
"rewards/accuracies": 0.75,
|
| 577 |
+
"rewards/chosen": 0.5546875,
|
| 578 |
+
"rewards/margins": 1.046875,
|
| 579 |
+
"rewards/rejected": -0.494140625,
|
| 580 |
+
"step": 360
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 1.182108626198083,
|
| 584 |
+
"grad_norm": 13.601144206004982,
|
| 585 |
+
"learning_rate": 3.366863905325444e-07,
|
| 586 |
+
"logits/chosen": -1.59375,
|
| 587 |
+
"logits/rejected": -1.6484375,
|
| 588 |
+
"logps/chosen": -237.0,
|
| 589 |
+
"logps/rejected": -296.0,
|
| 590 |
+
"loss": 0.3262,
|
| 591 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 592 |
+
"rewards/chosen": 0.75390625,
|
| 593 |
+
"rewards/margins": 1.328125,
|
| 594 |
+
"rewards/rejected": -0.57421875,
|
| 595 |
+
"step": 370
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.2140575079872205,
|
| 599 |
+
"grad_norm": 14.002962198108948,
|
| 600 |
+
"learning_rate": 3.3076923076923075e-07,
|
| 601 |
+
"logits/chosen": -1.625,
|
| 602 |
+
"logits/rejected": -1.65625,
|
| 603 |
+
"logps/chosen": -246.0,
|
| 604 |
+
"logps/rejected": -284.0,
|
| 605 |
+
"loss": 0.3248,
|
| 606 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 607 |
+
"rewards/chosen": 0.71484375,
|
| 608 |
+
"rewards/margins": 1.1953125,
|
| 609 |
+
"rewards/rejected": -0.484375,
|
| 610 |
+
"step": 380
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 1.2460063897763578,
|
| 614 |
+
"grad_norm": 16.239124276037153,
|
| 615 |
+
"learning_rate": 3.2485207100591716e-07,
|
| 616 |
+
"logits/chosen": -1.6171875,
|
| 617 |
+
"logits/rejected": -1.609375,
|
| 618 |
+
"logps/chosen": -237.0,
|
| 619 |
+
"logps/rejected": -298.0,
|
| 620 |
+
"loss": 0.3467,
|
| 621 |
+
"rewards/accuracies": 0.824999988079071,
|
| 622 |
+
"rewards/chosen": 0.84765625,
|
| 623 |
+
"rewards/margins": 1.4453125,
|
| 624 |
+
"rewards/rejected": -0.59375,
|
| 625 |
+
"step": 390
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.2779552715654952,
|
| 629 |
+
"grad_norm": 16.58669423900164,
|
| 630 |
+
"learning_rate": 3.189349112426035e-07,
|
| 631 |
+
"logits/chosen": -1.5546875,
|
| 632 |
+
"logits/rejected": -1.6171875,
|
| 633 |
+
"logps/chosen": -241.0,
|
| 634 |
+
"logps/rejected": -288.0,
|
| 635 |
+
"loss": 0.292,
|
| 636 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 637 |
+
"rewards/chosen": 0.8515625,
|
| 638 |
+
"rewards/margins": 1.4375,
|
| 639 |
+
"rewards/rejected": -0.58984375,
|
| 640 |
+
"step": 400
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.3099041533546325,
|
| 644 |
+
"grad_norm": 14.052814099588554,
|
| 645 |
+
"learning_rate": 3.130177514792899e-07,
|
| 646 |
+
"logits/chosen": -1.609375,
|
| 647 |
+
"logits/rejected": -1.640625,
|
| 648 |
+
"logps/chosen": -243.0,
|
| 649 |
+
"logps/rejected": -296.0,
|
| 650 |
+
"loss": 0.3533,
|
| 651 |
+
"rewards/accuracies": 0.800000011920929,
|
| 652 |
+
"rewards/chosen": 0.6875,
|
| 653 |
+
"rewards/margins": 1.390625,
|
| 654 |
+
"rewards/rejected": -0.703125,
|
| 655 |
+
"step": 410
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.34185303514377,
|
| 659 |
+
"grad_norm": 13.679395410389454,
|
| 660 |
+
"learning_rate": 3.071005917159763e-07,
|
| 661 |
+
"logits/chosen": -1.640625,
|
| 662 |
+
"logits/rejected": -1.703125,
|
| 663 |
+
"logps/chosen": -241.0,
|
| 664 |
+
"logps/rejected": -304.0,
|
| 665 |
+
"loss": 0.3461,
|
| 666 |
+
"rewards/accuracies": 0.862500011920929,
|
| 667 |
+
"rewards/chosen": 0.61328125,
|
| 668 |
+
"rewards/margins": 1.4765625,
|
| 669 |
+
"rewards/rejected": -0.86328125,
|
| 670 |
+
"step": 420
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 1.3738019169329074,
|
| 674 |
+
"grad_norm": 18.066697781388108,
|
| 675 |
+
"learning_rate": 3.011834319526627e-07,
|
| 676 |
+
"logits/chosen": -1.5859375,
|
| 677 |
+
"logits/rejected": -1.6953125,
|
| 678 |
+
"logps/chosen": -236.0,
|
| 679 |
+
"logps/rejected": -298.0,
|
| 680 |
+
"loss": 0.3133,
|
| 681 |
+
"rewards/accuracies": 0.862500011920929,
|
| 682 |
+
"rewards/chosen": 0.765625,
|
| 683 |
+
"rewards/margins": 1.578125,
|
| 684 |
+
"rewards/rejected": -0.8125,
|
| 685 |
+
"step": 430
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.4057507987220448,
|
| 689 |
+
"grad_norm": 12.738714916763403,
|
| 690 |
+
"learning_rate": 2.952662721893491e-07,
|
| 691 |
+
"logits/chosen": -1.609375,
|
| 692 |
+
"logits/rejected": -1.625,
|
| 693 |
+
"logps/chosen": -235.0,
|
| 694 |
+
"logps/rejected": -300.0,
|
| 695 |
+
"loss": 0.315,
|
| 696 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 697 |
+
"rewards/chosen": 0.67578125,
|
| 698 |
+
"rewards/margins": 1.515625,
|
| 699 |
+
"rewards/rejected": -0.83984375,
|
| 700 |
+
"step": 440
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 1.4376996805111821,
|
| 704 |
+
"grad_norm": 18.372879735041497,
|
| 705 |
+
"learning_rate": 2.893491124260355e-07,
|
| 706 |
+
"logits/chosen": -1.59375,
|
| 707 |
+
"logits/rejected": -1.6328125,
|
| 708 |
+
"logps/chosen": -240.0,
|
| 709 |
+
"logps/rejected": -296.0,
|
| 710 |
+
"loss": 0.3155,
|
| 711 |
+
"rewards/accuracies": 0.862500011920929,
|
| 712 |
+
"rewards/chosen": 0.60546875,
|
| 713 |
+
"rewards/margins": 1.5390625,
|
| 714 |
+
"rewards/rejected": -0.92578125,
|
| 715 |
+
"step": 450
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 1.4696485623003195,
|
| 719 |
+
"grad_norm": 16.824060938394247,
|
| 720 |
+
"learning_rate": 2.834319526627219e-07,
|
| 721 |
+
"logits/chosen": -1.5546875,
|
| 722 |
+
"logits/rejected": -1.609375,
|
| 723 |
+
"logps/chosen": -241.0,
|
| 724 |
+
"logps/rejected": -292.0,
|
| 725 |
+
"loss": 0.3328,
|
| 726 |
+
"rewards/accuracies": 0.800000011920929,
|
| 727 |
+
"rewards/chosen": 0.5859375,
|
| 728 |
+
"rewards/margins": 1.34375,
|
| 729 |
+
"rewards/rejected": -0.7578125,
|
| 730 |
+
"step": 460
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 1.5015974440894568,
|
| 734 |
+
"grad_norm": 15.847190331363642,
|
| 735 |
+
"learning_rate": 2.7751479289940824e-07,
|
| 736 |
+
"logits/chosen": -1.5546875,
|
| 737 |
+
"logits/rejected": -1.6171875,
|
| 738 |
+
"logps/chosen": -238.0,
|
| 739 |
+
"logps/rejected": -292.0,
|
| 740 |
+
"loss": 0.3183,
|
| 741 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 742 |
+
"rewards/chosen": 0.73828125,
|
| 743 |
+
"rewards/margins": 1.5390625,
|
| 744 |
+
"rewards/rejected": -0.80078125,
|
| 745 |
+
"step": 470
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.5335463258785942,
|
| 749 |
+
"grad_norm": 16.6941890115174,
|
| 750 |
+
"learning_rate": 2.715976331360947e-07,
|
| 751 |
+
"logits/chosen": -1.625,
|
| 752 |
+
"logits/rejected": -1.6171875,
|
| 753 |
+
"logps/chosen": -248.0,
|
| 754 |
+
"logps/rejected": -292.0,
|
| 755 |
+
"loss": 0.3562,
|
| 756 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 757 |
+
"rewards/chosen": 0.66796875,
|
| 758 |
+
"rewards/margins": 1.4453125,
|
| 759 |
+
"rewards/rejected": -0.78125,
|
| 760 |
+
"step": 480
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 1.5654952076677318,
|
| 764 |
+
"grad_norm": 17.58079153854403,
|
| 765 |
+
"learning_rate": 2.6568047337278106e-07,
|
| 766 |
+
"logits/chosen": -1.609375,
|
| 767 |
+
"logits/rejected": -1.6796875,
|
| 768 |
+
"logps/chosen": -234.0,
|
| 769 |
+
"logps/rejected": -296.0,
|
| 770 |
+
"loss": 0.3033,
|
| 771 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 772 |
+
"rewards/chosen": 0.7578125,
|
| 773 |
+
"rewards/margins": 1.5859375,
|
| 774 |
+
"rewards/rejected": -0.828125,
|
| 775 |
+
"step": 490
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 1.5974440894568689,
|
| 779 |
+
"grad_norm": 12.797873931195115,
|
| 780 |
+
"learning_rate": 2.5976331360946746e-07,
|
| 781 |
+
"logits/chosen": -1.640625,
|
| 782 |
+
"logits/rejected": -1.6875,
|
| 783 |
+
"logps/chosen": -235.0,
|
| 784 |
+
"logps/rejected": -294.0,
|
| 785 |
+
"loss": 0.3326,
|
| 786 |
+
"rewards/accuracies": 0.8125,
|
| 787 |
+
"rewards/chosen": 0.76171875,
|
| 788 |
+
"rewards/margins": 1.6171875,
|
| 789 |
+
"rewards/rejected": -0.85546875,
|
| 790 |
+
"step": 500
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 1.6293929712460065,
|
| 794 |
+
"grad_norm": 16.7014213517602,
|
| 795 |
+
"learning_rate": 2.538461538461538e-07,
|
| 796 |
+
"logits/chosen": -1.5703125,
|
| 797 |
+
"logits/rejected": -1.6796875,
|
| 798 |
+
"logps/chosen": -237.0,
|
| 799 |
+
"logps/rejected": -296.0,
|
| 800 |
+
"loss": 0.342,
|
| 801 |
+
"rewards/accuracies": 0.800000011920929,
|
| 802 |
+
"rewards/chosen": 0.474609375,
|
| 803 |
+
"rewards/margins": 1.3046875,
|
| 804 |
+
"rewards/rejected": -0.83203125,
|
| 805 |
+
"step": 510
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 1.6613418530351438,
|
| 809 |
+
"grad_norm": 16.319840171294768,
|
| 810 |
+
"learning_rate": 2.479289940828402e-07,
|
| 811 |
+
"logits/chosen": -1.59375,
|
| 812 |
+
"logits/rejected": -1.6875,
|
| 813 |
+
"logps/chosen": -233.0,
|
| 814 |
+
"logps/rejected": -298.0,
|
| 815 |
+
"loss": 0.3033,
|
| 816 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 817 |
+
"rewards/chosen": 0.734375,
|
| 818 |
+
"rewards/margins": 1.5703125,
|
| 819 |
+
"rewards/rejected": -0.828125,
|
| 820 |
+
"step": 520
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 1.6932907348242812,
|
| 824 |
+
"grad_norm": 17.196203110672243,
|
| 825 |
+
"learning_rate": 2.420118343195266e-07,
|
| 826 |
+
"logits/chosen": -1.625,
|
| 827 |
+
"logits/rejected": -1.6328125,
|
| 828 |
+
"logps/chosen": -240.0,
|
| 829 |
+
"logps/rejected": -292.0,
|
| 830 |
+
"loss": 0.3623,
|
| 831 |
+
"rewards/accuracies": 0.8125,
|
| 832 |
+
"rewards/chosen": 0.6796875,
|
| 833 |
+
"rewards/margins": 1.5234375,
|
| 834 |
+
"rewards/rejected": -0.84375,
|
| 835 |
+
"step": 530
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 1.7252396166134185,
|
| 839 |
+
"grad_norm": 20.669442243399377,
|
| 840 |
+
"learning_rate": 2.36094674556213e-07,
|
| 841 |
+
"logits/chosen": -1.6484375,
|
| 842 |
+
"logits/rejected": -1.6015625,
|
| 843 |
+
"logps/chosen": -237.0,
|
| 844 |
+
"logps/rejected": -302.0,
|
| 845 |
+
"loss": 0.35,
|
| 846 |
+
"rewards/accuracies": 0.887499988079071,
|
| 847 |
+
"rewards/chosen": 0.8203125,
|
| 848 |
+
"rewards/margins": 1.765625,
|
| 849 |
+
"rewards/rejected": -0.94921875,
|
| 850 |
+
"step": 540
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 1.7571884984025559,
|
| 854 |
+
"grad_norm": 17.765027713074762,
|
| 855 |
+
"learning_rate": 2.301775147928994e-07,
|
| 856 |
+
"logits/chosen": -1.6015625,
|
| 857 |
+
"logits/rejected": -1.6640625,
|
| 858 |
+
"logps/chosen": -242.0,
|
| 859 |
+
"logps/rejected": -302.0,
|
| 860 |
+
"loss": 0.3214,
|
| 861 |
+
"rewards/accuracies": 0.887499988079071,
|
| 862 |
+
"rewards/chosen": 0.5859375,
|
| 863 |
+
"rewards/margins": 1.4921875,
|
| 864 |
+
"rewards/rejected": -0.90234375,
|
| 865 |
+
"step": 550
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.7891373801916934,
|
| 869 |
+
"grad_norm": 18.979661871107176,
|
| 870 |
+
"learning_rate": 2.242603550295858e-07,
|
| 871 |
+
"logits/chosen": -1.5859375,
|
| 872 |
+
"logits/rejected": -1.625,
|
| 873 |
+
"logps/chosen": -240.0,
|
| 874 |
+
"logps/rejected": -292.0,
|
| 875 |
+
"loss": 0.35,
|
| 876 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 877 |
+
"rewards/chosen": 0.51171875,
|
| 878 |
+
"rewards/margins": 1.3203125,
|
| 879 |
+
"rewards/rejected": -0.8125,
|
| 880 |
+
"step": 560
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 1.8210862619808306,
|
| 884 |
+
"grad_norm": 14.344211337956814,
|
| 885 |
+
"learning_rate": 2.1834319526627217e-07,
|
| 886 |
+
"logits/chosen": -1.5703125,
|
| 887 |
+
"logits/rejected": -1.640625,
|
| 888 |
+
"logps/chosen": -234.0,
|
| 889 |
+
"logps/rejected": -304.0,
|
| 890 |
+
"loss": 0.3197,
|
| 891 |
+
"rewards/accuracies": 0.862500011920929,
|
| 892 |
+
"rewards/chosen": 0.44921875,
|
| 893 |
+
"rewards/margins": 1.6328125,
|
| 894 |
+
"rewards/rejected": -1.1796875,
|
| 895 |
+
"step": 570
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.8530351437699681,
|
| 899 |
+
"grad_norm": 18.365794275487666,
|
| 900 |
+
"learning_rate": 2.1242603550295858e-07,
|
| 901 |
+
"logits/chosen": -1.5546875,
|
| 902 |
+
"logits/rejected": -1.640625,
|
| 903 |
+
"logps/chosen": -239.0,
|
| 904 |
+
"logps/rejected": -300.0,
|
| 905 |
+
"loss": 0.3314,
|
| 906 |
+
"rewards/accuracies": 0.824999988079071,
|
| 907 |
+
"rewards/chosen": 0.359375,
|
| 908 |
+
"rewards/margins": 1.4609375,
|
| 909 |
+
"rewards/rejected": -1.109375,
|
| 910 |
+
"step": 580
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 1.8849840255591053,
|
| 914 |
+
"grad_norm": 16.19312631132269,
|
| 915 |
+
"learning_rate": 2.0650887573964496e-07,
|
| 916 |
+
"logits/chosen": -1.5703125,
|
| 917 |
+
"logits/rejected": -1.578125,
|
| 918 |
+
"logps/chosen": -243.0,
|
| 919 |
+
"logps/rejected": -298.0,
|
| 920 |
+
"loss": 0.2839,
|
| 921 |
+
"rewards/accuracies": 0.862500011920929,
|
| 922 |
+
"rewards/chosen": 0.5625,
|
| 923 |
+
"rewards/margins": 1.625,
|
| 924 |
+
"rewards/rejected": -1.0625,
|
| 925 |
+
"step": 590
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.9169329073482428,
|
| 929 |
+
"grad_norm": 12.485140486001677,
|
| 930 |
+
"learning_rate": 2.0059171597633133e-07,
|
| 931 |
+
"logits/chosen": -1.6015625,
|
| 932 |
+
"logits/rejected": -1.6484375,
|
| 933 |
+
"logps/chosen": -235.0,
|
| 934 |
+
"logps/rejected": -308.0,
|
| 935 |
+
"loss": 0.2658,
|
| 936 |
+
"rewards/accuracies": 0.925000011920929,
|
| 937 |
+
"rewards/chosen": 0.57421875,
|
| 938 |
+
"rewards/margins": 1.8046875,
|
| 939 |
+
"rewards/rejected": -1.234375,
|
| 940 |
+
"step": 600
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 1.9488817891373802,
|
| 944 |
+
"grad_norm": 12.432747016689122,
|
| 945 |
+
"learning_rate": 1.9467455621301774e-07,
|
| 946 |
+
"logits/chosen": -1.625,
|
| 947 |
+
"logits/rejected": -1.671875,
|
| 948 |
+
"logps/chosen": -236.0,
|
| 949 |
+
"logps/rejected": -308.0,
|
| 950 |
+
"loss": 0.2782,
|
| 951 |
+
"rewards/accuracies": 0.925000011920929,
|
| 952 |
+
"rewards/chosen": 0.71875,
|
| 953 |
+
"rewards/margins": 1.875,
|
| 954 |
+
"rewards/rejected": -1.15625,
|
| 955 |
+
"step": 610
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.9808306709265175,
|
| 959 |
+
"grad_norm": 18.34996934176123,
|
| 960 |
+
"learning_rate": 1.8875739644970412e-07,
|
| 961 |
+
"logits/chosen": -1.6015625,
|
| 962 |
+
"logits/rejected": -1.640625,
|
| 963 |
+
"logps/chosen": -233.0,
|
| 964 |
+
"logps/rejected": -302.0,
|
| 965 |
+
"loss": 0.3063,
|
| 966 |
+
"rewards/accuracies": 0.887499988079071,
|
| 967 |
+
"rewards/chosen": 0.71484375,
|
| 968 |
+
"rewards/margins": 1.9140625,
|
| 969 |
+
"rewards/rejected": -1.203125,
|
| 970 |
+
"step": 620
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 2.0,
|
| 974 |
+
"eval_logits/chosen": -1.640625,
|
| 975 |
+
"eval_logits/rejected": -1.6484375,
|
| 976 |
+
"eval_logps/chosen": -239.0,
|
| 977 |
+
"eval_logps/rejected": -300.0,
|
| 978 |
+
"eval_loss": 0.4112890660762787,
|
| 979 |
+
"eval_rewards/accuracies": 0.8035714030265808,
|
| 980 |
+
"eval_rewards/chosen": 0.40234375,
|
| 981 |
+
"eval_rewards/margins": 1.265625,
|
| 982 |
+
"eval_rewards/rejected": -0.859375,
|
| 983 |
+
"eval_runtime": 12.3166,
|
| 984 |
+
"eval_samples_per_second": 16.238,
|
| 985 |
+
"eval_steps_per_second": 0.568,
|
| 986 |
+
"step": 626
|
| 987 |
}
|
| 988 |
],
|
| 989 |
"logging_steps": 10,
|