Commit
•
c02310c
1
Parent(s):
f17a38d
Model save
Browse files- README.md +32 -22
- adapter_model.safetensors +1 -1
- all_results.json +5 -5
- train_results.json +5 -5
- trainer_state.json +2030 -90
README.md
CHANGED
|
@@ -1,13 +1,11 @@
|
|
| 1 |
---
|
| 2 |
library_name: peft
|
| 3 |
tags:
|
| 4 |
-
- alignment-handbook
|
| 5 |
- trl
|
| 6 |
- dpo
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
base_model: NbAiLab/nb-gpt-j-6B-v2
|
| 9 |
-
datasets:
|
| 10 |
-
- hugodk-sch/aftonposten_title_prefs
|
| 11 |
model-index:
|
| 12 |
- name: aftonposten-6b-align-scan
|
| 13 |
results: []
|
|
@@ -18,17 +16,17 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
|
| 19 |
# aftonposten-6b-align-scan
|
| 20 |
|
| 21 |
-
This model is a fine-tuned version of [
|
| 22 |
It achieves the following results on the evaluation set:
|
| 23 |
-
- Loss: 0.
|
| 24 |
-
- Rewards/chosen: -0.
|
| 25 |
-
- Rewards/rejected: -0.
|
| 26 |
-
- Rewards/accuracies: 0.
|
| 27 |
-
- Rewards/margins: 0.
|
| 28 |
-
- Logps/rejected: -37.
|
| 29 |
-
- Logps/chosen: -34.
|
| 30 |
-
- Logits/rejected: -2.
|
| 31 |
-
- Logits/chosen: -2.
|
| 32 |
|
| 33 |
## Model description
|
| 34 |
|
|
@@ -57,21 +55,33 @@ The following hyperparameters were used during training:
|
|
| 57 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 58 |
- lr_scheduler_type: cosine
|
| 59 |
- lr_scheduler_warmup_ratio: 0.1
|
| 60 |
-
- num_epochs:
|
| 61 |
|
| 62 |
### Training results
|
| 63 |
|
| 64 |
-
| Training Loss | Epoch | Step |
|
| 65 |
-
|
| 66 |
-
| 0.8913 | 0.26 | 100 |
|
| 67 |
-
| 0.7293 | 0.52 | 200 |
|
| 68 |
-
| 0.6144 | 0.78 | 300 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 69 |
|
| 70 |
|
| 71 |
### Framework versions
|
| 72 |
|
| 73 |
-
- PEFT 0.
|
| 74 |
-
- Transformers 4.
|
| 75 |
- Pytorch 2.1.2+cu121
|
| 76 |
-
- Datasets 2.
|
| 77 |
- Tokenizers 0.15.1
|
|
|
|
| 1 |
---
|
| 2 |
library_name: peft
|
| 3 |
tags:
|
|
|
|
| 4 |
- trl
|
| 5 |
- dpo
|
| 6 |
+
- alignment-handbook
|
| 7 |
- generated_from_trainer
|
| 8 |
base_model: NbAiLab/nb-gpt-j-6B-v2
|
|
|
|
|
|
|
| 9 |
model-index:
|
| 10 |
- name: aftonposten-6b-align-scan
|
| 11 |
results: []
|
|
|
|
| 16 |
|
| 17 |
# aftonposten-6b-align-scan
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [NbAiLab/nb-gpt-j-6B-v2](https://huggingface.co/NbAiLab/nb-gpt-j-6B-v2) on the None dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.9176
|
| 22 |
+
- Rewards/chosen: -0.0906
|
| 23 |
+
- Rewards/rejected: -0.1873
|
| 24 |
+
- Rewards/accuracies: 0.5627
|
| 25 |
+
- Rewards/margins: 0.0967
|
| 26 |
+
- Logps/rejected: -37.7248
|
| 27 |
+
- Logps/chosen: -34.1353
|
| 28 |
+
- Logits/rejected: -2.2000
|
| 29 |
+
- Logits/chosen: -2.2049
|
| 30 |
|
| 31 |
## Model description
|
| 32 |
|
|
|
|
| 55 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 56 |
- lr_scheduler_type: cosine
|
| 57 |
- lr_scheduler_warmup_ratio: 0.1
|
| 58 |
+
- num_epochs: 4
|
| 59 |
|
| 60 |
### Training results
|
| 61 |
|
| 62 |
+
| Training Loss | Epoch | Step | Logits/chosen | Logits/rejected | Logps/chosen | Logps/rejected | Validation Loss | Rewards/accuracies | Rewards/chosen | Rewards/margins | Rewards/rejected |
|
| 63 |
+
|:-------------:|:-----:|:----:|:-------------:|:---------------:|:------------:|:--------------:|:---------------:|:------------------:|:--------------:|:---------------:|:----------------:|
|
| 64 |
+
| 0.8913 | 0.26 | 100 | -2.2322 | -2.2273 | -34.0407 | -37.5405 | 0.9845 | 0.5195 | -0.0055 | 0.0159 | -0.0215 |
|
| 65 |
+
| 0.7293 | 0.52 | 200 | -2.2286 | -2.2238 | -34.0537 | -37.5811 | 0.9602 | 0.5714 | -0.0172 | 0.0408 | -0.0580 |
|
| 66 |
+
| 0.6144 | 0.78 | 300 | -2.2249 | -2.2201 | -34.0866 | -37.6032 | 0.9713 | 0.5282 | -0.0468 | 0.0310 | -0.0779 |
|
| 67 |
+
| 0.3632 | 1.04 | 400 | 0.9495 | -0.0909 | -0.1434 | 0.5602 | 0.0525 | -37.6760 | -34.1355 | -2.2076 | -2.2125 |
|
| 68 |
+
| 0.2994 | 1.3 | 500 | 0.9461 | -0.1647 | -0.2318 | 0.5540 | 0.0671 | -37.7742 | -34.2176 | -2.2162 | -2.2210 |
|
| 69 |
+
| 0.3408 | 1.56 | 600 | 0.9077 | -0.0675 | -0.1694 | 0.5868 | 0.1019 | -37.7048 | -34.1096 | -2.2017 | -2.2066 |
|
| 70 |
+
| 0.2796 | 1.82 | 700 | 0.9425 | -0.0929 | -0.1626 | 0.5569 | 0.0697 | -37.6973 | -34.1378 | -2.2012 | -2.2061 |
|
| 71 |
+
| 0.1052 | 2.08 | 800 | 0.9125 | -0.0848 | -0.1863 | 0.5926 | 0.1015 | -37.7236 | -34.1288 | -2.2003 | -2.2051 |
|
| 72 |
+
| 0.095 | 2.34 | 900 | 0.9005 | -0.0802 | -0.1942 | 0.5540 | 0.1140 | -37.7324 | -34.1237 | -2.2019 | -2.2067 |
|
| 73 |
+
| 0.123 | 2.6 | 1000 | 0.9194 | -0.0907 | -0.1876 | 0.5511 | 0.0969 | -37.7251 | -34.1353 | -2.1994 | -2.2043 |
|
| 74 |
+
| 0.0894 | 2.86 | 1100 | 0.9182 | -0.0915 | -0.1890 | 0.5336 | 0.0976 | -37.7267 | -34.1362 | -2.2001 | -2.2050 |
|
| 75 |
+
| 0.1086 | 3.12 | 1200 | 0.9023 | -0.0864 | -0.1976 | 0.5627 | 0.1112 | -37.7362 | -34.1306 | -2.2006 | -2.2054 |
|
| 76 |
+
| 0.0577 | 3.38 | 1300 | 0.9154 | -0.0922 | -0.1935 | 0.5598 | 0.1013 | -37.7317 | -34.1370 | -2.2002 | -2.2050 |
|
| 77 |
+
| 0.0375 | 3.64 | 1400 | 0.9233 | -0.0896 | -0.1810 | 0.5569 | 0.0914 | -37.7178 | -34.1342 | -2.2002 | -2.2050 |
|
| 78 |
+
| 0.0724 | 3.9 | 1500 | 0.9176 | -0.0906 | -0.1873 | 0.5627 | 0.0967 | -37.7248 | -34.1353 | -2.2000 | -2.2049 |
|
| 79 |
|
| 80 |
|
| 81 |
### Framework versions
|
| 82 |
|
| 83 |
+
- PEFT 0.10.0
|
| 84 |
+
- Transformers 4.39.0.dev0
|
| 85 |
- Pytorch 2.1.2+cu121
|
| 86 |
+
- Datasets 2.14.6
|
| 87 |
- Tokenizers 0.15.1
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 176183216
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:91d910fca51cc8ac20bb694b76c7d445531ace7d2e5055536929a38bcbb377dd
|
| 3 |
size 176183216
|
all_results.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
"eval_logits/chosen": -2.223867654800415,
|
| 4 |
"eval_logits/rejected": -2.219041109085083,
|
| 5 |
"eval_logps/chosen": -34.07316970825195,
|
|
@@ -13,9 +13,9 @@
|
|
| 13 |
"eval_samples": 343,
|
| 14 |
"eval_samples_per_second": 2.354,
|
| 15 |
"eval_steps_per_second": 0.295,
|
| 16 |
-
"train_loss": 0.
|
| 17 |
-
"train_runtime":
|
| 18 |
"train_samples": 3079,
|
| 19 |
-
"train_samples_per_second":
|
| 20 |
-
"train_steps_per_second": 0.
|
| 21 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.0,
|
| 3 |
"eval_logits/chosen": -2.223867654800415,
|
| 4 |
"eval_logits/rejected": -2.219041109085083,
|
| 5 |
"eval_logps/chosen": -34.07316970825195,
|
|
|
|
| 13 |
"eval_samples": 343,
|
| 14 |
"eval_samples_per_second": 2.354,
|
| 15 |
"eval_steps_per_second": 0.295,
|
| 16 |
+
"train_loss": 0.15107146371881683,
|
| 17 |
+
"train_runtime": 10740.8737,
|
| 18 |
"train_samples": 3079,
|
| 19 |
+
"train_samples_per_second": 1.147,
|
| 20 |
+
"train_steps_per_second": 0.143
|
| 21 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 3079,
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.0,
|
| 3 |
+
"train_loss": 0.15107146371881683,
|
| 4 |
+
"train_runtime": 10740.8737,
|
| 5 |
"train_samples": 3079,
|
| 6 |
+
"train_samples_per_second": 1.147,
|
| 7 |
+
"train_steps_per_second": 0.143
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -492,130 +492,2070 @@
|
|
| 492 |
},
|
| 493 |
{
|
| 494 |
"epoch": 0.81,
|
| 495 |
-
"
|
| 496 |
-
"
|
| 497 |
-
"logits/
|
| 498 |
-
"
|
| 499 |
-
"logps/
|
| 500 |
-
"
|
| 501 |
-
"
|
| 502 |
-
"rewards/
|
| 503 |
-
"rewards/
|
| 504 |
-
"rewards/
|
|
|
|
| 505 |
"step": 310
|
| 506 |
},
|
| 507 |
{
|
| 508 |
"epoch": 0.83,
|
| 509 |
-
"
|
| 510 |
-
"
|
| 511 |
-
"logits/
|
| 512 |
-
"
|
| 513 |
-
"logps/
|
| 514 |
-
"
|
| 515 |
-
"
|
| 516 |
-
"rewards/
|
| 517 |
-
"rewards/
|
| 518 |
-
"rewards/
|
|
|
|
| 519 |
"step": 320
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 0.86,
|
| 523 |
-
"
|
| 524 |
-
"
|
| 525 |
-
"logits/
|
| 526 |
-
"
|
| 527 |
-
"logps/
|
| 528 |
-
"
|
| 529 |
-
"
|
| 530 |
-
"rewards/
|
| 531 |
-
"rewards/
|
| 532 |
-
"rewards/
|
|
|
|
| 533 |
"step": 330
|
| 534 |
},
|
| 535 |
{
|
| 536 |
"epoch": 0.88,
|
| 537 |
-
"
|
| 538 |
-
"
|
| 539 |
-
"logits/
|
| 540 |
-
"
|
| 541 |
-
"logps/
|
| 542 |
-
"
|
| 543 |
-
"
|
| 544 |
-
"rewards/
|
| 545 |
-
"rewards/
|
| 546 |
-
"rewards/
|
|
|
|
| 547 |
"step": 340
|
| 548 |
},
|
| 549 |
{
|
| 550 |
"epoch": 0.91,
|
| 551 |
-
"
|
| 552 |
-
"
|
| 553 |
-
"logits/
|
| 554 |
-
"
|
| 555 |
-
"logps/
|
| 556 |
-
"
|
| 557 |
-
"
|
| 558 |
-
"rewards/
|
| 559 |
-
"rewards/
|
| 560 |
-
"rewards/
|
|
|
|
| 561 |
"step": 350
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 0.94,
|
| 565 |
-
"
|
| 566 |
-
"
|
| 567 |
-
"logits/
|
| 568 |
-
"
|
| 569 |
-
"logps/
|
| 570 |
-
"
|
| 571 |
-
"
|
| 572 |
-
"rewards/
|
| 573 |
-
"rewards/
|
| 574 |
-
"rewards/
|
|
|
|
| 575 |
"step": 360
|
| 576 |
},
|
| 577 |
{
|
| 578 |
"epoch": 0.96,
|
| 579 |
-
"
|
| 580 |
-
"
|
| 581 |
-
"logits/
|
| 582 |
-
"
|
| 583 |
-
"logps/
|
| 584 |
-
"
|
| 585 |
-
"
|
| 586 |
-
"rewards/
|
| 587 |
-
"rewards/
|
| 588 |
-
"rewards/
|
|
|
|
| 589 |
"step": 370
|
| 590 |
},
|
| 591 |
{
|
| 592 |
"epoch": 0.99,
|
| 593 |
-
"
|
| 594 |
-
"
|
| 595 |
-
"logits/
|
| 596 |
-
"
|
| 597 |
-
"logps/
|
| 598 |
-
"
|
| 599 |
-
"
|
| 600 |
-
"rewards/
|
| 601 |
-
"rewards/
|
| 602 |
-
"rewards/
|
|
|
|
| 603 |
"step": 380
|
| 604 |
},
|
| 605 |
{
|
| 606 |
-
"epoch": 1.
|
| 607 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 608 |
"total_flos": 0.0,
|
| 609 |
-
"train_loss": 0.
|
| 610 |
-
"train_runtime":
|
| 611 |
-
"train_samples_per_second":
|
| 612 |
-
"train_steps_per_second": 0.
|
| 613 |
}
|
| 614 |
],
|
| 615 |
"logging_steps": 10,
|
| 616 |
-
"max_steps":
|
| 617 |
"num_input_tokens_seen": 0,
|
| 618 |
-
"num_train_epochs":
|
| 619 |
"save_steps": 100,
|
| 620 |
"total_flos": 0.0,
|
| 621 |
"train_batch_size": 4,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 4.0,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 1540,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 492 |
},
|
| 493 |
{
|
| 494 |
"epoch": 0.81,
|
| 495 |
+
"grad_norm": 24.75,
|
| 496 |
+
"learning_rate": 4.84533120650964e-06,
|
| 497 |
+
"logits/chosen": -2.057966709136963,
|
| 498 |
+
"logits/rejected": -2.045175790786743,
|
| 499 |
+
"logps/chosen": -32.121585845947266,
|
| 500 |
+
"logps/rejected": -32.866825103759766,
|
| 501 |
+
"loss": 0.5062,
|
| 502 |
+
"rewards/accuracies": 0.875,
|
| 503 |
+
"rewards/chosen": 0.3661278784275055,
|
| 504 |
+
"rewards/margins": 0.5693171620368958,
|
| 505 |
+
"rewards/rejected": -0.20318928360939026,
|
| 506 |
"step": 310
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.83,
|
| 510 |
+
"grad_norm": 23.875,
|
| 511 |
+
"learning_rate": 4.825108134172131e-06,
|
| 512 |
+
"logits/chosen": -1.9696298837661743,
|
| 513 |
+
"logits/rejected": -1.9610140323638916,
|
| 514 |
+
"logps/chosen": -31.775466918945312,
|
| 515 |
+
"logps/rejected": -30.438364028930664,
|
| 516 |
+
"loss": 0.4574,
|
| 517 |
+
"rewards/accuracies": 0.862500011920929,
|
| 518 |
+
"rewards/chosen": 0.5215664505958557,
|
| 519 |
+
"rewards/margins": 0.6988391876220703,
|
| 520 |
+
"rewards/rejected": -0.1772727519273758,
|
| 521 |
"step": 320
|
| 522 |
},
|
| 523 |
{
|
| 524 |
"epoch": 0.86,
|
| 525 |
+
"grad_norm": 30.5,
|
| 526 |
+
"learning_rate": 4.80369052967602e-06,
|
| 527 |
+
"logits/chosen": -1.9035661220550537,
|
| 528 |
+
"logits/rejected": -1.9156697988510132,
|
| 529 |
+
"logps/chosen": -29.8248291015625,
|
| 530 |
+
"logps/rejected": -33.69133758544922,
|
| 531 |
+
"loss": 0.3788,
|
| 532 |
+
"rewards/accuracies": 0.862500011920929,
|
| 533 |
+
"rewards/chosen": 0.5889965295791626,
|
| 534 |
+
"rewards/margins": 0.8491708040237427,
|
| 535 |
+
"rewards/rejected": -0.2601741552352905,
|
| 536 |
"step": 330
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 0.88,
|
| 540 |
+
"grad_norm": 47.0,
|
| 541 |
+
"learning_rate": 4.781089396387968e-06,
|
| 542 |
+
"logits/chosen": -1.8689849376678467,
|
| 543 |
+
"logits/rejected": -1.8597549200057983,
|
| 544 |
+
"logps/chosen": -34.021995544433594,
|
| 545 |
+
"logps/rejected": -36.21257400512695,
|
| 546 |
+
"loss": 0.3569,
|
| 547 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 548 |
+
"rewards/chosen": 0.6092386841773987,
|
| 549 |
+
"rewards/margins": 0.9295026659965515,
|
| 550 |
+
"rewards/rejected": -0.32026407122612,
|
| 551 |
"step": 340
|
| 552 |
},
|
| 553 |
{
|
| 554 |
"epoch": 0.91,
|
| 555 |
+
"grad_norm": 27.0,
|
| 556 |
+
"learning_rate": 4.757316345716554e-06,
|
| 557 |
+
"logits/chosen": -1.9230518341064453,
|
| 558 |
+
"logits/rejected": -1.923689842224121,
|
| 559 |
+
"logps/chosen": -33.71295166015625,
|
| 560 |
+
"logps/rejected": -34.07986068725586,
|
| 561 |
+
"loss": 0.4213,
|
| 562 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 563 |
+
"rewards/chosen": 0.6230990886688232,
|
| 564 |
+
"rewards/margins": 0.8308955430984497,
|
| 565 |
+
"rewards/rejected": -0.20779642462730408,
|
| 566 |
"step": 350
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 0.94,
|
| 570 |
+
"grad_norm": 28.625,
|
| 571 |
+
"learning_rate": 4.73238359114687e-06,
|
| 572 |
+
"logits/chosen": -2.050320863723755,
|
| 573 |
+
"logits/rejected": -2.056476593017578,
|
| 574 |
+
"logps/chosen": -31.107620239257812,
|
| 575 |
+
"logps/rejected": -32.95317840576172,
|
| 576 |
+
"loss": 0.4902,
|
| 577 |
+
"rewards/accuracies": 0.862500011920929,
|
| 578 |
+
"rewards/chosen": 0.41906651854515076,
|
| 579 |
+
"rewards/margins": 0.6754721403121948,
|
| 580 |
+
"rewards/rejected": -0.25640562176704407,
|
| 581 |
"step": 360
|
| 582 |
},
|
| 583 |
{
|
| 584 |
"epoch": 0.96,
|
| 585 |
+
"grad_norm": 41.75,
|
| 586 |
+
"learning_rate": 4.706303941965804e-06,
|
| 587 |
+
"logits/chosen": -1.9822814464569092,
|
| 588 |
+
"logits/rejected": -1.9818360805511475,
|
| 589 |
+
"logps/chosen": -32.8140869140625,
|
| 590 |
+
"logps/rejected": -36.2825813293457,
|
| 591 |
+
"loss": 0.4132,
|
| 592 |
+
"rewards/accuracies": 0.862500011920929,
|
| 593 |
+
"rewards/chosen": 0.5677691102027893,
|
| 594 |
+
"rewards/margins": 0.7772952318191528,
|
| 595 |
+
"rewards/rejected": -0.2095261812210083,
|
| 596 |
"step": 370
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 0.99,
|
| 600 |
+
"grad_norm": 31.125,
|
| 601 |
+
"learning_rate": 4.679090796681225e-06,
|
| 602 |
+
"logits/chosen": -2.0170412063598633,
|
| 603 |
+
"logits/rejected": -2.012453556060791,
|
| 604 |
+
"logps/chosen": -30.147674560546875,
|
| 605 |
+
"logps/rejected": -29.48748207092285,
|
| 606 |
+
"loss": 0.4139,
|
| 607 |
+
"rewards/accuracies": 0.887499988079071,
|
| 608 |
+
"rewards/chosen": 0.506630539894104,
|
| 609 |
+
"rewards/margins": 0.7528629302978516,
|
| 610 |
+
"rewards/rejected": -0.24623243510723114,
|
| 611 |
"step": 380
|
| 612 |
},
|
| 613 |
{
|
| 614 |
+
"epoch": 1.01,
|
| 615 |
+
"grad_norm": 31.75,
|
| 616 |
+
"learning_rate": 4.650758136138454e-06,
|
| 617 |
+
"logits/chosen": -1.7892462015151978,
|
| 618 |
+
"logits/rejected": -1.7956597805023193,
|
| 619 |
+
"logps/chosen": -31.721576690673828,
|
| 620 |
+
"logps/rejected": -36.598609924316406,
|
| 621 |
+
"loss": 0.3273,
|
| 622 |
+
"rewards/accuracies": 0.875,
|
| 623 |
+
"rewards/chosen": 0.5977230668067932,
|
| 624 |
+
"rewards/margins": 1.0300120115280151,
|
| 625 |
+
"rewards/rejected": -0.4322889745235443,
|
| 626 |
+
"step": 390
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.04,
|
| 630 |
+
"grad_norm": 22.625,
|
| 631 |
+
"learning_rate": 4.621320516337559e-06,
|
| 632 |
+
"logits/chosen": -1.9460976123809814,
|
| 633 |
+
"logits/rejected": -1.9397977590560913,
|
| 634 |
+
"logps/chosen": -33.06171417236328,
|
| 635 |
+
"logps/rejected": -32.60498809814453,
|
| 636 |
+
"loss": 0.3632,
|
| 637 |
+
"rewards/accuracies": 0.824999988079071,
|
| 638 |
+
"rewards/chosen": 0.6938852667808533,
|
| 639 |
+
"rewards/margins": 0.9603848457336426,
|
| 640 |
+
"rewards/rejected": -0.2664996087551117,
|
| 641 |
+
"step": 400
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 1.04,
|
| 645 |
+
"eval_logits/chosen": -2.212498426437378,
|
| 646 |
+
"eval_logits/rejected": -2.207648992538452,
|
| 647 |
+
"eval_logps/chosen": -34.13554000854492,
|
| 648 |
+
"eval_logps/rejected": -37.67595291137695,
|
| 649 |
+
"eval_loss": 0.9494805932044983,
|
| 650 |
+
"eval_rewards/accuracies": 0.560215950012207,
|
| 651 |
+
"eval_rewards/chosen": -0.09088955074548721,
|
| 652 |
+
"eval_rewards/margins": 0.05251322314143181,
|
| 653 |
+
"eval_rewards/rejected": -0.14340277016162872,
|
| 654 |
+
"eval_runtime": 145.4418,
|
| 655 |
+
"eval_samples_per_second": 2.358,
|
| 656 |
+
"eval_steps_per_second": 0.296,
|
| 657 |
+
"step": 400
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.06,
|
| 661 |
+
"grad_norm": 29.25,
|
| 662 |
+
"learning_rate": 4.590793060955158e-06,
|
| 663 |
+
"logits/chosen": -1.9469482898712158,
|
| 664 |
+
"logits/rejected": -1.9542169570922852,
|
| 665 |
+
"logps/chosen": -28.586200714111328,
|
| 666 |
+
"logps/rejected": -29.615833282470703,
|
| 667 |
+
"loss": 0.3369,
|
| 668 |
+
"rewards/accuracies": 0.949999988079071,
|
| 669 |
+
"rewards/chosen": 0.4807574152946472,
|
| 670 |
+
"rewards/margins": 0.9466059803962708,
|
| 671 |
+
"rewards/rejected": -0.4658486247062683,
|
| 672 |
+
"step": 410
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.09,
|
| 676 |
+
"grad_norm": 29.625,
|
| 677 |
+
"learning_rate": 4.559191453574582e-06,
|
| 678 |
+
"logits/chosen": -1.967563271522522,
|
| 679 |
+
"logits/rejected": -1.9665197134017944,
|
| 680 |
+
"logps/chosen": -33.637752532958984,
|
| 681 |
+
"logps/rejected": -31.163837432861328,
|
| 682 |
+
"loss": 0.4787,
|
| 683 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 684 |
+
"rewards/chosen": 0.551658034324646,
|
| 685 |
+
"rewards/margins": 0.7503688335418701,
|
| 686 |
+
"rewards/rejected": -0.19871078431606293,
|
| 687 |
+
"step": 420
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 1.12,
|
| 691 |
+
"grad_norm": 26.125,
|
| 692 |
+
"learning_rate": 4.52653192962838e-06,
|
| 693 |
+
"logits/chosen": -1.9623167514801025,
|
| 694 |
+
"logits/rejected": -1.9450109004974365,
|
| 695 |
+
"logps/chosen": -30.47269630432129,
|
| 696 |
+
"logps/rejected": -33.48539733886719,
|
| 697 |
+
"loss": 0.3673,
|
| 698 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 699 |
+
"rewards/chosen": 0.4540596604347229,
|
| 700 |
+
"rewards/margins": 0.9572601318359375,
|
| 701 |
+
"rewards/rejected": -0.5032004117965698,
|
| 702 |
+
"step": 430
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 1.14,
|
| 706 |
+
"grad_norm": 18.125,
|
| 707 |
+
"learning_rate": 4.492831268057307e-06,
|
| 708 |
+
"logits/chosen": -1.9934860467910767,
|
| 709 |
+
"logits/rejected": -1.9955333471298218,
|
| 710 |
+
"logps/chosen": -35.85193634033203,
|
| 711 |
+
"logps/rejected": -35.39922332763672,
|
| 712 |
+
"loss": 0.2297,
|
| 713 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 714 |
+
"rewards/chosen": 0.6843839883804321,
|
| 715 |
+
"rewards/margins": 1.1697149276733398,
|
| 716 |
+
"rewards/rejected": -0.4853309988975525,
|
| 717 |
+
"step": 440
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 1.17,
|
| 721 |
+
"grad_norm": 26.625,
|
| 722 |
+
"learning_rate": 4.458106782690094e-06,
|
| 723 |
+
"logits/chosen": -2.072458267211914,
|
| 724 |
+
"logits/rejected": -2.0722086429595947,
|
| 725 |
+
"logps/chosen": -31.9611759185791,
|
| 726 |
+
"logps/rejected": -33.9387321472168,
|
| 727 |
+
"loss": 0.3876,
|
| 728 |
+
"rewards/accuracies": 0.887499988079071,
|
| 729 |
+
"rewards/chosen": 0.6049578189849854,
|
| 730 |
+
"rewards/margins": 0.9457138776779175,
|
| 731 |
+
"rewards/rejected": -0.3407560884952545,
|
| 732 |
+
"step": 450
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 1.19,
|
| 736 |
+
"grad_norm": 30.625,
|
| 737 |
+
"learning_rate": 4.422376313348405e-06,
|
| 738 |
+
"logits/chosen": -2.0193634033203125,
|
| 739 |
+
"logits/rejected": -2.011868715286255,
|
| 740 |
+
"logps/chosen": -31.689464569091797,
|
| 741 |
+
"logps/rejected": -36.604766845703125,
|
| 742 |
+
"loss": 0.2964,
|
| 743 |
+
"rewards/accuracies": 0.875,
|
| 744 |
+
"rewards/chosen": 0.6448441743850708,
|
| 745 |
+
"rewards/margins": 1.140498399734497,
|
| 746 |
+
"rewards/rejected": -0.4956541061401367,
|
| 747 |
+
"step": 460
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 1.22,
|
| 751 |
+
"grad_norm": 28.75,
|
| 752 |
+
"learning_rate": 4.3856582166815696e-06,
|
| 753 |
+
"logits/chosen": -1.9231773614883423,
|
| 754 |
+
"logits/rejected": -1.9196126461029053,
|
| 755 |
+
"logps/chosen": -33.275428771972656,
|
| 756 |
+
"logps/rejected": -33.32958221435547,
|
| 757 |
+
"loss": 0.3447,
|
| 758 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 759 |
+
"rewards/chosen": 0.5911833047866821,
|
| 760 |
+
"rewards/margins": 0.9563148617744446,
|
| 761 |
+
"rewards/rejected": -0.36513155698776245,
|
| 762 |
+
"step": 470
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 1.25,
|
| 766 |
+
"grad_norm": 31.0,
|
| 767 |
+
"learning_rate": 4.347971356735789e-06,
|
| 768 |
+
"logits/chosen": -2.050720453262329,
|
| 769 |
+
"logits/rejected": -2.0436971187591553,
|
| 770 |
+
"logps/chosen": -30.2794132232666,
|
| 771 |
+
"logps/rejected": -32.480506896972656,
|
| 772 |
+
"loss": 0.3824,
|
| 773 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 774 |
+
"rewards/chosen": 0.49421030282974243,
|
| 775 |
+
"rewards/margins": 0.9325267672538757,
|
| 776 |
+
"rewards/rejected": -0.4383164346218109,
|
| 777 |
+
"step": 480
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 1.27,
|
| 781 |
+
"grad_norm": 14.3125,
|
| 782 |
+
"learning_rate": 4.309335095262675e-06,
|
| 783 |
+
"logits/chosen": -1.996100664138794,
|
| 784 |
+
"logits/rejected": -1.997636079788208,
|
| 785 |
+
"logps/chosen": -34.52549743652344,
|
| 786 |
+
"logps/rejected": -34.070865631103516,
|
| 787 |
+
"loss": 0.2553,
|
| 788 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 789 |
+
"rewards/chosen": 0.7449228763580322,
|
| 790 |
+
"rewards/margins": 1.1313400268554688,
|
| 791 |
+
"rewards/rejected": -0.38641709089279175,
|
| 792 |
+
"step": 490
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 1.3,
|
| 796 |
+
"grad_norm": 36.25,
|
| 797 |
+
"learning_rate": 4.269769281772082e-06,
|
| 798 |
+
"logits/chosen": -1.8836076259613037,
|
| 799 |
+
"logits/rejected": -1.8812100887298584,
|
| 800 |
+
"logps/chosen": -32.4937629699707,
|
| 801 |
+
"logps/rejected": -37.06315231323242,
|
| 802 |
+
"loss": 0.2994,
|
| 803 |
+
"rewards/accuracies": 0.887499988079071,
|
| 804 |
+
"rewards/chosen": 0.6106754541397095,
|
| 805 |
+
"rewards/margins": 1.149621605873108,
|
| 806 |
+
"rewards/rejected": -0.538946270942688,
|
| 807 |
+
"step": 500
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 1.3,
|
| 811 |
+
"eval_logits/chosen": -2.221041202545166,
|
| 812 |
+
"eval_logits/rejected": -2.2161660194396973,
|
| 813 |
+
"eval_logps/chosen": -34.217567443847656,
|
| 814 |
+
"eval_logps/rejected": -37.774173736572266,
|
| 815 |
+
"eval_loss": 0.9461130499839783,
|
| 816 |
+
"eval_rewards/accuracies": 0.5539867281913757,
|
| 817 |
+
"eval_rewards/chosen": -0.16471394896507263,
|
| 818 |
+
"eval_rewards/margins": 0.0670846477150917,
|
| 819 |
+
"eval_rewards/rejected": -0.23179861903190613,
|
| 820 |
+
"eval_runtime": 145.1678,
|
| 821 |
+
"eval_samples_per_second": 2.363,
|
| 822 |
+
"eval_steps_per_second": 0.296,
|
| 823 |
+
"step": 500
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.32,
|
| 827 |
+
"grad_norm": 28.125,
|
| 828 |
+
"learning_rate": 4.22929424333435e-06,
|
| 829 |
+
"logits/chosen": -1.9868427515029907,
|
| 830 |
+
"logits/rejected": -1.9915640354156494,
|
| 831 |
+
"logps/chosen": -32.677276611328125,
|
| 832 |
+
"logps/rejected": -32.02418899536133,
|
| 833 |
+
"loss": 0.3257,
|
| 834 |
+
"rewards/accuracies": 0.875,
|
| 835 |
+
"rewards/chosen": 0.702068030834198,
|
| 836 |
+
"rewards/margins": 1.0423411130905151,
|
| 837 |
+
"rewards/rejected": -0.3402729630470276,
|
| 838 |
+
"step": 510
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.35,
|
| 842 |
+
"grad_norm": 31.25,
|
| 843 |
+
"learning_rate": 4.1879307741372085e-06,
|
| 844 |
+
"logits/chosen": -2.0190398693084717,
|
| 845 |
+
"logits/rejected": -2.0300133228302,
|
| 846 |
+
"logps/chosen": -30.644947052001953,
|
| 847 |
+
"logps/rejected": -32.037715911865234,
|
| 848 |
+
"loss": 0.2992,
|
| 849 |
+
"rewards/accuracies": 0.9375,
|
| 850 |
+
"rewards/chosen": 0.6709223389625549,
|
| 851 |
+
"rewards/margins": 1.147947907447815,
|
| 852 |
+
"rewards/rejected": -0.4770255982875824,
|
| 853 |
+
"step": 520
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.38,
|
| 857 |
+
"grad_norm": 16.625,
|
| 858 |
+
"learning_rate": 4.145700124802693e-06,
|
| 859 |
+
"logits/chosen": -1.9455944299697876,
|
| 860 |
+
"logits/rejected": -1.9422004222869873,
|
| 861 |
+
"logps/chosen": -31.76815414428711,
|
| 862 |
+
"logps/rejected": -32.878482818603516,
|
| 863 |
+
"loss": 0.3351,
|
| 864 |
+
"rewards/accuracies": 0.925000011920929,
|
| 865 |
+
"rewards/chosen": 0.5871203541755676,
|
| 866 |
+
"rewards/margins": 1.0512535572052002,
|
| 867 |
+
"rewards/rejected": -0.4641333222389221,
|
| 868 |
+
"step": 530
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 1.4,
|
| 872 |
+
"grad_norm": 23.875,
|
| 873 |
+
"learning_rate": 4.102623991469562e-06,
|
| 874 |
+
"logits/chosen": -1.8109652996063232,
|
| 875 |
+
"logits/rejected": -1.8201634883880615,
|
| 876 |
+
"logps/chosen": -31.86029624938965,
|
| 877 |
+
"logps/rejected": -32.3470458984375,
|
| 878 |
+
"loss": 0.3393,
|
| 879 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 880 |
+
"rewards/chosen": 0.6381845474243164,
|
| 881 |
+
"rewards/margins": 1.0801360607147217,
|
| 882 |
+
"rewards/rejected": -0.4419515132904053,
|
| 883 |
+
"step": 540
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.43,
|
| 887 |
+
"grad_norm": 27.25,
|
| 888 |
+
"learning_rate": 4.058724504646834e-06,
|
| 889 |
+
"logits/chosen": -1.9057979583740234,
|
| 890 |
+
"logits/rejected": -1.8994461297988892,
|
| 891 |
+
"logps/chosen": -32.93577575683594,
|
| 892 |
+
"logps/rejected": -31.36001968383789,
|
| 893 |
+
"loss": 0.3481,
|
| 894 |
+
"rewards/accuracies": 0.887499988079071,
|
| 895 |
+
"rewards/chosen": 0.6952788233757019,
|
| 896 |
+
"rewards/margins": 1.0270812511444092,
|
| 897 |
+
"rewards/rejected": -0.33180245757102966,
|
| 898 |
+
"step": 550
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 1.45,
|
| 902 |
+
"grad_norm": 36.5,
|
| 903 |
+
"learning_rate": 4.014024217844167e-06,
|
| 904 |
+
"logits/chosen": -1.9947702884674072,
|
| 905 |
+
"logits/rejected": -1.9926884174346924,
|
| 906 |
+
"logps/chosen": -33.76757049560547,
|
| 907 |
+
"logps/rejected": -31.845256805419922,
|
| 908 |
+
"loss": 0.3511,
|
| 909 |
+
"rewards/accuracies": 0.875,
|
| 910 |
+
"rewards/chosen": 0.7232670187950134,
|
| 911 |
+
"rewards/margins": 1.0363487005233765,
|
| 912 |
+
"rewards/rejected": -0.3130818009376526,
|
| 913 |
+
"step": 560
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.48,
|
| 917 |
+
"grad_norm": 24.875,
|
| 918 |
+
"learning_rate": 3.968546095984911e-06,
|
| 919 |
+
"logits/chosen": -1.8271764516830444,
|
| 920 |
+
"logits/rejected": -1.8250176906585693,
|
| 921 |
+
"logps/chosen": -32.03794479370117,
|
| 922 |
+
"logps/rejected": -31.38640785217285,
|
| 923 |
+
"loss": 0.4057,
|
| 924 |
+
"rewards/accuracies": 0.862500011920929,
|
| 925 |
+
"rewards/chosen": 0.6826205253601074,
|
| 926 |
+
"rewards/margins": 1.0283124446868896,
|
| 927 |
+
"rewards/rejected": -0.34569185972213745,
|
| 928 |
+
"step": 570
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 1.51,
|
| 932 |
+
"grad_norm": 18.625,
|
| 933 |
+
"learning_rate": 3.922313503607806e-06,
|
| 934 |
+
"logits/chosen": -1.9618475437164307,
|
| 935 |
+
"logits/rejected": -1.9585418701171875,
|
| 936 |
+
"logps/chosen": -30.302867889404297,
|
| 937 |
+
"logps/rejected": -35.241397857666016,
|
| 938 |
+
"loss": 0.2905,
|
| 939 |
+
"rewards/accuracies": 0.875,
|
| 940 |
+
"rewards/chosen": 0.7215973138809204,
|
| 941 |
+
"rewards/margins": 1.2753074169158936,
|
| 942 |
+
"rewards/rejected": -0.5537099242210388,
|
| 943 |
+
"step": 580
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.53,
|
| 947 |
+
"grad_norm": 22.0,
|
| 948 |
+
"learning_rate": 3.875350192863368e-06,
|
| 949 |
+
"logits/chosen": -1.902073860168457,
|
| 950 |
+
"logits/rejected": -1.9056593179702759,
|
| 951 |
+
"logps/chosen": -28.98307228088379,
|
| 952 |
+
"logps/rejected": -31.06064796447754,
|
| 953 |
+
"loss": 0.389,
|
| 954 |
+
"rewards/accuracies": 0.862500011920929,
|
| 955 |
+
"rewards/chosen": 0.56876540184021,
|
| 956 |
+
"rewards/margins": 0.9303674697875977,
|
| 957 |
+
"rewards/rejected": -0.36160212755203247,
|
| 958 |
+
"step": 590
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 1.56,
|
| 962 |
+
"grad_norm": 26.25,
|
| 963 |
+
"learning_rate": 3.8276802913111436e-06,
|
| 964 |
+
"logits/chosen": -1.9391911029815674,
|
| 965 |
+
"logits/rejected": -1.939034104347229,
|
| 966 |
+
"logps/chosen": -31.3705997467041,
|
| 967 |
+
"logps/rejected": -31.58413314819336,
|
| 968 |
+
"loss": 0.3408,
|
| 969 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 970 |
+
"rewards/chosen": 0.7658407092094421,
|
| 971 |
+
"rewards/margins": 1.1178474426269531,
|
| 972 |
+
"rewards/rejected": -0.352006733417511,
|
| 973 |
+
"step": 600
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.56,
|
| 977 |
+
"eval_logits/chosen": -2.206589460372925,
|
| 978 |
+
"eval_logits/rejected": -2.201718807220459,
|
| 979 |
+
"eval_logps/chosen": -34.10955047607422,
|
| 980 |
+
"eval_logps/rejected": -37.704795837402344,
|
| 981 |
+
"eval_loss": 0.9076833724975586,
|
| 982 |
+
"eval_rewards/accuracies": 0.5867940187454224,
|
| 983 |
+
"eval_rewards/chosen": -0.06749669462442398,
|
| 984 |
+
"eval_rewards/margins": 0.10186038166284561,
|
| 985 |
+
"eval_rewards/rejected": -0.1693570613861084,
|
| 986 |
+
"eval_runtime": 145.2918,
|
| 987 |
+
"eval_samples_per_second": 2.361,
|
| 988 |
+
"eval_steps_per_second": 0.296,
|
| 989 |
+
"step": 600
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.58,
|
| 993 |
+
"grad_norm": 29.125,
|
| 994 |
+
"learning_rate": 3.7793282895240927e-06,
|
| 995 |
+
"logits/chosen": -2.0020594596862793,
|
| 996 |
+
"logits/rejected": -2.0028228759765625,
|
| 997 |
+
"logps/chosen": -34.12175750732422,
|
| 998 |
+
"logps/rejected": -33.66775894165039,
|
| 999 |
+
"loss": 0.2586,
|
| 1000 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1001 |
+
"rewards/chosen": 0.8175474405288696,
|
| 1002 |
+
"rewards/margins": 1.3427560329437256,
|
| 1003 |
+
"rewards/rejected": -0.5252088308334351,
|
| 1004 |
+
"step": 610
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 1.61,
|
| 1008 |
+
"grad_norm": 22.875,
|
| 1009 |
+
"learning_rate": 3.730319028506478e-06,
|
| 1010 |
+
"logits/chosen": -1.964914321899414,
|
| 1011 |
+
"logits/rejected": -1.962311029434204,
|
| 1012 |
+
"logps/chosen": -32.368412017822266,
|
| 1013 |
+
"logps/rejected": -32.58821487426758,
|
| 1014 |
+
"loss": 0.2969,
|
| 1015 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1016 |
+
"rewards/chosen": 0.6812867522239685,
|
| 1017 |
+
"rewards/margins": 1.1477558612823486,
|
| 1018 |
+
"rewards/rejected": -0.4664689898490906,
|
| 1019 |
+
"step": 620
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.64,
|
| 1023 |
+
"grad_norm": 41.75,
|
| 1024 |
+
"learning_rate": 3.6806776869317074e-06,
|
| 1025 |
+
"logits/chosen": -1.978055715560913,
|
| 1026 |
+
"logits/rejected": -1.9689325094223022,
|
| 1027 |
+
"logps/chosen": -31.888168334960938,
|
| 1028 |
+
"logps/rejected": -31.396778106689453,
|
| 1029 |
+
"loss": 0.3729,
|
| 1030 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 1031 |
+
"rewards/chosen": 0.7991548180580139,
|
| 1032 |
+
"rewards/margins": 1.214425802230835,
|
| 1033 |
+
"rewards/rejected": -0.4152711033821106,
|
| 1034 |
+
"step": 630
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 1.66,
|
| 1038 |
+
"grad_norm": 38.0,
|
| 1039 |
+
"learning_rate": 3.6304297682067146e-06,
|
| 1040 |
+
"logits/chosen": -1.9791650772094727,
|
| 1041 |
+
"logits/rejected": -1.9759613275527954,
|
| 1042 |
+
"logps/chosen": -31.385793685913086,
|
| 1043 |
+
"logps/rejected": -32.911964416503906,
|
| 1044 |
+
"loss": 0.2822,
|
| 1045 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1046 |
+
"rewards/chosen": 0.7313672304153442,
|
| 1047 |
+
"rewards/margins": 1.1746617555618286,
|
| 1048 |
+
"rewards/rejected": -0.44329458475112915,
|
| 1049 |
+
"step": 640
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 1.69,
|
| 1053 |
+
"grad_norm": 23.875,
|
| 1054 |
+
"learning_rate": 3.579601087369492e-06,
|
| 1055 |
+
"logits/chosen": -1.9867624044418335,
|
| 1056 |
+
"logits/rejected": -1.9891353845596313,
|
| 1057 |
+
"logps/chosen": -32.82892608642578,
|
| 1058 |
+
"logps/rejected": -34.27159881591797,
|
| 1059 |
+
"loss": 0.2431,
|
| 1060 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1061 |
+
"rewards/chosen": 0.7423025369644165,
|
| 1062 |
+
"rewards/margins": 1.2716872692108154,
|
| 1063 |
+
"rewards/rejected": -0.5293846726417542,
|
| 1064 |
+
"step": 650
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 1.71,
|
| 1068 |
+
"grad_norm": 37.75,
|
| 1069 |
+
"learning_rate": 3.5282177578265295e-06,
|
| 1070 |
+
"logits/chosen": -1.899405837059021,
|
| 1071 |
+
"logits/rejected": -1.899852991104126,
|
| 1072 |
+
"logps/chosen": -33.07904815673828,
|
| 1073 |
+
"logps/rejected": -32.00278091430664,
|
| 1074 |
+
"loss": 0.3078,
|
| 1075 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1076 |
+
"rewards/chosen": 0.7810029983520508,
|
| 1077 |
+
"rewards/margins": 1.136976957321167,
|
| 1078 |
+
"rewards/rejected": -0.3559740483760834,
|
| 1079 |
+
"step": 660
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 1.74,
|
| 1083 |
+
"grad_norm": 18.875,
|
| 1084 |
+
"learning_rate": 3.476306177936961e-06,
|
| 1085 |
+
"logits/chosen": -1.941097617149353,
|
| 1086 |
+
"logits/rejected": -1.9313017129898071,
|
| 1087 |
+
"logps/chosen": -33.12556838989258,
|
| 1088 |
+
"logps/rejected": -32.835227966308594,
|
| 1089 |
+
"loss": 0.1901,
|
| 1090 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1091 |
+
"rewards/chosen": 0.788142740726471,
|
| 1092 |
+
"rewards/margins": 1.2628332376480103,
|
| 1093 |
+
"rewards/rejected": -0.4746905267238617,
|
| 1094 |
+
"step": 670
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 1.77,
|
| 1098 |
+
"grad_norm": 34.0,
|
| 1099 |
+
"learning_rate": 3.423893017450324e-06,
|
| 1100 |
+
"logits/chosen": -1.8415699005126953,
|
| 1101 |
+
"logits/rejected": -1.838326096534729,
|
| 1102 |
+
"logps/chosen": -30.559249877929688,
|
| 1103 |
+
"logps/rejected": -34.762123107910156,
|
| 1104 |
+
"loss": 0.2502,
|
| 1105 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1106 |
+
"rewards/chosen": 0.8115372657775879,
|
| 1107 |
+
"rewards/margins": 1.2107248306274414,
|
| 1108 |
+
"rewards/rejected": -0.3991875946521759,
|
| 1109 |
+
"step": 680
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.79,
|
| 1113 |
+
"grad_norm": 17.75,
|
| 1114 |
+
"learning_rate": 3.3710052038048794e-06,
|
| 1115 |
+
"logits/chosen": -1.902061104774475,
|
| 1116 |
+
"logits/rejected": -1.9013097286224365,
|
| 1117 |
+
"logps/chosen": -34.259620666503906,
|
| 1118 |
+
"logps/rejected": -35.9239501953125,
|
| 1119 |
+
"loss": 0.183,
|
| 1120 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1121 |
+
"rewards/chosen": 0.9095993041992188,
|
| 1122 |
+
"rewards/margins": 1.42271089553833,
|
| 1123 |
+
"rewards/rejected": -0.5131114721298218,
|
| 1124 |
+
"step": 690
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 1.82,
|
| 1128 |
+
"grad_norm": 12.625,
|
| 1129 |
+
"learning_rate": 3.3176699082935546e-06,
|
| 1130 |
+
"logits/chosen": -1.8813689947128296,
|
| 1131 |
+
"logits/rejected": -1.884253740310669,
|
| 1132 |
+
"logps/chosen": -31.644283294677734,
|
| 1133 |
+
"logps/rejected": -36.16461181640625,
|
| 1134 |
+
"loss": 0.2796,
|
| 1135 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1136 |
+
"rewards/chosen": 0.8627067804336548,
|
| 1137 |
+
"rewards/margins": 1.3333566188812256,
|
| 1138 |
+
"rewards/rejected": -0.4706498980522156,
|
| 1139 |
+
"step": 700
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 1.82,
|
| 1143 |
+
"eval_logits/chosen": -2.206071615219116,
|
| 1144 |
+
"eval_logits/rejected": -2.2012383937835693,
|
| 1145 |
+
"eval_logps/chosen": -34.13777542114258,
|
| 1146 |
+
"eval_logps/rejected": -37.69728088378906,
|
| 1147 |
+
"eval_loss": 0.9425109624862671,
|
| 1148 |
+
"eval_rewards/accuracies": 0.5568937063217163,
|
| 1149 |
+
"eval_rewards/chosen": -0.0929015502333641,
|
| 1150 |
+
"eval_rewards/margins": 0.06969217956066132,
|
| 1151 |
+
"eval_rewards/rejected": -0.16259372234344482,
|
| 1152 |
+
"eval_runtime": 145.1331,
|
| 1153 |
+
"eval_samples_per_second": 2.363,
|
| 1154 |
+
"eval_steps_per_second": 0.296,
|
| 1155 |
+
"step": 700
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 1.84,
|
| 1159 |
+
"grad_norm": 28.75,
|
| 1160 |
+
"learning_rate": 3.2639145321045933e-06,
|
| 1161 |
+
"logits/chosen": -1.9908039569854736,
|
| 1162 |
+
"logits/rejected": -1.9935470819473267,
|
| 1163 |
+
"logps/chosen": -33.92770767211914,
|
| 1164 |
+
"logps/rejected": -34.66431427001953,
|
| 1165 |
+
"loss": 0.3611,
|
| 1166 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 1167 |
+
"rewards/chosen": 0.8710969090461731,
|
| 1168 |
+
"rewards/margins": 1.2258384227752686,
|
| 1169 |
+
"rewards/rejected": -0.3547416031360626,
|
| 1170 |
+
"step": 710
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 1.87,
|
| 1174 |
+
"grad_norm": 37.5,
|
| 1175 |
+
"learning_rate": 3.2097666922441107e-06,
|
| 1176 |
+
"logits/chosen": -1.845144271850586,
|
| 1177 |
+
"logits/rejected": -1.8392305374145508,
|
| 1178 |
+
"logps/chosen": -33.98247146606445,
|
| 1179 |
+
"logps/rejected": -32.777137756347656,
|
| 1180 |
+
"loss": 0.3525,
|
| 1181 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1182 |
+
"rewards/chosen": 0.7656866908073425,
|
| 1183 |
+
"rewards/margins": 1.1484500169754028,
|
| 1184 |
+
"rewards/rejected": -0.38276320695877075,
|
| 1185 |
+
"step": 720
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 1.9,
|
| 1189 |
+
"grad_norm": 18.375,
|
| 1190 |
+
"learning_rate": 3.1552542073477554e-06,
|
| 1191 |
+
"logits/chosen": -2.0191843509674072,
|
| 1192 |
+
"logits/rejected": -2.0161468982696533,
|
| 1193 |
+
"logps/chosen": -29.76511573791504,
|
| 1194 |
+
"logps/rejected": -32.01858901977539,
|
| 1195 |
+
"loss": 0.2082,
|
| 1196 |
+
"rewards/accuracies": 0.9375,
|
| 1197 |
+
"rewards/chosen": 0.7990841865539551,
|
| 1198 |
+
"rewards/margins": 1.3124363422393799,
|
| 1199 |
+
"rewards/rejected": -0.5133520364761353,
|
| 1200 |
+
"step": 730
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.92,
|
| 1204 |
+
"grad_norm": 34.0,
|
| 1205 |
+
"learning_rate": 3.100405083388799e-06,
|
| 1206 |
+
"logits/chosen": -1.8623883724212646,
|
| 1207 |
+
"logits/rejected": -1.8623119592666626,
|
| 1208 |
+
"logps/chosen": -32.44645309448242,
|
| 1209 |
+
"logps/rejected": -38.02064895629883,
|
| 1210 |
+
"loss": 0.2773,
|
| 1211 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1212 |
+
"rewards/chosen": 0.9150499105453491,
|
| 1213 |
+
"rewards/margins": 1.2686070203781128,
|
| 1214 |
+
"rewards/rejected": -0.3535570502281189,
|
| 1215 |
+
"step": 740
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 1.95,
|
| 1219 |
+
"grad_norm": 14.5,
|
| 1220 |
+
"learning_rate": 3.0452474992899645e-06,
|
| 1221 |
+
"logits/chosen": -1.7451330423355103,
|
| 1222 |
+
"logits/rejected": -1.7503840923309326,
|
| 1223 |
+
"logps/chosen": -35.91836166381836,
|
| 1224 |
+
"logps/rejected": -34.34960174560547,
|
| 1225 |
+
"loss": 0.3131,
|
| 1226 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1227 |
+
"rewards/chosen": 0.9126898646354675,
|
| 1228 |
+
"rewards/margins": 1.2522337436676025,
|
| 1229 |
+
"rewards/rejected": -0.339543879032135,
|
| 1230 |
+
"step": 750
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 1.97,
|
| 1234 |
+
"grad_norm": 33.5,
|
| 1235 |
+
"learning_rate": 2.989809792446417e-06,
|
| 1236 |
+
"logits/chosen": -1.9522911310195923,
|
| 1237 |
+
"logits/rejected": -1.9537473917007446,
|
| 1238 |
+
"logps/chosen": -31.657379150390625,
|
| 1239 |
+
"logps/rejected": -33.18722915649414,
|
| 1240 |
+
"loss": 0.2682,
|
| 1241 |
+
"rewards/accuracies": 0.9375,
|
| 1242 |
+
"rewards/chosen": 0.853469967842102,
|
| 1243 |
+
"rewards/margins": 1.2206826210021973,
|
| 1244 |
+
"rewards/rejected": -0.36721271276474,
|
| 1245 |
+
"step": 760
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 2.0,
|
| 1249 |
+
"grad_norm": 25.125,
|
| 1250 |
+
"learning_rate": 2.9341204441673267e-06,
|
| 1251 |
+
"logits/chosen": -1.9250198602676392,
|
| 1252 |
+
"logits/rejected": -1.9241039752960205,
|
| 1253 |
+
"logps/chosen": -31.101261138916016,
|
| 1254 |
+
"logps/rejected": -35.2042350769043,
|
| 1255 |
+
"loss": 0.2804,
|
| 1256 |
+
"rewards/accuracies": 0.8958333730697632,
|
| 1257 |
+
"rewards/chosen": 0.9669157266616821,
|
| 1258 |
+
"rewards/margins": 1.1253904104232788,
|
| 1259 |
+
"rewards/rejected": -0.1584748923778534,
|
| 1260 |
+
"step": 770
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 2.03,
|
| 1264 |
+
"grad_norm": 6.28125,
|
| 1265 |
+
"learning_rate": 2.878208065043501e-06,
|
| 1266 |
+
"logits/chosen": -1.9248275756835938,
|
| 1267 |
+
"logits/rejected": -1.9241101741790771,
|
| 1268 |
+
"logps/chosen": -33.638389587402344,
|
| 1269 |
+
"logps/rejected": -32.00913619995117,
|
| 1270 |
+
"loss": 0.1397,
|
| 1271 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1272 |
+
"rewards/chosen": 0.9090627431869507,
|
| 1273 |
+
"rewards/margins": 1.4397172927856445,
|
| 1274 |
+
"rewards/rejected": -0.5306544899940491,
|
| 1275 |
+
"step": 780
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 2.05,
|
| 1279 |
+
"grad_norm": 22.875,
|
| 1280 |
+
"learning_rate": 2.8221013802485974e-06,
|
| 1281 |
+
"logits/chosen": -1.9553604125976562,
|
| 1282 |
+
"logits/rejected": -1.9538936614990234,
|
| 1283 |
+
"logps/chosen": -28.537487030029297,
|
| 1284 |
+
"logps/rejected": -33.33250045776367,
|
| 1285 |
+
"loss": 0.084,
|
| 1286 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1287 |
+
"rewards/chosen": 0.9231246709823608,
|
| 1288 |
+
"rewards/margins": 1.5496938228607178,
|
| 1289 |
+
"rewards/rejected": -0.6265692114830017,
|
| 1290 |
+
"step": 790
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 2.08,
|
| 1294 |
+
"grad_norm": 0.0135498046875,
|
| 1295 |
+
"learning_rate": 2.76582921478147e-06,
|
| 1296 |
+
"logits/chosen": -2.0081448554992676,
|
| 1297 |
+
"logits/rejected": -2.0046160221099854,
|
| 1298 |
+
"logps/chosen": -31.324649810791016,
|
| 1299 |
+
"logps/rejected": -34.89632797241211,
|
| 1300 |
+
"loss": 0.1052,
|
| 1301 |
+
"rewards/accuracies": 0.9375,
|
| 1302 |
+
"rewards/chosen": 0.9923051595687866,
|
| 1303 |
+
"rewards/margins": 1.6198686361312866,
|
| 1304 |
+
"rewards/rejected": -0.6275635957717896,
|
| 1305 |
+
"step": 800
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 2.08,
|
| 1309 |
+
"eval_logits/chosen": -2.205134868621826,
|
| 1310 |
+
"eval_logits/rejected": -2.200315475463867,
|
| 1311 |
+
"eval_logps/chosen": -34.128814697265625,
|
| 1312 |
+
"eval_logps/rejected": -37.72360610961914,
|
| 1313 |
+
"eval_loss": 0.9124869704246521,
|
| 1314 |
+
"eval_rewards/accuracies": 0.5926079750061035,
|
| 1315 |
+
"eval_rewards/chosen": -0.08483708649873734,
|
| 1316 |
+
"eval_rewards/margins": 0.10145010054111481,
|
| 1317 |
+
"eval_rewards/rejected": -0.18628719449043274,
|
| 1318 |
+
"eval_runtime": 145.104,
|
| 1319 |
+
"eval_samples_per_second": 2.364,
|
| 1320 |
+
"eval_steps_per_second": 0.296,
|
| 1321 |
+
"step": 800
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 2.1,
|
| 1325 |
+
"grad_norm": 15.1875,
|
| 1326 |
+
"learning_rate": 2.7094204786572254e-06,
|
| 1327 |
+
"logits/chosen": -1.8404695987701416,
|
| 1328 |
+
"logits/rejected": -1.8326365947723389,
|
| 1329 |
+
"logps/chosen": -33.10143280029297,
|
| 1330 |
+
"logps/rejected": -35.159244537353516,
|
| 1331 |
+
"loss": 0.1258,
|
| 1332 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1333 |
+
"rewards/chosen": 1.2193273305892944,
|
| 1334 |
+
"rewards/margins": 1.7585445642471313,
|
| 1335 |
+
"rewards/rejected": -0.5392170548439026,
|
| 1336 |
+
"step": 810
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.13,
|
| 1340 |
+
"grad_norm": 21.125,
|
| 1341 |
+
"learning_rate": 2.6529041520546072e-06,
|
| 1342 |
+
"logits/chosen": -1.9172642230987549,
|
| 1343 |
+
"logits/rejected": -1.927819013595581,
|
| 1344 |
+
"logps/chosen": -34.42330551147461,
|
| 1345 |
+
"logps/rejected": -32.97254180908203,
|
| 1346 |
+
"loss": 0.1226,
|
| 1347 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1348 |
+
"rewards/chosen": 1.1451458930969238,
|
| 1349 |
+
"rewards/margins": 1.632136583328247,
|
| 1350 |
+
"rewards/rejected": -0.48699086904525757,
|
| 1351 |
+
"step": 820
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 2.16,
|
| 1355 |
+
"grad_norm": 28.5,
|
| 1356 |
+
"learning_rate": 2.5963092704273302e-06,
|
| 1357 |
+
"logits/chosen": -1.9668289422988892,
|
| 1358 |
+
"logits/rejected": -1.971239686012268,
|
| 1359 |
+
"logps/chosen": -33.753257751464844,
|
| 1360 |
+
"logps/rejected": -29.762826919555664,
|
| 1361 |
+
"loss": 0.1291,
|
| 1362 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1363 |
+
"rewards/chosen": 1.034569501876831,
|
| 1364 |
+
"rewards/margins": 1.479695439338684,
|
| 1365 |
+
"rewards/rejected": -0.44512590765953064,
|
| 1366 |
+
"step": 830
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 2.18,
|
| 1370 |
+
"grad_norm": 14.5,
|
| 1371 |
+
"learning_rate": 2.53966490958702e-06,
|
| 1372 |
+
"logits/chosen": -1.9671608209609985,
|
| 1373 |
+
"logits/rejected": -1.9750837087631226,
|
| 1374 |
+
"logps/chosen": -33.52558135986328,
|
| 1375 |
+
"logps/rejected": -30.574993133544922,
|
| 1376 |
+
"loss": 0.071,
|
| 1377 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1378 |
+
"rewards/chosen": 1.1362085342407227,
|
| 1379 |
+
"rewards/margins": 1.7907711267471313,
|
| 1380 |
+
"rewards/rejected": -0.6545625329017639,
|
| 1381 |
+
"step": 840
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 2.21,
|
| 1385 |
+
"grad_norm": 22.75,
|
| 1386 |
+
"learning_rate": 2.4830001707654135e-06,
|
| 1387 |
+
"logits/chosen": -1.8968629837036133,
|
| 1388 |
+
"logits/rejected": -1.8873090744018555,
|
| 1389 |
+
"logps/chosen": -30.793508529663086,
|
| 1390 |
+
"logps/rejected": -32.640350341796875,
|
| 1391 |
+
"loss": 0.1021,
|
| 1392 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1393 |
+
"rewards/chosen": 0.8997215032577515,
|
| 1394 |
+
"rewards/margins": 1.4636071920394897,
|
| 1395 |
+
"rewards/rejected": -0.5638857483863831,
|
| 1396 |
+
"step": 850
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 2.23,
|
| 1400 |
+
"grad_norm": 15.4375,
|
| 1401 |
+
"learning_rate": 2.4263441656635054e-06,
|
| 1402 |
+
"logits/chosen": -2.0392231941223145,
|
| 1403 |
+
"logits/rejected": -2.02933931350708,
|
| 1404 |
+
"logps/chosen": -25.14605140686035,
|
| 1405 |
+
"logps/rejected": -30.546062469482422,
|
| 1406 |
+
"loss": 0.1161,
|
| 1407 |
+
"rewards/accuracies": 0.9375,
|
| 1408 |
+
"rewards/chosen": 0.9702181816101074,
|
| 1409 |
+
"rewards/margins": 1.576568603515625,
|
| 1410 |
+
"rewards/rejected": -0.6063503623008728,
|
| 1411 |
+
"step": 860
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 2.26,
|
| 1415 |
+
"grad_norm": 19.375,
|
| 1416 |
+
"learning_rate": 2.3697260014953107e-06,
|
| 1417 |
+
"logits/chosen": -1.8909342288970947,
|
| 1418 |
+
"logits/rejected": -1.8921743631362915,
|
| 1419 |
+
"logps/chosen": -33.14449691772461,
|
| 1420 |
+
"logps/rejected": -30.5850772857666,
|
| 1421 |
+
"loss": 0.0889,
|
| 1422 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1423 |
+
"rewards/chosen": 1.0696533918380737,
|
| 1424 |
+
"rewards/margins": 1.61197030544281,
|
| 1425 |
+
"rewards/rejected": -0.5423170328140259,
|
| 1426 |
+
"step": 870
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 2.29,
|
| 1430 |
+
"grad_norm": 10.875,
|
| 1431 |
+
"learning_rate": 2.3131747660339396e-06,
|
| 1432 |
+
"logits/chosen": -1.9035065174102783,
|
| 1433 |
+
"logits/rejected": -1.9043560028076172,
|
| 1434 |
+
"logps/chosen": -31.356319427490234,
|
| 1435 |
+
"logps/rejected": -33.71599197387695,
|
| 1436 |
+
"loss": 0.2119,
|
| 1437 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1438 |
+
"rewards/chosen": 0.929639458656311,
|
| 1439 |
+
"rewards/margins": 1.5282022953033447,
|
| 1440 |
+
"rewards/rejected": -0.5985628962516785,
|
| 1441 |
+
"step": 880
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 2.31,
|
| 1445 |
+
"grad_norm": 28.75,
|
| 1446 |
+
"learning_rate": 2.256719512667651e-06,
|
| 1447 |
+
"logits/chosen": -1.8071092367172241,
|
| 1448 |
+
"logits/rejected": -1.8054319620132446,
|
| 1449 |
+
"logps/chosen": -34.50264358520508,
|
| 1450 |
+
"logps/rejected": -36.736839294433594,
|
| 1451 |
+
"loss": 0.091,
|
| 1452 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1453 |
+
"rewards/chosen": 1.0460432767868042,
|
| 1454 |
+
"rewards/margins": 1.8196884393692017,
|
| 1455 |
+
"rewards/rejected": -0.7736451029777527,
|
| 1456 |
+
"step": 890
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 2.34,
|
| 1460 |
+
"grad_norm": 11.75,
|
| 1461 |
+
"learning_rate": 2.2003892454735786e-06,
|
| 1462 |
+
"logits/chosen": -1.9551169872283936,
|
| 1463 |
+
"logits/rejected": -1.9482837915420532,
|
| 1464 |
+
"logps/chosen": -30.755849838256836,
|
| 1465 |
+
"logps/rejected": -33.490779876708984,
|
| 1466 |
+
"loss": 0.095,
|
| 1467 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1468 |
+
"rewards/chosen": 1.122107982635498,
|
| 1469 |
+
"rewards/margins": 1.781549096107483,
|
| 1470 |
+
"rewards/rejected": -0.6594408750534058,
|
| 1471 |
+
"step": 900
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 2.34,
|
| 1475 |
+
"eval_logits/chosen": -2.2067153453826904,
|
| 1476 |
+
"eval_logits/rejected": -2.2018842697143555,
|
| 1477 |
+
"eval_logps/chosen": -34.123661041259766,
|
| 1478 |
+
"eval_logps/rejected": -37.73240280151367,
|
| 1479 |
+
"eval_loss": 0.900454580783844,
|
| 1480 |
+
"eval_rewards/accuracies": 0.5539867281913757,
|
| 1481 |
+
"eval_rewards/chosen": -0.08019853383302689,
|
| 1482 |
+
"eval_rewards/margins": 0.11400878429412842,
|
| 1483 |
+
"eval_rewards/rejected": -0.19420728087425232,
|
| 1484 |
+
"eval_runtime": 145.2038,
|
| 1485 |
+
"eval_samples_per_second": 2.362,
|
| 1486 |
+
"eval_steps_per_second": 0.296,
|
| 1487 |
+
"step": 900
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 2.36,
|
| 1491 |
+
"grad_norm": 25.625,
|
| 1492 |
+
"learning_rate": 2.1442129043167877e-06,
|
| 1493 |
+
"logits/chosen": -1.9985930919647217,
|
| 1494 |
+
"logits/rejected": -1.9938539266586304,
|
| 1495 |
+
"logps/chosen": -32.36955261230469,
|
| 1496 |
+
"logps/rejected": -35.18764877319336,
|
| 1497 |
+
"loss": 0.091,
|
| 1498 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1499 |
+
"rewards/chosen": 1.0693137645721436,
|
| 1500 |
+
"rewards/margins": 1.6624386310577393,
|
| 1501 |
+
"rewards/rejected": -0.5931245684623718,
|
| 1502 |
+
"step": 910
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 2.39,
|
| 1506 |
+
"grad_norm": 9.6875,
|
| 1507 |
+
"learning_rate": 2.088219349982323e-06,
|
| 1508 |
+
"logits/chosen": -1.9170243740081787,
|
| 1509 |
+
"logits/rejected": -1.9220654964447021,
|
| 1510 |
+
"logps/chosen": -34.132789611816406,
|
| 1511 |
+
"logps/rejected": -33.121482849121094,
|
| 1512 |
+
"loss": 0.1312,
|
| 1513 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1514 |
+
"rewards/chosen": 1.0942142009735107,
|
| 1515 |
+
"rewards/margins": 1.6536308526992798,
|
| 1516 |
+
"rewards/rejected": -0.5594164133071899,
|
| 1517 |
+
"step": 920
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 2.42,
|
| 1521 |
+
"grad_norm": 16.75,
|
| 1522 |
+
"learning_rate": 2.0324373493478803e-06,
|
| 1523 |
+
"logits/chosen": -2.0138721466064453,
|
| 1524 |
+
"logits/rejected": -2.004511833190918,
|
| 1525 |
+
"logps/chosen": -31.145587921142578,
|
| 1526 |
+
"logps/rejected": -35.140323638916016,
|
| 1527 |
+
"loss": 0.09,
|
| 1528 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1529 |
+
"rewards/chosen": 1.0254095792770386,
|
| 1530 |
+
"rewards/margins": 1.653183937072754,
|
| 1531 |
+
"rewards/rejected": -0.6277744174003601,
|
| 1532 |
+
"step": 930
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 2.44,
|
| 1536 |
+
"grad_norm": 17.0,
|
| 1537 |
+
"learning_rate": 1.976895560604729e-06,
|
| 1538 |
+
"logits/chosen": -1.9402869939804077,
|
| 1539 |
+
"logits/rejected": -1.9371931552886963,
|
| 1540 |
+
"logps/chosen": -30.240589141845703,
|
| 1541 |
+
"logps/rejected": -33.107887268066406,
|
| 1542 |
+
"loss": 0.1529,
|
| 1543 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1544 |
+
"rewards/chosen": 0.9430824518203735,
|
| 1545 |
+
"rewards/margins": 1.5213444232940674,
|
| 1546 |
+
"rewards/rejected": -0.578262209892273,
|
| 1547 |
+
"step": 940
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 2.47,
|
| 1551 |
+
"grad_norm": 4.25,
|
| 1552 |
+
"learning_rate": 1.921622518534466e-06,
|
| 1553 |
+
"logits/chosen": -1.8814115524291992,
|
| 1554 |
+
"logits/rejected": -1.8888870477676392,
|
| 1555 |
+
"logps/chosen": -31.7093448638916,
|
| 1556 |
+
"logps/rejected": -36.24521255493164,
|
| 1557 |
+
"loss": 0.181,
|
| 1558 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1559 |
+
"rewards/chosen": 1.0455348491668701,
|
| 1560 |
+
"rewards/margins": 1.671287178993225,
|
| 1561 |
+
"rewards/rejected": -0.625752329826355,
|
| 1562 |
+
"step": 950
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 2.49,
|
| 1566 |
+
"grad_norm": 0.04150390625,
|
| 1567 |
+
"learning_rate": 1.8666466198491794e-06,
|
| 1568 |
+
"logits/chosen": -1.8903650045394897,
|
| 1569 |
+
"logits/rejected": -1.8835119009017944,
|
| 1570 |
+
"logps/chosen": -32.601585388183594,
|
| 1571 |
+
"logps/rejected": -37.4316520690918,
|
| 1572 |
+
"loss": 0.0724,
|
| 1573 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1574 |
+
"rewards/chosen": 1.104241132736206,
|
| 1575 |
+
"rewards/margins": 1.853803277015686,
|
| 1576 |
+
"rewards/rejected": -0.7495620846748352,
|
| 1577 |
+
"step": 960
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 2.52,
|
| 1581 |
+
"grad_norm": 0.0,
|
| 1582 |
+
"learning_rate": 1.8119961086025376e-06,
|
| 1583 |
+
"logits/chosen": -1.8931461572647095,
|
| 1584 |
+
"logits/rejected": -1.8937944173812866,
|
| 1585 |
+
"logps/chosen": -30.099145889282227,
|
| 1586 |
+
"logps/rejected": -33.60529327392578,
|
| 1587 |
+
"loss": 0.1061,
|
| 1588 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1589 |
+
"rewards/chosen": 1.0235319137573242,
|
| 1590 |
+
"rewards/margins": 1.585160493850708,
|
| 1591 |
+
"rewards/rejected": -0.5616283416748047,
|
| 1592 |
+
"step": 970
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 2.55,
|
| 1596 |
+
"grad_norm": 9.75,
|
| 1597 |
+
"learning_rate": 1.7576990616793139e-06,
|
| 1598 |
+
"logits/chosen": -1.8976783752441406,
|
| 1599 |
+
"logits/rejected": -1.9089406728744507,
|
| 1600 |
+
"logps/chosen": -31.712087631225586,
|
| 1601 |
+
"logps/rejected": -34.73202133178711,
|
| 1602 |
+
"loss": 0.1285,
|
| 1603 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1604 |
+
"rewards/chosen": 1.2281408309936523,
|
| 1605 |
+
"rewards/margins": 1.7688732147216797,
|
| 1606 |
+
"rewards/rejected": -0.5407322645187378,
|
| 1607 |
+
"step": 980
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 2.57,
|
| 1611 |
+
"grad_norm": 9.875,
|
| 1612 |
+
"learning_rate": 1.7037833743707892e-06,
|
| 1613 |
+
"logits/chosen": -1.9692922830581665,
|
| 1614 |
+
"logits/rejected": -1.966718077659607,
|
| 1615 |
+
"logps/chosen": -34.942657470703125,
|
| 1616 |
+
"logps/rejected": -32.915771484375,
|
| 1617 |
+
"loss": 0.1929,
|
| 1618 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1619 |
+
"rewards/chosen": 1.0605528354644775,
|
| 1620 |
+
"rewards/margins": 1.525442361831665,
|
| 1621 |
+
"rewards/rejected": -0.4648895263671875,
|
| 1622 |
+
"step": 990
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 2.6,
|
| 1626 |
+
"grad_norm": 19.25,
|
| 1627 |
+
"learning_rate": 1.6502767460434588e-06,
|
| 1628 |
+
"logits/chosen": -1.9397118091583252,
|
| 1629 |
+
"logits/rejected": -1.9436838626861572,
|
| 1630 |
+
"logps/chosen": -33.27742385864258,
|
| 1631 |
+
"logps/rejected": -35.068485260009766,
|
| 1632 |
+
"loss": 0.123,
|
| 1633 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1634 |
+
"rewards/chosen": 1.0672835111618042,
|
| 1635 |
+
"rewards/margins": 1.5914794206619263,
|
| 1636 |
+
"rewards/rejected": -0.5241960287094116,
|
| 1637 |
+
"step": 1000
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 2.6,
|
| 1641 |
+
"eval_logits/chosen": -2.2042503356933594,
|
| 1642 |
+
"eval_logits/rejected": -2.199423313140869,
|
| 1643 |
+
"eval_logps/chosen": -34.13529586791992,
|
| 1644 |
+
"eval_logps/rejected": -37.72508239746094,
|
| 1645 |
+
"eval_loss": 0.9194008708000183,
|
| 1646 |
+
"eval_rewards/accuracies": 0.5510797500610352,
|
| 1647 |
+
"eval_rewards/chosen": -0.09066839516162872,
|
| 1648 |
+
"eval_rewards/margins": 0.09694948047399521,
|
| 1649 |
+
"eval_rewards/rejected": -0.18761785328388214,
|
| 1650 |
+
"eval_runtime": 145.3351,
|
| 1651 |
+
"eval_samples_per_second": 2.36,
|
| 1652 |
+
"eval_steps_per_second": 0.296,
|
| 1653 |
+
"step": 1000
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 2.62,
|
| 1657 |
+
"grad_norm": 28.0,
|
| 1658 |
+
"learning_rate": 1.5972066659083796e-06,
|
| 1659 |
+
"logits/chosen": -1.9486963748931885,
|
| 1660 |
+
"logits/rejected": -1.9537280797958374,
|
| 1661 |
+
"logps/chosen": -31.729883193969727,
|
| 1662 |
+
"logps/rejected": -33.12669372558594,
|
| 1663 |
+
"loss": 0.1503,
|
| 1664 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1665 |
+
"rewards/chosen": 0.9634879231452942,
|
| 1666 |
+
"rewards/margins": 1.5083791017532349,
|
| 1667 |
+
"rewards/rejected": -0.5448910593986511,
|
| 1668 |
+
"step": 1010
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 2.65,
|
| 1672 |
+
"grad_norm": 21.125,
|
| 1673 |
+
"learning_rate": 1.5446003988985041e-06,
|
| 1674 |
+
"logits/chosen": -1.9090359210968018,
|
| 1675 |
+
"logits/rejected": -1.911655068397522,
|
| 1676 |
+
"logps/chosen": -29.359201431274414,
|
| 1677 |
+
"logps/rejected": -32.33182144165039,
|
| 1678 |
+
"loss": 0.1044,
|
| 1679 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1680 |
+
"rewards/chosen": 0.9777968525886536,
|
| 1681 |
+
"rewards/margins": 1.5471875667572021,
|
| 1682 |
+
"rewards/rejected": -0.5693906545639038,
|
| 1683 |
+
"step": 1020
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 2.68,
|
| 1687 |
+
"grad_norm": 0.0,
|
| 1688 |
+
"learning_rate": 1.4924849716612211e-06,
|
| 1689 |
+
"logits/chosen": -1.96365225315094,
|
| 1690 |
+
"logits/rejected": -1.9579105377197266,
|
| 1691 |
+
"logps/chosen": -32.81294631958008,
|
| 1692 |
+
"logps/rejected": -34.20869064331055,
|
| 1693 |
+
"loss": 0.1032,
|
| 1694 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1695 |
+
"rewards/chosen": 1.0585477352142334,
|
| 1696 |
+
"rewards/margins": 1.6801731586456299,
|
| 1697 |
+
"rewards/rejected": -0.6216254830360413,
|
| 1698 |
+
"step": 1030
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 2.7,
|
| 1702 |
+
"grad_norm": 28.875,
|
| 1703 |
+
"learning_rate": 1.440887158673332e-06,
|
| 1704 |
+
"logits/chosen": -1.8949559926986694,
|
| 1705 |
+
"logits/rejected": -1.8988254070281982,
|
| 1706 |
+
"logps/chosen": -34.670326232910156,
|
| 1707 |
+
"logps/rejected": -35.01986312866211,
|
| 1708 |
+
"loss": 0.1415,
|
| 1709 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1710 |
+
"rewards/chosen": 0.9750639200210571,
|
| 1711 |
+
"rewards/margins": 1.515320062637329,
|
| 1712 |
+
"rewards/rejected": -0.5402560830116272,
|
| 1713 |
+
"step": 1040
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 2.73,
|
| 1717 |
+
"grad_norm": 0.0,
|
| 1718 |
+
"learning_rate": 1.3898334684855647e-06,
|
| 1719 |
+
"logits/chosen": -1.8945505619049072,
|
| 1720 |
+
"logits/rejected": -1.9076248407363892,
|
| 1721 |
+
"logps/chosen": -30.79265785217285,
|
| 1722 |
+
"logps/rejected": -33.660736083984375,
|
| 1723 |
+
"loss": 0.1249,
|
| 1724 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1725 |
+
"rewards/chosen": 0.9632073640823364,
|
| 1726 |
+
"rewards/margins": 1.5398341417312622,
|
| 1727 |
+
"rewards/rejected": -0.5766268372535706,
|
| 1728 |
+
"step": 1050
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 2.75,
|
| 1732 |
+
"grad_norm": 25.75,
|
| 1733 |
+
"learning_rate": 1.3393501301037245e-06,
|
| 1734 |
+
"logits/chosen": -1.8843669891357422,
|
| 1735 |
+
"logits/rejected": -1.8777261972427368,
|
| 1736 |
+
"logps/chosen": -31.016742706298828,
|
| 1737 |
+
"logps/rejected": -33.897727966308594,
|
| 1738 |
+
"loss": 0.0762,
|
| 1739 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1740 |
+
"rewards/chosen": 1.028328776359558,
|
| 1741 |
+
"rewards/margins": 1.5501177310943604,
|
| 1742 |
+
"rewards/rejected": -0.5217889547348022,
|
| 1743 |
+
"step": 1060
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 2.78,
|
| 1747 |
+
"grad_norm": 0.0,
|
| 1748 |
+
"learning_rate": 1.2894630795134454e-06,
|
| 1749 |
+
"logits/chosen": -2.0126426219940186,
|
| 1750 |
+
"logits/rejected": -2.0134384632110596,
|
| 1751 |
+
"logps/chosen": -32.44770050048828,
|
| 1752 |
+
"logps/rejected": -33.501129150390625,
|
| 1753 |
+
"loss": 0.0927,
|
| 1754 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1755 |
+
"rewards/chosen": 0.98944091796875,
|
| 1756 |
+
"rewards/margins": 1.5802329778671265,
|
| 1757 |
+
"rewards/rejected": -0.590792179107666,
|
| 1758 |
+
"step": 1070
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 2.81,
|
| 1762 |
+
"grad_norm": 26.125,
|
| 1763 |
+
"learning_rate": 1.2401979463554984e-06,
|
| 1764 |
+
"logits/chosen": -1.9425004720687866,
|
| 1765 |
+
"logits/rejected": -1.9413063526153564,
|
| 1766 |
+
"logps/chosen": -33.02809524536133,
|
| 1767 |
+
"logps/rejected": -33.649681091308594,
|
| 1768 |
+
"loss": 0.2202,
|
| 1769 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1770 |
+
"rewards/chosen": 0.9555650949478149,
|
| 1771 |
+
"rewards/margins": 1.3676296472549438,
|
| 1772 |
+
"rewards/rejected": -0.4120645523071289,
|
| 1773 |
+
"step": 1080
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 2.83,
|
| 1777 |
+
"grad_norm": 0.0,
|
| 1778 |
+
"learning_rate": 1.1915800407584705e-06,
|
| 1779 |
+
"logits/chosen": -1.970933198928833,
|
| 1780 |
+
"logits/rejected": -1.9633777141571045,
|
| 1781 |
+
"logps/chosen": -33.092559814453125,
|
| 1782 |
+
"logps/rejected": -31.60101890563965,
|
| 1783 |
+
"loss": 0.0991,
|
| 1784 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1785 |
+
"rewards/chosen": 1.0199841260910034,
|
| 1786 |
+
"rewards/margins": 1.5235220193862915,
|
| 1787 |
+
"rewards/rejected": -0.5035377740859985,
|
| 1788 |
+
"step": 1090
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 2.86,
|
| 1792 |
+
"grad_norm": 25.625,
|
| 1793 |
+
"learning_rate": 1.1436343403356019e-06,
|
| 1794 |
+
"logits/chosen": -1.939095139503479,
|
| 1795 |
+
"logits/rejected": -1.9381242990493774,
|
| 1796 |
+
"logps/chosen": -34.21430206298828,
|
| 1797 |
+
"logps/rejected": -37.063575744628906,
|
| 1798 |
+
"loss": 0.0894,
|
| 1799 |
+
"rewards/accuracies": 1.0,
|
| 1800 |
+
"rewards/chosen": 1.083695650100708,
|
| 1801 |
+
"rewards/margins": 1.7234748601913452,
|
| 1802 |
+
"rewards/rejected": -0.6397790908813477,
|
| 1803 |
+
"step": 1100
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 2.86,
|
| 1807 |
+
"eval_logits/chosen": -2.2049529552459717,
|
| 1808 |
+
"eval_logits/rejected": -2.200129747390747,
|
| 1809 |
+
"eval_logps/chosen": -34.136192321777344,
|
| 1810 |
+
"eval_logps/rejected": -37.72666931152344,
|
| 1811 |
+
"eval_loss": 0.9181937575340271,
|
| 1812 |
+
"eval_rewards/accuracies": 0.5336378812789917,
|
| 1813 |
+
"eval_rewards/chosen": -0.09147636592388153,
|
| 1814 |
+
"eval_rewards/margins": 0.09756775945425034,
|
| 1815 |
+
"eval_rewards/rejected": -0.18904413282871246,
|
| 1816 |
+
"eval_runtime": 145.1219,
|
| 1817 |
+
"eval_samples_per_second": 2.364,
|
| 1818 |
+
"eval_steps_per_second": 0.296,
|
| 1819 |
+
"step": 1100
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"epoch": 2.88,
|
| 1823 |
+
"grad_norm": 16.625,
|
| 1824 |
+
"learning_rate": 1.0963854773524548e-06,
|
| 1825 |
+
"logits/chosen": -1.9868457317352295,
|
| 1826 |
+
"logits/rejected": -1.9922664165496826,
|
| 1827 |
+
"logps/chosen": -34.364707946777344,
|
| 1828 |
+
"logps/rejected": -36.20539093017578,
|
| 1829 |
+
"loss": 0.1547,
|
| 1830 |
+
"rewards/accuracies": 0.9375,
|
| 1831 |
+
"rewards/chosen": 0.992165744304657,
|
| 1832 |
+
"rewards/margins": 1.5877904891967773,
|
| 1833 |
+
"rewards/rejected": -0.5956246852874756,
|
| 1834 |
+
"step": 1110
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 2.91,
|
| 1838 |
+
"grad_norm": 29.0,
|
| 1839 |
+
"learning_rate": 1.049857726072005e-06,
|
| 1840 |
+
"logits/chosen": -1.9590438604354858,
|
| 1841 |
+
"logits/rejected": -1.9570538997650146,
|
| 1842 |
+
"logps/chosen": -31.564861297607422,
|
| 1843 |
+
"logps/rejected": -33.61270523071289,
|
| 1844 |
+
"loss": 0.1168,
|
| 1845 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1846 |
+
"rewards/chosen": 0.9462112188339233,
|
| 1847 |
+
"rewards/margins": 1.5131675004959106,
|
| 1848 |
+
"rewards/rejected": -0.5669562220573425,
|
| 1849 |
+
"step": 1120
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 2.94,
|
| 1853 |
+
"grad_norm": 17.0,
|
| 1854 |
+
"learning_rate": 1.0040749902836508e-06,
|
| 1855 |
+
"logits/chosen": -1.8618561029434204,
|
| 1856 |
+
"logits/rejected": -1.8646554946899414,
|
| 1857 |
+
"logps/chosen": -28.602588653564453,
|
| 1858 |
+
"logps/rejected": -31.00104331970215,
|
| 1859 |
+
"loss": 0.0896,
|
| 1860 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1861 |
+
"rewards/chosen": 1.0314428806304932,
|
| 1862 |
+
"rewards/margins": 1.7118492126464844,
|
| 1863 |
+
"rewards/rejected": -0.6804062128067017,
|
| 1864 |
+
"step": 1130
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 2.96,
|
| 1868 |
+
"grad_norm": 11.6875,
|
| 1869 |
+
"learning_rate": 9.59060791022566e-07,
|
| 1870 |
+
"logits/chosen": -1.9582477807998657,
|
| 1871 |
+
"logits/rejected": -1.955122709274292,
|
| 1872 |
+
"logps/chosen": -32.12145233154297,
|
| 1873 |
+
"logps/rejected": -33.0306282043457,
|
| 1874 |
+
"loss": 0.1423,
|
| 1875 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1876 |
+
"rewards/chosen": 0.9740654826164246,
|
| 1877 |
+
"rewards/margins": 1.4274402856826782,
|
| 1878 |
+
"rewards/rejected": -0.4533747136592865,
|
| 1879 |
+
"step": 1140
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 2.99,
|
| 1883 |
+
"grad_norm": 11.875,
|
| 1884 |
+
"learning_rate": 9.148382544856885e-07,
|
| 1885 |
+
"logits/chosen": -1.878685712814331,
|
| 1886 |
+
"logits/rejected": -1.8723747730255127,
|
| 1887 |
+
"logps/chosen": -27.11667251586914,
|
| 1888 |
+
"logps/rejected": -32.9290771484375,
|
| 1889 |
+
"loss": 0.1014,
|
| 1890 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1891 |
+
"rewards/chosen": 0.9191849827766418,
|
| 1892 |
+
"rewards/margins": 1.6218817234039307,
|
| 1893 |
+
"rewards/rejected": -0.7026965618133545,
|
| 1894 |
+
"step": 1150
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 3.01,
|
| 1898 |
+
"grad_norm": 0.0,
|
| 1899 |
+
"learning_rate": 8.714301001505568e-07,
|
| 1900 |
+
"logits/chosen": -1.9950920343399048,
|
| 1901 |
+
"logits/rejected": -1.9926071166992188,
|
| 1902 |
+
"logps/chosen": -31.993701934814453,
|
| 1903 |
+
"logps/rejected": -35.8871955871582,
|
| 1904 |
+
"loss": 0.0971,
|
| 1905 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1906 |
+
"rewards/chosen": 0.9921283721923828,
|
| 1907 |
+
"rewards/margins": 1.7082250118255615,
|
| 1908 |
+
"rewards/rejected": -0.7160967588424683,
|
| 1909 |
+
"step": 1160
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 3.04,
|
| 1913 |
+
"grad_norm": 0.0,
|
| 1914 |
+
"learning_rate": 8.288586291031025e-07,
|
| 1915 |
+
"logits/chosen": -1.9799201488494873,
|
| 1916 |
+
"logits/rejected": -1.977949857711792,
|
| 1917 |
+
"logps/chosen": -30.49740982055664,
|
| 1918 |
+
"logps/rejected": -33.171287536621094,
|
| 1919 |
+
"loss": 0.0419,
|
| 1920 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1921 |
+
"rewards/chosen": 1.005298376083374,
|
| 1922 |
+
"rewards/margins": 1.7318332195281982,
|
| 1923 |
+
"rewards/rejected": -0.7265347242355347,
|
| 1924 |
+
"step": 1170
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 3.06,
|
| 1928 |
+
"grad_norm": 12.9375,
|
| 1929 |
+
"learning_rate": 7.871457125803897e-07,
|
| 1930 |
+
"logits/chosen": -1.972078561782837,
|
| 1931 |
+
"logits/rejected": -1.959705114364624,
|
| 1932 |
+
"logps/chosen": -34.829872131347656,
|
| 1933 |
+
"logps/rejected": -34.836021423339844,
|
| 1934 |
+
"loss": 0.0889,
|
| 1935 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1936 |
+
"rewards/chosen": 1.0938866138458252,
|
| 1937 |
+
"rewards/margins": 1.7819774150848389,
|
| 1938 |
+
"rewards/rejected": -0.6880909204483032,
|
| 1939 |
+
"step": 1180
|
| 1940 |
+
},
|
| 1941 |
+
{
|
| 1942 |
+
"epoch": 3.09,
|
| 1943 |
+
"grad_norm": 0.0,
|
| 1944 |
+
"learning_rate": 7.463127807341966e-07,
|
| 1945 |
+
"logits/chosen": -1.865128755569458,
|
| 1946 |
+
"logits/rejected": -1.8597571849822998,
|
| 1947 |
+
"logps/chosen": -34.058189392089844,
|
| 1948 |
+
"logps/rejected": -34.86783981323242,
|
| 1949 |
+
"loss": 0.0598,
|
| 1950 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1951 |
+
"rewards/chosen": 1.1623505353927612,
|
| 1952 |
+
"rewards/margins": 1.729365587234497,
|
| 1953 |
+
"rewards/rejected": -0.5670150518417358,
|
| 1954 |
+
"step": 1190
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 3.12,
|
| 1958 |
+
"grad_norm": 13.8125,
|
| 1959 |
+
"learning_rate": 7.063808116212021e-07,
|
| 1960 |
+
"logits/chosen": -1.9072997570037842,
|
| 1961 |
+
"logits/rejected": -1.9104669094085693,
|
| 1962 |
+
"logps/chosen": -31.529245376586914,
|
| 1963 |
+
"logps/rejected": -32.14056396484375,
|
| 1964 |
+
"loss": 0.1086,
|
| 1965 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1966 |
+
"rewards/chosen": 0.9772661328315735,
|
| 1967 |
+
"rewards/margins": 1.5662124156951904,
|
| 1968 |
+
"rewards/rejected": -0.5889462232589722,
|
| 1969 |
+
"step": 1200
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 3.12,
|
| 1973 |
+
"eval_logits/chosen": -2.205395221710205,
|
| 1974 |
+
"eval_logits/rejected": -2.2005577087402344,
|
| 1975 |
+
"eval_logps/chosen": -34.13056945800781,
|
| 1976 |
+
"eval_logps/rejected": -37.736209869384766,
|
| 1977 |
+
"eval_loss": 0.9022552967071533,
|
| 1978 |
+
"eval_rewards/accuracies": 0.5627076625823975,
|
| 1979 |
+
"eval_rewards/chosen": -0.08641883730888367,
|
| 1980 |
+
"eval_rewards/margins": 0.11121704429388046,
|
| 1981 |
+
"eval_rewards/rejected": -0.19763588905334473,
|
| 1982 |
+
"eval_runtime": 145.103,
|
| 1983 |
+
"eval_samples_per_second": 2.364,
|
| 1984 |
+
"eval_steps_per_second": 0.296,
|
| 1985 |
+
"step": 1200
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 3.14,
|
| 1989 |
+
"grad_norm": 23.125,
|
| 1990 |
+
"learning_rate": 6.673703204254348e-07,
|
| 1991 |
+
"logits/chosen": -1.9353859424591064,
|
| 1992 |
+
"logits/rejected": -1.9308178424835205,
|
| 1993 |
+
"logps/chosen": -30.187076568603516,
|
| 1994 |
+
"logps/rejected": -31.651758193969727,
|
| 1995 |
+
"loss": 0.0719,
|
| 1996 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1997 |
+
"rewards/chosen": 0.9779340028762817,
|
| 1998 |
+
"rewards/margins": 1.647231101989746,
|
| 1999 |
+
"rewards/rejected": -0.6692970991134644,
|
| 2000 |
+
"step": 1210
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 3.17,
|
| 2004 |
+
"grad_norm": 0.0,
|
| 2005 |
+
"learning_rate": 6.293013489185315e-07,
|
| 2006 |
+
"logits/chosen": -1.9491630792617798,
|
| 2007 |
+
"logits/rejected": -1.9436299800872803,
|
| 2008 |
+
"logps/chosen": -34.10167694091797,
|
| 2009 |
+
"logps/rejected": -35.294559478759766,
|
| 2010 |
+
"loss": 0.0867,
|
| 2011 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2012 |
+
"rewards/chosen": 1.055037260055542,
|
| 2013 |
+
"rewards/margins": 1.7072465419769287,
|
| 2014 |
+
"rewards/rejected": -0.6522093415260315,
|
| 2015 |
+
"step": 1220
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 3.19,
|
| 2019 |
+
"grad_norm": 0.0,
|
| 2020 |
+
"learning_rate": 5.921934551632086e-07,
|
| 2021 |
+
"logits/chosen": -1.9506566524505615,
|
| 2022 |
+
"logits/rejected": -1.9374538660049438,
|
| 2023 |
+
"logps/chosen": -31.99393081665039,
|
| 2024 |
+
"logps/rejected": -34.90449142456055,
|
| 2025 |
+
"loss": 0.1538,
|
| 2026 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2027 |
+
"rewards/chosen": 0.9900773167610168,
|
| 2028 |
+
"rewards/margins": 1.6962600946426392,
|
| 2029 |
+
"rewards/rejected": -0.706182599067688,
|
| 2030 |
+
"step": 1230
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 3.22,
|
| 2034 |
+
"grad_norm": 0.0,
|
| 2035 |
+
"learning_rate": 5.560657034652405e-07,
|
| 2036 |
+
"logits/chosen": -1.9950263500213623,
|
| 2037 |
+
"logits/rejected": -1.9922975301742554,
|
| 2038 |
+
"logps/chosen": -33.79475402832031,
|
| 2039 |
+
"logps/rejected": -32.4438362121582,
|
| 2040 |
+
"loss": 0.0771,
|
| 2041 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2042 |
+
"rewards/chosen": 1.114330530166626,
|
| 2043 |
+
"rewards/margins": 1.6578998565673828,
|
| 2044 |
+
"rewards/rejected": -0.5435694456100464,
|
| 2045 |
+
"step": 1240
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 3.25,
|
| 2049 |
+
"grad_norm": 16.5,
|
| 2050 |
+
"learning_rate": 5.2093665457911e-07,
|
| 2051 |
+
"logits/chosen": -1.9237819910049438,
|
| 2052 |
+
"logits/rejected": -1.9209115505218506,
|
| 2053 |
+
"logps/chosen": -33.26008987426758,
|
| 2054 |
+
"logps/rejected": -35.61740493774414,
|
| 2055 |
+
"loss": 0.0971,
|
| 2056 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2057 |
+
"rewards/chosen": 1.0224542617797852,
|
| 2058 |
+
"rewards/margins": 1.7607052326202393,
|
| 2059 |
+
"rewards/rejected": -0.7382508516311646,
|
| 2060 |
+
"step": 1250
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 3.27,
|
| 2064 |
+
"grad_norm": 15.125,
|
| 2065 |
+
"learning_rate": 4.868243561723535e-07,
|
| 2066 |
+
"logits/chosen": -1.988936185836792,
|
| 2067 |
+
"logits/rejected": -1.9837830066680908,
|
| 2068 |
+
"logps/chosen": -29.6236629486084,
|
| 2069 |
+
"logps/rejected": -33.01030349731445,
|
| 2070 |
+
"loss": 0.0567,
|
| 2071 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2072 |
+
"rewards/chosen": 1.0234204530715942,
|
| 2073 |
+
"rewards/margins": 1.5883967876434326,
|
| 2074 |
+
"rewards/rejected": -0.5649763345718384,
|
| 2075 |
+
"step": 1260
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 3.3,
|
| 2079 |
+
"grad_norm": 0.0,
|
| 2080 |
+
"learning_rate": 4.537463335535161e-07,
|
| 2081 |
+
"logits/chosen": -2.0158066749572754,
|
| 2082 |
+
"logits/rejected": -2.0205185413360596,
|
| 2083 |
+
"logps/chosen": -31.42230796813965,
|
| 2084 |
+
"logps/rejected": -32.35493087768555,
|
| 2085 |
+
"loss": 0.077,
|
| 2086 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2087 |
+
"rewards/chosen": 1.1598228216171265,
|
| 2088 |
+
"rewards/margins": 1.766034483909607,
|
| 2089 |
+
"rewards/rejected": -0.6062116026878357,
|
| 2090 |
+
"step": 1270
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 3.32,
|
| 2094 |
+
"grad_norm": 14.25,
|
| 2095 |
+
"learning_rate": 4.217195806684629e-07,
|
| 2096 |
+
"logits/chosen": -1.9084632396697998,
|
| 2097 |
+
"logits/rejected": -1.91574227809906,
|
| 2098 |
+
"logps/chosen": -33.739585876464844,
|
| 2099 |
+
"logps/rejected": -33.33063507080078,
|
| 2100 |
+
"loss": 0.0795,
|
| 2101 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2102 |
+
"rewards/chosen": 1.0508931875228882,
|
| 2103 |
+
"rewards/margins": 1.7100213766098022,
|
| 2104 |
+
"rewards/rejected": -0.6591281890869141,
|
| 2105 |
+
"step": 1280
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 3.35,
|
| 2109 |
+
"grad_norm": 15.5,
|
| 2110 |
+
"learning_rate": 3.907605513696808e-07,
|
| 2111 |
+
"logits/chosen": -1.8143936395645142,
|
| 2112 |
+
"logits/rejected": -1.8166534900665283,
|
| 2113 |
+
"logps/chosen": -31.8719539642334,
|
| 2114 |
+
"logps/rejected": -37.17786407470703,
|
| 2115 |
+
"loss": 0.0853,
|
| 2116 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2117 |
+
"rewards/chosen": 1.1339490413665771,
|
| 2118 |
+
"rewards/margins": 1.6951805353164673,
|
| 2119 |
+
"rewards/rejected": -0.5612314343452454,
|
| 2120 |
+
"step": 1290
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 3.38,
|
| 2124 |
+
"grad_norm": 28.25,
|
| 2125 |
+
"learning_rate": 3.6088515096305675e-07,
|
| 2126 |
+
"logits/chosen": -1.867759346961975,
|
| 2127 |
+
"logits/rejected": -1.8711694478988647,
|
| 2128 |
+
"logps/chosen": -32.002017974853516,
|
| 2129 |
+
"logps/rejected": -33.140037536621094,
|
| 2130 |
+
"loss": 0.0577,
|
| 2131 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2132 |
+
"rewards/chosen": 1.0931856632232666,
|
| 2133 |
+
"rewards/margins": 1.7014715671539307,
|
| 2134 |
+
"rewards/rejected": -0.6082859039306641,
|
| 2135 |
+
"step": 1300
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 3.38,
|
| 2139 |
+
"eval_logits/chosen": -2.204974412918091,
|
| 2140 |
+
"eval_logits/rejected": -2.2001500129699707,
|
| 2141 |
+
"eval_logps/chosen": -34.137020111083984,
|
| 2142 |
+
"eval_logps/rejected": -37.73167037963867,
|
| 2143 |
+
"eval_loss": 0.9154069423675537,
|
| 2144 |
+
"eval_rewards/accuracies": 0.5598006844520569,
|
| 2145 |
+
"eval_rewards/chosen": -0.09222196787595749,
|
| 2146 |
+
"eval_rewards/margins": 0.10132110863924026,
|
| 2147 |
+
"eval_rewards/rejected": -0.19354306161403656,
|
| 2148 |
+
"eval_runtime": 145.1705,
|
| 2149 |
+
"eval_samples_per_second": 2.363,
|
| 2150 |
+
"eval_steps_per_second": 0.296,
|
| 2151 |
+
"step": 1300
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 3.4,
|
| 2155 |
+
"grad_norm": 0.0301513671875,
|
| 2156 |
+
"learning_rate": 3.321087280364757e-07,
|
| 2157 |
+
"logits/chosen": -1.97760808467865,
|
| 2158 |
+
"logits/rejected": -1.9573789834976196,
|
| 2159 |
+
"logps/chosen": -29.92989158630371,
|
| 2160 |
+
"logps/rejected": -35.94379425048828,
|
| 2161 |
+
"loss": 0.1028,
|
| 2162 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2163 |
+
"rewards/chosen": 1.065155029296875,
|
| 2164 |
+
"rewards/margins": 1.764804482460022,
|
| 2165 |
+
"rewards/rejected": -0.6996492743492126,
|
| 2166 |
+
"step": 1310
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 3.43,
|
| 2170 |
+
"grad_norm": 3.5,
|
| 2171 |
+
"learning_rate": 3.044460665744284e-07,
|
| 2172 |
+
"logits/chosen": -1.9561742544174194,
|
| 2173 |
+
"logits/rejected": -1.9619709253311157,
|
| 2174 |
+
"logps/chosen": -31.752899169921875,
|
| 2175 |
+
"logps/rejected": -33.09962463378906,
|
| 2176 |
+
"loss": 0.0827,
|
| 2177 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2178 |
+
"rewards/chosen": 1.15280020236969,
|
| 2179 |
+
"rewards/margins": 1.7007869482040405,
|
| 2180 |
+
"rewards/rejected": -0.5479868650436401,
|
| 2181 |
+
"step": 1320
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 3.45,
|
| 2185 |
+
"grad_norm": 0.0211181640625,
|
| 2186 |
+
"learning_rate": 2.779113783626916e-07,
|
| 2187 |
+
"logits/chosen": -1.9276313781738281,
|
| 2188 |
+
"logits/rejected": -1.9227180480957031,
|
| 2189 |
+
"logps/chosen": -31.962305068969727,
|
| 2190 |
+
"logps/rejected": -35.37018585205078,
|
| 2191 |
+
"loss": 0.0646,
|
| 2192 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2193 |
+
"rewards/chosen": 0.9541842341423035,
|
| 2194 |
+
"rewards/margins": 1.6722910404205322,
|
| 2195 |
+
"rewards/rejected": -0.7181065678596497,
|
| 2196 |
+
"step": 1330
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 3.48,
|
| 2200 |
+
"grad_norm": 0.0,
|
| 2201 |
+
"learning_rate": 2.5251829568697204e-07,
|
| 2202 |
+
"logits/chosen": -1.790858507156372,
|
| 2203 |
+
"logits/rejected": -1.8005390167236328,
|
| 2204 |
+
"logps/chosen": -32.53480911254883,
|
| 2205 |
+
"logps/rejected": -31.600765228271484,
|
| 2206 |
+
"loss": 0.0607,
|
| 2207 |
+
"rewards/accuracies": 1.0,
|
| 2208 |
+
"rewards/chosen": 1.1383044719696045,
|
| 2209 |
+
"rewards/margins": 1.6482902765274048,
|
| 2210 |
+
"rewards/rejected": -0.5099858045578003,
|
| 2211 |
+
"step": 1340
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 3.51,
|
| 2215 |
+
"grad_norm": 0.0,
|
| 2216 |
+
"learning_rate": 2.2827986432927774e-07,
|
| 2217 |
+
"logits/chosen": -1.8518263101577759,
|
| 2218 |
+
"logits/rejected": -1.8565566539764404,
|
| 2219 |
+
"logps/chosen": -32.678627014160156,
|
| 2220 |
+
"logps/rejected": -33.76792526245117,
|
| 2221 |
+
"loss": 0.0315,
|
| 2222 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2223 |
+
"rewards/chosen": 1.1450836658477783,
|
| 2224 |
+
"rewards/margins": 1.6518867015838623,
|
| 2225 |
+
"rewards/rejected": -0.5068029165267944,
|
| 2226 |
+
"step": 1350
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 3.53,
|
| 2230 |
+
"grad_norm": 31.375,
|
| 2231 |
+
"learning_rate": 2.0520853686560177e-07,
|
| 2232 |
+
"logits/chosen": -1.8785499334335327,
|
| 2233 |
+
"logits/rejected": -1.8840558528900146,
|
| 2234 |
+
"logps/chosen": -32.6619873046875,
|
| 2235 |
+
"logps/rejected": -35.23099136352539,
|
| 2236 |
+
"loss": 0.0754,
|
| 2237 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2238 |
+
"rewards/chosen": 1.0527045726776123,
|
| 2239 |
+
"rewards/margins": 1.6538400650024414,
|
| 2240 |
+
"rewards/rejected": -0.6011354923248291,
|
| 2241 |
+
"step": 1360
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 3.56,
|
| 2245 |
+
"grad_norm": 0.02587890625,
|
| 2246 |
+
"learning_rate": 1.833161662683672e-07,
|
| 2247 |
+
"logits/chosen": -1.9697166681289673,
|
| 2248 |
+
"logits/rejected": -1.9651731252670288,
|
| 2249 |
+
"logps/chosen": -31.484020233154297,
|
| 2250 |
+
"logps/rejected": -31.601449966430664,
|
| 2251 |
+
"loss": 0.0507,
|
| 2252 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2253 |
+
"rewards/chosen": 1.0364978313446045,
|
| 2254 |
+
"rewards/margins": 1.657091498374939,
|
| 2255 |
+
"rewards/rejected": -0.6205938458442688,
|
| 2256 |
+
"step": 1370
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 3.58,
|
| 2260 |
+
"grad_norm": 16.875,
|
| 2261 |
+
"learning_rate": 1.626139998169246e-07,
|
| 2262 |
+
"logits/chosen": -1.888527274131775,
|
| 2263 |
+
"logits/rejected": -1.8903592824935913,
|
| 2264 |
+
"logps/chosen": -28.53927993774414,
|
| 2265 |
+
"logps/rejected": -31.388671875,
|
| 2266 |
+
"loss": 0.0877,
|
| 2267 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2268 |
+
"rewards/chosen": 1.0080214738845825,
|
| 2269 |
+
"rewards/margins": 1.5880650281906128,
|
| 2270 |
+
"rewards/rejected": -0.5800435543060303,
|
| 2271 |
+
"step": 1380
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 3.61,
|
| 2275 |
+
"grad_norm": 21.125,
|
| 2276 |
+
"learning_rate": 1.4311267331922535e-07,
|
| 2277 |
+
"logits/chosen": -1.8796398639678955,
|
| 2278 |
+
"logits/rejected": -1.8842071294784546,
|
| 2279 |
+
"logps/chosen": -30.890060424804688,
|
| 2280 |
+
"logps/rejected": -33.532711029052734,
|
| 2281 |
+
"loss": 0.0687,
|
| 2282 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2283 |
+
"rewards/chosen": 1.0164906978607178,
|
| 2284 |
+
"rewards/margins": 1.50667405128479,
|
| 2285 |
+
"rewards/rejected": -0.49018335342407227,
|
| 2286 |
+
"step": 1390
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 3.64,
|
| 2290 |
+
"grad_norm": 0.0,
|
| 2291 |
+
"learning_rate": 1.2482220564763669e-07,
|
| 2292 |
+
"logits/chosen": -1.9481089115142822,
|
| 2293 |
+
"logits/rejected": -1.9513651132583618,
|
| 2294 |
+
"logps/chosen": -33.666908264160156,
|
| 2295 |
+
"logps/rejected": -34.179168701171875,
|
| 2296 |
+
"loss": 0.0375,
|
| 2297 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2298 |
+
"rewards/chosen": 1.1372981071472168,
|
| 2299 |
+
"rewards/margins": 1.8173433542251587,
|
| 2300 |
+
"rewards/rejected": -0.6800452470779419,
|
| 2301 |
+
"step": 1400
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 3.64,
|
| 2305 |
+
"eval_logits/chosen": -2.204983949661255,
|
| 2306 |
+
"eval_logits/rejected": -2.200157403945923,
|
| 2307 |
+
"eval_logps/chosen": -34.134159088134766,
|
| 2308 |
+
"eval_logps/rejected": -37.7177848815918,
|
| 2309 |
+
"eval_loss": 0.9233020544052124,
|
| 2310 |
+
"eval_rewards/accuracies": 0.5568937063217163,
|
| 2311 |
+
"eval_rewards/chosen": -0.08964409679174423,
|
| 2312 |
+
"eval_rewards/margins": 0.09140493720769882,
|
| 2313 |
+
"eval_rewards/rejected": -0.18104901909828186,
|
| 2314 |
+
"eval_runtime": 145.1261,
|
| 2315 |
+
"eval_samples_per_second": 2.363,
|
| 2316 |
+
"eval_steps_per_second": 0.296,
|
| 2317 |
+
"step": 1400
|
| 2318 |
+
},
|
| 2319 |
+
{
|
| 2320 |
+
"epoch": 3.66,
|
| 2321 |
+
"grad_norm": 14.875,
|
| 2322 |
+
"learning_rate": 1.0775199359171346e-07,
|
| 2323 |
+
"logits/chosen": -2.017505407333374,
|
| 2324 |
+
"logits/rejected": -2.010617256164551,
|
| 2325 |
+
"logps/chosen": -32.54536056518555,
|
| 2326 |
+
"logps/rejected": -34.80001449584961,
|
| 2327 |
+
"loss": 0.0811,
|
| 2328 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2329 |
+
"rewards/chosen": 1.131469488143921,
|
| 2330 |
+
"rewards/margins": 1.6313743591308594,
|
| 2331 |
+
"rewards/rejected": -0.4999050498008728,
|
| 2332 |
+
"step": 1410
|
| 2333 |
+
},
|
| 2334 |
+
{
|
| 2335 |
+
"epoch": 3.69,
|
| 2336 |
+
"grad_norm": 18.5,
|
| 2337 |
+
"learning_rate": 9.191080703056604e-08,
|
| 2338 |
+
"logits/chosen": -1.9044952392578125,
|
| 2339 |
+
"logits/rejected": -1.9153999090194702,
|
| 2340 |
+
"logps/chosen": -32.92113494873047,
|
| 2341 |
+
"logps/rejected": -34.28981018066406,
|
| 2342 |
+
"loss": 0.0577,
|
| 2343 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2344 |
+
"rewards/chosen": 1.1053516864776611,
|
| 2345 |
+
"rewards/margins": 1.7342084646224976,
|
| 2346 |
+
"rewards/rejected": -0.6288568377494812,
|
| 2347 |
+
"step": 1420
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 3.71,
|
| 2351 |
+
"grad_norm": 19.0,
|
| 2352 |
+
"learning_rate": 7.730678442730539e-08,
|
| 2353 |
+
"logits/chosen": -1.9869391918182373,
|
| 2354 |
+
"logits/rejected": -1.9988434314727783,
|
| 2355 |
+
"logps/chosen": -33.78189468383789,
|
| 2356 |
+
"logps/rejected": -34.36466598510742,
|
| 2357 |
+
"loss": 0.0555,
|
| 2358 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2359 |
+
"rewards/chosen": 1.1510204076766968,
|
| 2360 |
+
"rewards/margins": 1.8047173023223877,
|
| 2361 |
+
"rewards/rejected": -0.6536968350410461,
|
| 2362 |
+
"step": 1430
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 3.74,
|
| 2366 |
+
"grad_norm": 16.5,
|
| 2367 |
+
"learning_rate": 6.394742864787806e-08,
|
| 2368 |
+
"logits/chosen": -1.977538824081421,
|
| 2369 |
+
"logits/rejected": -1.979612112045288,
|
| 2370 |
+
"logps/chosen": -32.0444450378418,
|
| 2371 |
+
"logps/rejected": -34.551475524902344,
|
| 2372 |
+
"loss": 0.1098,
|
| 2373 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2374 |
+
"rewards/chosen": 1.0475465059280396,
|
| 2375 |
+
"rewards/margins": 1.6699031591415405,
|
| 2376 |
+
"rewards/rejected": -0.6223568320274353,
|
| 2377 |
+
"step": 1440
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 3.77,
|
| 2381 |
+
"grad_norm": 0.01312255859375,
|
| 2382 |
+
"learning_rate": 5.183960310644748e-08,
|
| 2383 |
+
"logits/chosen": -1.9596837759017944,
|
| 2384 |
+
"logits/rejected": -1.9525411128997803,
|
| 2385 |
+
"logps/chosen": -34.377952575683594,
|
| 2386 |
+
"logps/rejected": -34.37388229370117,
|
| 2387 |
+
"loss": 0.0864,
|
| 2388 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2389 |
+
"rewards/chosen": 1.0713341236114502,
|
| 2390 |
+
"rewards/margins": 1.6144660711288452,
|
| 2391 |
+
"rewards/rejected": -0.5431317090988159,
|
| 2392 |
+
"step": 1450
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 3.79,
|
| 2396 |
+
"grad_norm": 9.6875,
|
| 2397 |
+
"learning_rate": 4.098952823928693e-08,
|
| 2398 |
+
"logits/chosen": -1.9263242483139038,
|
| 2399 |
+
"logits/rejected": -1.9324493408203125,
|
| 2400 |
+
"logps/chosen": -30.057865142822266,
|
| 2401 |
+
"logps/rejected": -34.518531799316406,
|
| 2402 |
+
"loss": 0.0742,
|
| 2403 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2404 |
+
"rewards/chosen": 1.0397825241088867,
|
| 2405 |
+
"rewards/margins": 1.8034608364105225,
|
| 2406 |
+
"rewards/rejected": -0.7636784315109253,
|
| 2407 |
+
"step": 1460
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 3.82,
|
| 2411 |
+
"grad_norm": 11.9375,
|
| 2412 |
+
"learning_rate": 3.1402778309014284e-08,
|
| 2413 |
+
"logits/chosen": -1.8729248046875,
|
| 2414 |
+
"logits/rejected": -1.8707479238510132,
|
| 2415 |
+
"logps/chosen": -29.638717651367188,
|
| 2416 |
+
"logps/rejected": -31.215627670288086,
|
| 2417 |
+
"loss": 0.0943,
|
| 2418 |
+
"rewards/accuracies": 0.9375,
|
| 2419 |
+
"rewards/chosen": 1.0861549377441406,
|
| 2420 |
+
"rewards/margins": 1.6454105377197266,
|
| 2421 |
+
"rewards/rejected": -0.5592554211616516,
|
| 2422 |
+
"step": 1470
|
| 2423 |
+
},
|
| 2424 |
+
{
|
| 2425 |
+
"epoch": 3.84,
|
| 2426 |
+
"grad_norm": 0.0,
|
| 2427 |
+
"learning_rate": 2.3084278540791427e-08,
|
| 2428 |
+
"logits/chosen": -2.048567295074463,
|
| 2429 |
+
"logits/rejected": -2.042868137359619,
|
| 2430 |
+
"logps/chosen": -34.20145797729492,
|
| 2431 |
+
"logps/rejected": -32.80525588989258,
|
| 2432 |
+
"loss": 0.0547,
|
| 2433 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2434 |
+
"rewards/chosen": 0.9754201173782349,
|
| 2435 |
+
"rewards/margins": 1.649107575416565,
|
| 2436 |
+
"rewards/rejected": -0.673687756061554,
|
| 2437 |
+
"step": 1480
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 3.87,
|
| 2441 |
+
"grad_norm": 24.5,
|
| 2442 |
+
"learning_rate": 1.6038302591975807e-08,
|
| 2443 |
+
"logits/chosen": -1.939672827720642,
|
| 2444 |
+
"logits/rejected": -1.9418550729751587,
|
| 2445 |
+
"logps/chosen": -27.397912979125977,
|
| 2446 |
+
"logps/rejected": -28.60453224182129,
|
| 2447 |
+
"loss": 0.1225,
|
| 2448 |
+
"rewards/accuracies": 0.9375,
|
| 2449 |
+
"rewards/chosen": 0.9706217050552368,
|
| 2450 |
+
"rewards/margins": 1.4946484565734863,
|
| 2451 |
+
"rewards/rejected": -0.5240268111228943,
|
| 2452 |
+
"step": 1490
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 3.9,
|
| 2456 |
+
"grad_norm": 15.5625,
|
| 2457 |
+
"learning_rate": 1.0268470356514237e-08,
|
| 2458 |
+
"logits/chosen": -1.9486587047576904,
|
| 2459 |
+
"logits/rejected": -1.9430547952651978,
|
| 2460 |
+
"logps/chosen": -31.683019638061523,
|
| 2461 |
+
"logps/rejected": -32.714141845703125,
|
| 2462 |
+
"loss": 0.0724,
|
| 2463 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2464 |
+
"rewards/chosen": 1.0059928894042969,
|
| 2465 |
+
"rewards/margins": 1.6766607761383057,
|
| 2466 |
+
"rewards/rejected": -0.6706677675247192,
|
| 2467 |
+
"step": 1500
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 3.9,
|
| 2471 |
+
"eval_logits/chosen": -2.204876184463501,
|
| 2472 |
+
"eval_logits/rejected": -2.200045108795166,
|
| 2473 |
+
"eval_logps/chosen": -34.13526153564453,
|
| 2474 |
+
"eval_logps/rejected": -37.724769592285156,
|
| 2475 |
+
"eval_loss": 0.9175900816917419,
|
| 2476 |
+
"eval_rewards/accuracies": 0.5627076625823975,
|
| 2477 |
+
"eval_rewards/chosen": -0.09064043313264847,
|
| 2478 |
+
"eval_rewards/margins": 0.09669183194637299,
|
| 2479 |
+
"eval_rewards/rejected": -0.18733225762844086,
|
| 2480 |
+
"eval_runtime": 145.3111,
|
| 2481 |
+
"eval_samples_per_second": 2.36,
|
| 2482 |
+
"eval_steps_per_second": 0.296,
|
| 2483 |
+
"step": 1500
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 3.92,
|
| 2487 |
+
"grad_norm": 18.625,
|
| 2488 |
+
"learning_rate": 5.777746105209147e-09,
|
| 2489 |
+
"logits/chosen": -1.87344491481781,
|
| 2490 |
+
"logits/rejected": -1.8777220249176025,
|
| 2491 |
+
"logps/chosen": -33.06303405761719,
|
| 2492 |
+
"logps/rejected": -35.00675964355469,
|
| 2493 |
+
"loss": 0.1331,
|
| 2494 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2495 |
+
"rewards/chosen": 0.9971674084663391,
|
| 2496 |
+
"rewards/margins": 1.5339066982269287,
|
| 2497 |
+
"rewards/rejected": -0.5367392301559448,
|
| 2498 |
+
"step": 1510
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 3.95,
|
| 2502 |
+
"grad_norm": 6.90625,
|
| 2503 |
+
"learning_rate": 2.5684369628148352e-09,
|
| 2504 |
+
"logits/chosen": -1.9323539733886719,
|
| 2505 |
+
"logits/rejected": -1.9306873083114624,
|
| 2506 |
+
"logps/chosen": -29.582763671875,
|
| 2507 |
+
"logps/rejected": -33.175575256347656,
|
| 2508 |
+
"loss": 0.1185,
|
| 2509 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2510 |
+
"rewards/chosen": 1.0362732410430908,
|
| 2511 |
+
"rewards/margins": 1.5635396242141724,
|
| 2512 |
+
"rewards/rejected": -0.5272663831710815,
|
| 2513 |
+
"step": 1520
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 3.97,
|
| 2517 |
+
"grad_norm": 18.875,
|
| 2518 |
+
"learning_rate": 6.421917227455999e-10,
|
| 2519 |
+
"logits/chosen": -2.030973434448242,
|
| 2520 |
+
"logits/rejected": -2.0232560634613037,
|
| 2521 |
+
"logps/chosen": -27.034753799438477,
|
| 2522 |
+
"logps/rejected": -29.34645652770996,
|
| 2523 |
+
"loss": 0.0492,
|
| 2524 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2525 |
+
"rewards/chosen": 0.9253705143928528,
|
| 2526 |
+
"rewards/margins": 1.5327354669570923,
|
| 2527 |
+
"rewards/rejected": -0.6073648929595947,
|
| 2528 |
+
"step": 1530
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 4.0,
|
| 2532 |
+
"grad_norm": 14.4375,
|
| 2533 |
+
"learning_rate": 0.0,
|
| 2534 |
+
"logits/chosen": -1.9311290979385376,
|
| 2535 |
+
"logits/rejected": -1.9209285974502563,
|
| 2536 |
+
"logps/chosen": -31.977828979492188,
|
| 2537 |
+
"logps/rejected": -36.17755126953125,
|
| 2538 |
+
"loss": 0.0394,
|
| 2539 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2540 |
+
"rewards/chosen": 0.979533851146698,
|
| 2541 |
+
"rewards/margins": 1.6692752838134766,
|
| 2542 |
+
"rewards/rejected": -0.689741313457489,
|
| 2543 |
+
"step": 1540
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 4.0,
|
| 2547 |
+
"step": 1540,
|
| 2548 |
"total_flos": 0.0,
|
| 2549 |
+
"train_loss": 0.15107146371881683,
|
| 2550 |
+
"train_runtime": 10740.8737,
|
| 2551 |
+
"train_samples_per_second": 1.147,
|
| 2552 |
+
"train_steps_per_second": 0.143
|
| 2553 |
}
|
| 2554 |
],
|
| 2555 |
"logging_steps": 10,
|
| 2556 |
+
"max_steps": 1540,
|
| 2557 |
"num_input_tokens_seen": 0,
|
| 2558 |
+
"num_train_epochs": 4,
|
| 2559 |
"save_steps": 100,
|
| 2560 |
"total_flos": 0.0,
|
| 2561 |
"train_batch_size": 4,
|