Commit
•
9d33454
1
Parent(s):
f92adb3
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:
|
| 25 |
-
- Rewards/rejected:
|
| 26 |
-
- Rewards/accuracies: 0.
|
| 27 |
-
- Rewards/margins: 0.
|
| 28 |
-
- Logps/rejected: -37.
|
| 29 |
-
- Logps/chosen: -
|
| 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 |
-
| 0.4711 | 0.26 | 100 |
|
| 67 |
-
| 0.5061 | 0.52 | 200 |
|
| 68 |
-
| 0.3371 | 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: 0.5566
|
| 22 |
+
- Rewards/chosen: 0.0778
|
| 23 |
+
- Rewards/rejected: 0.0602
|
| 24 |
+
- Rewards/accuracies: 0.5050
|
| 25 |
+
- Rewards/margins: 0.0176
|
| 26 |
+
- Logps/rejected: -37.4306
|
| 27 |
+
- Logps/chosen: -33.9234
|
| 28 |
+
- Logits/rejected: -2.2423
|
| 29 |
+
- Logits/chosen: -2.2472
|
| 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 |
+
| 0.4711 | 0.26 | 100 | -2.2401 | -2.2352 | -34.0113 | -37.4979 | 0.5755 | 0.5195 | 0.0163 | 0.0032 | 0.0131 |
|
| 65 |
+
| 0.5061 | 0.52 | 200 | -2.2385 | -2.2337 | -34.0500 | -37.5455 | 0.5877 | 0.4992 | -0.0108 | 0.0094 | -0.0202 |
|
| 66 |
+
| 0.3371 | 0.78 | 300 | -2.2371 | -2.2322 | -34.0344 | -37.5353 | 0.5843 | 0.5278 | 0.0001 | 0.0132 | -0.0131 |
|
| 67 |
+
| 0.4001 | 1.04 | 400 | 0.6350 | -0.0073 | 0.0033 | 0.4838 | -0.0106 | -37.5120 | -34.0450 | -2.2353 | -2.2402 |
|
| 68 |
+
| 0.3401 | 1.3 | 500 | 0.6238 | -0.0135 | -0.0193 | 0.5141 | 0.0058 | -37.5443 | -34.0539 | -2.2353 | -2.2402 |
|
| 69 |
+
| 0.433 | 1.56 | 600 | 0.6143 | 0.0129 | 0.0108 | 0.5245 | 0.0021 | -37.5011 | -34.0161 | -2.2421 | -2.2469 |
|
| 70 |
+
| 0.3298 | 1.82 | 700 | 0.5790 | 0.0633 | 0.0499 | 0.5195 | 0.0134 | -37.4453 | -33.9442 | -2.2401 | -2.2450 |
|
| 71 |
+
| 0.14 | 2.08 | 800 | 0.5904 | 0.0586 | 0.0544 | 0.5162 | 0.0041 | -37.4389 | -33.9509 | -2.2423 | -2.2472 |
|
| 72 |
+
| 0.2302 | 2.34 | 900 | 0.5758 | 0.0851 | 0.0740 | 0.5544 | 0.0111 | -37.4109 | -33.9130 | -2.2448 | -2.2497 |
|
| 73 |
+
| 0.2296 | 2.6 | 1000 | 0.5750 | 0.0631 | 0.0552 | 0.5075 | 0.0080 | -37.4378 | -33.9444 | -2.2440 | -2.2489 |
|
| 74 |
+
| 0.2798 | 2.86 | 1100 | 0.5483 | 0.0729 | 0.0545 | 0.5428 | 0.0184 | -37.4387 | -33.9303 | -2.2419 | -2.2468 |
|
| 75 |
+
| 0.1195 | 3.12 | 1200 | 0.5759 | 0.0672 | 0.0613 | 0.5137 | 0.0059 | -37.4291 | -33.9386 | -2.2424 | -2.2473 |
|
| 76 |
+
| 0.1371 | 3.38 | 1300 | 0.5592 | 0.0733 | 0.0574 | 0.5494 | 0.0159 | -37.4346 | -33.9299 | -2.2434 | -2.2483 |
|
| 77 |
+
| 0.0993 | 3.64 | 1400 | 0.6130 | 0.0546 | 0.0598 | 0.4871 | -0.0053 | -37.4311 | -33.9566 | -2.2422 | -2.2471 |
|
| 78 |
+
| 0.18 | 3.9 | 1500 | 0.5566 | 0.0778 | 0.0602 | 0.5050 | 0.0176 | -37.4306 | -33.9234 | -2.2423 | -2.2472 |
|
| 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 sha256:6237739007c13404c7f7a42da3b6e3f9e0db4cf6b72197b38009d5c4989014b6
|
| 3 |
size 176183216
|
all_results.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
"eval_logits/chosen": -2.236751079559326,
|
| 4 |
"eval_logits/rejected": -2.2318832874298096,
|
| 5 |
"eval_logps/chosen": -34.04561233520508,
|
|
@@ -13,9 +13,9 @@
|
|
| 13 |
"eval_samples": 343,
|
| 14 |
"eval_samples_per_second": 2.356,
|
| 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": -2.236751079559326,
|
| 4 |
"eval_logits/rejected": -2.2318832874298096,
|
| 5 |
"eval_logps/chosen": -34.04561233520508,
|
|
|
|
| 13 |
"eval_samples": 343,
|
| 14 |
"eval_samples_per_second": 2.356,
|
| 15 |
"eval_steps_per_second": 0.295,
|
| 16 |
+
"train_loss": 0.20226844080856868,
|
| 17 |
+
"train_runtime": 10768.956,
|
| 18 |
"train_samples": 3079,
|
| 19 |
+
"train_samples_per_second": 1.144,
|
| 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": 0.20226844080856868,
|
| 4 |
+
"train_runtime": 10768.956,
|
| 5 |
"train_samples": 3079,
|
| 6 |
+
"train_samples_per_second": 1.144,
|
| 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": 101.5,
|
| 496 |
+
"learning_rate": 4.84533120650964e-06,
|
| 497 |
+
"logits/chosen": -2.0717549324035645,
|
| 498 |
+
"logits/rejected": -2.058924913406372,
|
| 499 |
+
"logps/chosen": -32.21288299560547,
|
| 500 |
+
"logps/rejected": -32.803016662597656,
|
| 501 |
+
"loss": 0.2644,
|
| 502 |
+
"rewards/accuracies": 0.887499988079071,
|
| 503 |
+
"rewards/chosen": 0.22085602581501007,
|
| 504 |
+
"rewards/margins": 0.3342258632183075,
|
| 505 |
+
"rewards/rejected": -0.11336983740329742,
|
| 506 |
"step": 310
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.83,
|
| 510 |
+
"grad_norm": 28.625,
|
| 511 |
+
"learning_rate": 4.825108134172131e-06,
|
| 512 |
+
"logits/chosen": -1.98604416847229,
|
| 513 |
+
"logits/rejected": -1.9773375988006592,
|
| 514 |
+
"logps/chosen": -32.002044677734375,
|
| 515 |
+
"logps/rejected": -30.32120132446289,
|
| 516 |
+
"loss": 0.2537,
|
| 517 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 518 |
+
"rewards/chosen": 0.24706146121025085,
|
| 519 |
+
"rewards/margins": 0.3029271960258484,
|
| 520 |
+
"rewards/rejected": -0.055865705013275146,
|
| 521 |
"step": 320
|
| 522 |
},
|
| 523 |
{
|
| 524 |
"epoch": 0.86,
|
| 525 |
+
"grad_norm": 23.125,
|
| 526 |
+
"learning_rate": 4.80369052967602e-06,
|
| 527 |
+
"logits/chosen": -1.923841118812561,
|
| 528 |
+
"logits/rejected": -1.935959815979004,
|
| 529 |
+
"logps/chosen": -30.132516860961914,
|
| 530 |
+
"logps/rejected": -33.53127670288086,
|
| 531 |
+
"loss": 0.3016,
|
| 532 |
+
"rewards/accuracies": 0.824999988079071,
|
| 533 |
+
"rewards/chosen": 0.2427288293838501,
|
| 534 |
+
"rewards/margins": 0.3330441117286682,
|
| 535 |
+
"rewards/rejected": -0.0903153270483017,
|
| 536 |
"step": 330
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 0.88,
|
| 540 |
+
"grad_norm": 56.5,
|
| 541 |
+
"learning_rate": 4.781089396387968e-06,
|
| 542 |
+
"logits/chosen": -1.8918178081512451,
|
| 543 |
+
"logits/rejected": -1.8825798034667969,
|
| 544 |
+
"logps/chosen": -34.266075134277344,
|
| 545 |
+
"logps/rejected": -36.046722412109375,
|
| 546 |
+
"loss": 0.283,
|
| 547 |
+
"rewards/accuracies": 0.887499988079071,
|
| 548 |
+
"rewards/chosen": 0.3029957413673401,
|
| 549 |
+
"rewards/margins": 0.4359920620918274,
|
| 550 |
+
"rewards/rejected": -0.1329963356256485,
|
| 551 |
"step": 340
|
| 552 |
},
|
| 553 |
{
|
| 554 |
"epoch": 0.91,
|
| 555 |
+
"grad_norm": 44.0,
|
| 556 |
+
"learning_rate": 4.757316345716554e-06,
|
| 557 |
+
"logits/chosen": -1.950615644454956,
|
| 558 |
+
"logits/rejected": -1.951266884803772,
|
| 559 |
+
"logps/chosen": -34.006526947021484,
|
| 560 |
+
"logps/rejected": -33.97644805908203,
|
| 561 |
+
"loss": 0.2869,
|
| 562 |
+
"rewards/accuracies": 0.862500011920929,
|
| 563 |
+
"rewards/chosen": 0.2791282534599304,
|
| 564 |
+
"rewards/margins": 0.3683601915836334,
|
| 565 |
+
"rewards/rejected": -0.08923190832138062,
|
| 566 |
"step": 350
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 0.94,
|
| 570 |
+
"grad_norm": 54.25,
|
| 571 |
+
"learning_rate": 4.73238359114687e-06,
|
| 572 |
+
"logits/chosen": -2.07924222946167,
|
| 573 |
+
"logits/rejected": -2.0854721069335938,
|
| 574 |
+
"logps/chosen": -31.282617568969727,
|
| 575 |
+
"logps/rejected": -32.800941467285156,
|
| 576 |
+
"loss": 0.3061,
|
| 577 |
+
"rewards/accuracies": 0.800000011920929,
|
| 578 |
+
"rewards/chosen": 0.2034403532743454,
|
| 579 |
+
"rewards/margins": 0.296302855014801,
|
| 580 |
+
"rewards/rejected": -0.09286249428987503,
|
| 581 |
"step": 360
|
| 582 |
},
|
| 583 |
{
|
| 584 |
"epoch": 0.96,
|
| 585 |
+
"grad_norm": 172.0,
|
| 586 |
+
"learning_rate": 4.706303941965804e-06,
|
| 587 |
+
"logits/chosen": -2.011671781539917,
|
| 588 |
+
"logits/rejected": -2.0111806392669678,
|
| 589 |
+
"logps/chosen": -33.01143264770508,
|
| 590 |
+
"logps/rejected": -36.115821838378906,
|
| 591 |
+
"loss": 0.4197,
|
| 592 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 593 |
+
"rewards/chosen": 0.3034539818763733,
|
| 594 |
+
"rewards/margins": 0.349683940410614,
|
| 595 |
+
"rewards/rejected": -0.04622991383075714,
|
| 596 |
"step": 370
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 0.99,
|
| 600 |
+
"grad_norm": 32.5,
|
| 601 |
+
"learning_rate": 4.679090796681225e-06,
|
| 602 |
+
"logits/chosen": -2.047156572341919,
|
| 603 |
+
"logits/rejected": -2.04251766204834,
|
| 604 |
+
"logps/chosen": -30.326885223388672,
|
| 605 |
+
"logps/rejected": -29.325164794921875,
|
| 606 |
+
"loss": 0.2942,
|
| 607 |
+
"rewards/accuracies": 0.887499988079071,
|
| 608 |
+
"rewards/chosen": 0.26859989762306213,
|
| 609 |
+
"rewards/margins": 0.3464903235435486,
|
| 610 |
+
"rewards/rejected": -0.07789044082164764,
|
| 611 |
"step": 380
|
| 612 |
},
|
| 613 |
{
|
| 614 |
+
"epoch": 1.01,
|
| 615 |
+
"grad_norm": 52.25,
|
| 616 |
+
"learning_rate": 4.650758136138454e-06,
|
| 617 |
+
"logits/chosen": -1.8196346759796143,
|
| 618 |
+
"logits/rejected": -1.8260679244995117,
|
| 619 |
+
"logps/chosen": -32.0097541809082,
|
| 620 |
+
"logps/rejected": -36.38646697998047,
|
| 621 |
+
"loss": 0.4763,
|
| 622 |
+
"rewards/accuracies": 0.887499988079071,
|
| 623 |
+
"rewards/chosen": 0.26317089796066284,
|
| 624 |
+
"rewards/margins": 0.450895220041275,
|
| 625 |
+
"rewards/rejected": -0.1877242624759674,
|
| 626 |
+
"step": 390
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.04,
|
| 630 |
+
"grad_norm": 55.75,
|
| 631 |
+
"learning_rate": 4.621320516337559e-06,
|
| 632 |
+
"logits/chosen": -1.9777911901474,
|
| 633 |
+
"logits/rejected": -1.9714463949203491,
|
| 634 |
+
"logps/chosen": -33.3974609375,
|
| 635 |
+
"logps/rejected": -32.44751739501953,
|
| 636 |
+
"loss": 0.4001,
|
| 637 |
+
"rewards/accuracies": 0.8125,
|
| 638 |
+
"rewards/chosen": 0.3046652674674988,
|
| 639 |
+
"rewards/margins": 0.40171337127685547,
|
| 640 |
+
"rewards/rejected": -0.09704810380935669,
|
| 641 |
+
"step": 400
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 1.04,
|
| 645 |
+
"eval_logits/chosen": -2.2401604652404785,
|
| 646 |
+
"eval_logits/rejected": -2.235311508178711,
|
| 647 |
+
"eval_logps/chosen": -34.04499435424805,
|
| 648 |
+
"eval_logps/rejected": -37.51197052001953,
|
| 649 |
+
"eval_loss": 0.6349518299102783,
|
| 650 |
+
"eval_rewards/accuracies": 0.48380398750305176,
|
| 651 |
+
"eval_rewards/chosen": -0.00731015345081687,
|
| 652 |
+
"eval_rewards/margins": -0.010565591044723988,
|
| 653 |
+
"eval_rewards/rejected": 0.003255437593907118,
|
| 654 |
+
"eval_runtime": 145.4219,
|
| 655 |
+
"eval_samples_per_second": 2.359,
|
| 656 |
+
"eval_steps_per_second": 0.296,
|
| 657 |
+
"step": 400
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.06,
|
| 661 |
+
"grad_norm": 32.75,
|
| 662 |
+
"learning_rate": 4.590793060955158e-06,
|
| 663 |
+
"logits/chosen": -1.9763590097427368,
|
| 664 |
+
"logits/rejected": -1.9835665225982666,
|
| 665 |
+
"logps/chosen": -28.756824493408203,
|
| 666 |
+
"logps/rejected": -29.339588165283203,
|
| 667 |
+
"loss": 0.3007,
|
| 668 |
+
"rewards/accuracies": 0.9375,
|
| 669 |
+
"rewards/chosen": 0.2544856667518616,
|
| 670 |
+
"rewards/margins": 0.42344123125076294,
|
| 671 |
+
"rewards/rejected": -0.16895556449890137,
|
| 672 |
+
"step": 410
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.09,
|
| 676 |
+
"grad_norm": 83.0,
|
| 677 |
+
"learning_rate": 4.559191453574582e-06,
|
| 678 |
+
"logits/chosen": -1.9966684579849243,
|
| 679 |
+
"logits/rejected": -1.9955562353134155,
|
| 680 |
+
"logps/chosen": -33.83961868286133,
|
| 681 |
+
"logps/rejected": -31.041629791259766,
|
| 682 |
+
"loss": 0.3581,
|
| 683 |
+
"rewards/accuracies": 0.824999988079071,
|
| 684 |
+
"rewards/chosen": 0.28776243329048157,
|
| 685 |
+
"rewards/margins": 0.3567716181278229,
|
| 686 |
+
"rewards/rejected": -0.06900922954082489,
|
| 687 |
+
"step": 420
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 1.12,
|
| 691 |
+
"grad_norm": 96.0,
|
| 692 |
+
"learning_rate": 4.52653192962838e-06,
|
| 693 |
+
"logits/chosen": -1.9937080144882202,
|
| 694 |
+
"logits/rejected": -1.9763784408569336,
|
| 695 |
+
"logps/chosen": -30.642507553100586,
|
| 696 |
+
"logps/rejected": -33.117218017578125,
|
| 697 |
+
"loss": 0.5267,
|
| 698 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 699 |
+
"rewards/chosen": 0.2342870533466339,
|
| 700 |
+
"rewards/margins": 0.36793988943099976,
|
| 701 |
+
"rewards/rejected": -0.13365286588668823,
|
| 702 |
+
"step": 430
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 1.14,
|
| 706 |
+
"grad_norm": 43.75,
|
| 707 |
+
"learning_rate": 4.492831268057307e-06,
|
| 708 |
+
"logits/chosen": -2.0257625579833984,
|
| 709 |
+
"logits/rejected": -2.0276401042938232,
|
| 710 |
+
"logps/chosen": -36.22648239135742,
|
| 711 |
+
"logps/rejected": -35.008148193359375,
|
| 712 |
+
"loss": 0.361,
|
| 713 |
+
"rewards/accuracies": 0.875,
|
| 714 |
+
"rewards/chosen": 0.2701130509376526,
|
| 715 |
+
"rewards/margins": 0.37383943796157837,
|
| 716 |
+
"rewards/rejected": -0.10372640937566757,
|
| 717 |
+
"step": 440
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 1.17,
|
| 721 |
+
"grad_norm": 24.875,
|
| 722 |
+
"learning_rate": 4.458106782690094e-06,
|
| 723 |
+
"logits/chosen": -2.1039767265319824,
|
| 724 |
+
"logits/rejected": -2.103851079940796,
|
| 725 |
+
"logps/chosen": -32.201255798339844,
|
| 726 |
+
"logps/rejected": -33.596595764160156,
|
| 727 |
+
"loss": 0.3431,
|
| 728 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 729 |
+
"rewards/chosen": 0.30246877670288086,
|
| 730 |
+
"rewards/margins": 0.3280053436756134,
|
| 731 |
+
"rewards/rejected": -0.025536546483635902,
|
| 732 |
+
"step": 450
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 1.19,
|
| 736 |
+
"grad_norm": 73.0,
|
| 737 |
+
"learning_rate": 4.422376313348405e-06,
|
| 738 |
+
"logits/chosen": -2.0463452339172363,
|
| 739 |
+
"logits/rejected": -2.0388851165771484,
|
| 740 |
+
"logps/chosen": -32.02504348754883,
|
| 741 |
+
"logps/rejected": -36.26923370361328,
|
| 742 |
+
"loss": 0.504,
|
| 743 |
+
"rewards/accuracies": 0.800000011920929,
|
| 744 |
+
"rewards/chosen": 0.2666401267051697,
|
| 745 |
+
"rewards/margins": 0.41727694869041443,
|
| 746 |
+
"rewards/rejected": -0.15063682198524475,
|
| 747 |
+
"step": 460
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 1.22,
|
| 751 |
+
"grad_norm": 74.0,
|
| 752 |
+
"learning_rate": 4.3856582166815696e-06,
|
| 753 |
+
"logits/chosen": -1.9466873407363892,
|
| 754 |
+
"logits/rejected": -1.9430792331695557,
|
| 755 |
+
"logps/chosen": -33.49795913696289,
|
| 756 |
+
"logps/rejected": -33.04600143432617,
|
| 757 |
+
"loss": 0.2362,
|
| 758 |
+
"rewards/accuracies": 0.887499988079071,
|
| 759 |
+
"rewards/chosen": 0.3040338456630707,
|
| 760 |
+
"rewards/margins": 0.38951820135116577,
|
| 761 |
+
"rewards/rejected": -0.08548439294099808,
|
| 762 |
+
"step": 470
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 1.25,
|
| 766 |
+
"grad_norm": 74.0,
|
| 767 |
+
"learning_rate": 4.347971356735789e-06,
|
| 768 |
+
"logits/chosen": -2.0715551376342773,
|
| 769 |
+
"logits/rejected": -2.0645508766174316,
|
| 770 |
+
"logps/chosen": -30.481578826904297,
|
| 771 |
+
"logps/rejected": -32.14142990112305,
|
| 772 |
+
"loss": 0.4149,
|
| 773 |
+
"rewards/accuracies": 0.800000011920929,
|
| 774 |
+
"rewards/chosen": 0.242868572473526,
|
| 775 |
+
"rewards/margins": 0.3464292287826538,
|
| 776 |
+
"rewards/rejected": -0.10356064140796661,
|
| 777 |
+
"step": 480
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 1.27,
|
| 781 |
+
"grad_norm": 97.5,
|
| 782 |
+
"learning_rate": 4.309335095262675e-06,
|
| 783 |
+
"logits/chosen": -2.018632411956787,
|
| 784 |
+
"logits/rejected": -2.020221710205078,
|
| 785 |
+
"logps/chosen": -34.90156936645508,
|
| 786 |
+
"logps/rejected": -33.80845260620117,
|
| 787 |
+
"loss": 0.3635,
|
| 788 |
+
"rewards/accuracies": 0.824999988079071,
|
| 789 |
+
"rewards/chosen": 0.3161340653896332,
|
| 790 |
+
"rewards/margins": 0.4329928457736969,
|
| 791 |
+
"rewards/rejected": -0.11685874313116074,
|
| 792 |
+
"step": 490
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 1.3,
|
| 796 |
+
"grad_norm": 64.0,
|
| 797 |
+
"learning_rate": 4.269769281772082e-06,
|
| 798 |
+
"logits/chosen": -1.9072506427764893,
|
| 799 |
+
"logits/rejected": -1.9048874378204346,
|
| 800 |
+
"logps/chosen": -32.74415588378906,
|
| 801 |
+
"logps/rejected": -36.70649337768555,
|
| 802 |
+
"loss": 0.3401,
|
| 803 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 804 |
+
"rewards/chosen": 0.2996928095817566,
|
| 805 |
+
"rewards/margins": 0.469214528799057,
|
| 806 |
+
"rewards/rejected": -0.16952167451381683,
|
| 807 |
+
"step": 500
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 1.3,
|
| 811 |
+
"eval_logits/chosen": -2.2401528358459473,
|
| 812 |
+
"eval_logits/rejected": -2.235276937484741,
|
| 813 |
+
"eval_logps/chosen": -34.05385971069336,
|
| 814 |
+
"eval_logps/rejected": -37.54425811767578,
|
| 815 |
+
"eval_loss": 0.623750627040863,
|
| 816 |
+
"eval_rewards/accuracies": 0.5141196250915527,
|
| 817 |
+
"eval_rewards/chosen": -0.013515940867364407,
|
| 818 |
+
"eval_rewards/margins": 0.005829531699419022,
|
| 819 |
+
"eval_rewards/rejected": -0.019345473498106003,
|
| 820 |
+
"eval_runtime": 145.1627,
|
| 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": 43.75,
|
| 828 |
+
"learning_rate": 4.22929424333435e-06,
|
| 829 |
+
"logits/chosen": -2.0122382640838623,
|
| 830 |
+
"logits/rejected": -2.0169613361358643,
|
| 831 |
+
"logps/chosen": -33.01191329956055,
|
| 832 |
+
"logps/rejected": -31.77316665649414,
|
| 833 |
+
"loss": 0.3867,
|
| 834 |
+
"rewards/accuracies": 0.800000011920929,
|
| 835 |
+
"rewards/chosen": 0.31180983781814575,
|
| 836 |
+
"rewards/margins": 0.40075111389160156,
|
| 837 |
+
"rewards/rejected": -0.08894126117229462,
|
| 838 |
+
"step": 510
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.35,
|
| 842 |
+
"grad_norm": 47.0,
|
| 843 |
+
"learning_rate": 4.1879307741372085e-06,
|
| 844 |
+
"logits/chosen": -2.045281171798706,
|
| 845 |
+
"logits/rejected": -2.0561985969543457,
|
| 846 |
+
"logps/chosen": -30.94211196899414,
|
| 847 |
+
"logps/rejected": -31.70987892150879,
|
| 848 |
+
"loss": 0.4115,
|
| 849 |
+
"rewards/accuracies": 0.862500011920929,
|
| 850 |
+
"rewards/chosen": 0.31381234526634216,
|
| 851 |
+
"rewards/margins": 0.455346018075943,
|
| 852 |
+
"rewards/rejected": -0.14153364300727844,
|
| 853 |
+
"step": 520
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.38,
|
| 857 |
+
"grad_norm": 69.0,
|
| 858 |
+
"learning_rate": 4.145700124802693e-06,
|
| 859 |
+
"logits/chosen": -1.9737331867218018,
|
| 860 |
+
"logits/rejected": -1.9702527523040771,
|
| 861 |
+
"logps/chosen": -31.990686416625977,
|
| 862 |
+
"logps/rejected": -32.61437225341797,
|
| 863 |
+
"loss": 0.4131,
|
| 864 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 865 |
+
"rewards/chosen": 0.3008785843849182,
|
| 866 |
+
"rewards/margins": 0.4769964814186096,
|
| 867 |
+
"rewards/rejected": -0.1761178821325302,
|
| 868 |
+
"step": 530
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 1.4,
|
| 872 |
+
"grad_norm": 52.75,
|
| 873 |
+
"learning_rate": 4.102623991469562e-06,
|
| 874 |
+
"logits/chosen": -1.8443644046783447,
|
| 875 |
+
"logits/rejected": -1.853546142578125,
|
| 876 |
+
"logps/chosen": -32.251625061035156,
|
| 877 |
+
"logps/rejected": -32.053138732910156,
|
| 878 |
+
"loss": 0.3856,
|
| 879 |
+
"rewards/accuracies": 0.824999988079071,
|
| 880 |
+
"rewards/chosen": 0.22243666648864746,
|
| 881 |
+
"rewards/margins": 0.3604408800601959,
|
| 882 |
+
"rewards/rejected": -0.13800422847270966,
|
| 883 |
+
"step": 540
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.43,
|
| 887 |
+
"grad_norm": 58.75,
|
| 888 |
+
"learning_rate": 4.058724504646834e-06,
|
| 889 |
+
"logits/chosen": -1.9435665607452393,
|
| 890 |
+
"logits/rejected": -1.9371446371078491,
|
| 891 |
+
"logps/chosen": -33.34047317504883,
|
| 892 |
+
"logps/rejected": -31.121173858642578,
|
| 893 |
+
"loss": 0.3465,
|
| 894 |
+
"rewards/accuracies": 0.8125,
|
| 895 |
+
"rewards/chosen": 0.2574843168258667,
|
| 896 |
+
"rewards/margins": 0.34836047887802124,
|
| 897 |
+
"rewards/rejected": -0.09087616950273514,
|
| 898 |
+
"step": 550
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 1.45,
|
| 902 |
+
"grad_norm": 135.0,
|
| 903 |
+
"learning_rate": 4.014024217844167e-06,
|
| 904 |
+
"logits/chosen": -2.0361862182617188,
|
| 905 |
+
"logits/rejected": -2.0341010093688965,
|
| 906 |
+
"logps/chosen": -34.12422561645508,
|
| 907 |
+
"logps/rejected": -31.634075164794922,
|
| 908 |
+
"loss": 0.6034,
|
| 909 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 910 |
+
"rewards/chosen": 0.3128809928894043,
|
| 911 |
+
"rewards/margins": 0.40856558084487915,
|
| 912 |
+
"rewards/rejected": -0.09568462520837784,
|
| 913 |
+
"step": 560
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.48,
|
| 917 |
+
"grad_norm": 52.75,
|
| 918 |
+
"learning_rate": 3.968546095984911e-06,
|
| 919 |
+
"logits/chosen": -1.8710432052612305,
|
| 920 |
+
"logits/rejected": -1.8688808679580688,
|
| 921 |
+
"logps/chosen": -32.49556350708008,
|
| 922 |
+
"logps/rejected": -31.1130428314209,
|
| 923 |
+
"loss": 0.4937,
|
| 924 |
+
"rewards/accuracies": 0.8125,
|
| 925 |
+
"rewards/chosen": 0.21059513092041016,
|
| 926 |
+
"rewards/margins": 0.2881118357181549,
|
| 927 |
+
"rewards/rejected": -0.07751675695180893,
|
| 928 |
+
"step": 570
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 1.51,
|
| 932 |
+
"grad_norm": 95.5,
|
| 933 |
+
"learning_rate": 3.922313503607806e-06,
|
| 934 |
+
"logits/chosen": -2.005981922149658,
|
| 935 |
+
"logits/rejected": -2.0027260780334473,
|
| 936 |
+
"logps/chosen": -30.789072036743164,
|
| 937 |
+
"logps/rejected": -34.820518493652344,
|
| 938 |
+
"loss": 0.4043,
|
| 939 |
+
"rewards/accuracies": 0.800000011920929,
|
| 940 |
+
"rewards/chosen": 0.22090163826942444,
|
| 941 |
+
"rewards/margins": 0.356947124004364,
|
| 942 |
+
"rewards/rejected": -0.13604548573493958,
|
| 943 |
+
"step": 580
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.53,
|
| 947 |
+
"grad_norm": 20.25,
|
| 948 |
+
"learning_rate": 3.875350192863368e-06,
|
| 949 |
+
"logits/chosen": -1.948760986328125,
|
| 950 |
+
"logits/rejected": -1.9524637460708618,
|
| 951 |
+
"logps/chosen": -29.25394630432129,
|
| 952 |
+
"logps/rejected": -30.857736587524414,
|
| 953 |
+
"loss": 0.3681,
|
| 954 |
+
"rewards/accuracies": 0.8125,
|
| 955 |
+
"rewards/chosen": 0.2527596354484558,
|
| 956 |
+
"rewards/margins": 0.3919641375541687,
|
| 957 |
+
"rewards/rejected": -0.1392044574022293,
|
| 958 |
+
"step": 590
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 1.56,
|
| 962 |
+
"grad_norm": 28.125,
|
| 963 |
+
"learning_rate": 3.8276802913111436e-06,
|
| 964 |
+
"logits/chosen": -1.9878416061401367,
|
| 965 |
+
"logits/rejected": -1.9875671863555908,
|
| 966 |
+
"logps/chosen": -31.794857025146484,
|
| 967 |
+
"logps/rejected": -31.33709716796875,
|
| 968 |
+
"loss": 0.433,
|
| 969 |
+
"rewards/accuracies": 0.887499988079071,
|
| 970 |
+
"rewards/chosen": 0.2986726462841034,
|
| 971 |
+
"rewards/margins": 0.3995324373245239,
|
| 972 |
+
"rewards/rejected": -0.10085982084274292,
|
| 973 |
+
"step": 600
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.56,
|
| 977 |
+
"eval_logits/chosen": -2.2469263076782227,
|
| 978 |
+
"eval_logits/rejected": -2.2420639991760254,
|
| 979 |
+
"eval_logps/chosen": -34.01606369018555,
|
| 980 |
+
"eval_logps/rejected": -37.50114822387695,
|
| 981 |
+
"eval_loss": 0.6143299341201782,
|
| 982 |
+
"eval_rewards/accuracies": 0.5245016813278198,
|
| 983 |
+
"eval_rewards/chosen": 0.012940882705152035,
|
| 984 |
+
"eval_rewards/margins": 0.002107805572450161,
|
| 985 |
+
"eval_rewards/rejected": 0.010833078064024448,
|
| 986 |
+
"eval_runtime": 145.3485,
|
| 987 |
+
"eval_samples_per_second": 2.36,
|
| 988 |
+
"eval_steps_per_second": 0.296,
|
| 989 |
+
"step": 600
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.58,
|
| 993 |
+
"grad_norm": 77.5,
|
| 994 |
+
"learning_rate": 3.7793282895240927e-06,
|
| 995 |
+
"logits/chosen": -2.05083966255188,
|
| 996 |
+
"logits/rejected": -2.0515644550323486,
|
| 997 |
+
"logps/chosen": -34.6113395690918,
|
| 998 |
+
"logps/rejected": -33.167572021484375,
|
| 999 |
+
"loss": 0.3998,
|
| 1000 |
+
"rewards/accuracies": 0.8125,
|
| 1001 |
+
"rewards/chosen": 0.29316502809524536,
|
| 1002 |
+
"rewards/margins": 0.3515344560146332,
|
| 1003 |
+
"rewards/rejected": -0.058369409292936325,
|
| 1004 |
+
"step": 610
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 1.61,
|
| 1008 |
+
"grad_norm": 91.5,
|
| 1009 |
+
"learning_rate": 3.730319028506478e-06,
|
| 1010 |
+
"logits/chosen": -2.01389741897583,
|
| 1011 |
+
"logits/rejected": -2.0112407207489014,
|
| 1012 |
+
"logps/chosen": -32.689491271972656,
|
| 1013 |
+
"logps/rejected": -32.25727081298828,
|
| 1014 |
+
"loss": 0.3825,
|
| 1015 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1016 |
+
"rewards/chosen": 0.3051331341266632,
|
| 1017 |
+
"rewards/margins": 0.4362831711769104,
|
| 1018 |
+
"rewards/rejected": -0.1311500072479248,
|
| 1019 |
+
"step": 620
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.64,
|
| 1023 |
+
"grad_norm": 39.25,
|
| 1024 |
+
"learning_rate": 3.6806776869317074e-06,
|
| 1025 |
+
"logits/chosen": -2.028050184249878,
|
| 1026 |
+
"logits/rejected": -2.018892765045166,
|
| 1027 |
+
"logps/chosen": -32.39860153198242,
|
| 1028 |
+
"logps/rejected": -30.971405029296875,
|
| 1029 |
+
"loss": 0.5713,
|
| 1030 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1031 |
+
"rewards/chosen": 0.26426073908805847,
|
| 1032 |
+
"rewards/margins": 0.28948652744293213,
|
| 1033 |
+
"rewards/rejected": -0.02522580698132515,
|
| 1034 |
+
"step": 630
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 1.66,
|
| 1038 |
+
"grad_norm": 32.25,
|
| 1039 |
+
"learning_rate": 3.6304297682067146e-06,
|
| 1040 |
+
"logits/chosen": -2.0308046340942383,
|
| 1041 |
+
"logits/rejected": -2.027677059173584,
|
| 1042 |
+
"logps/chosen": -31.706838607788086,
|
| 1043 |
+
"logps/rejected": -32.60099792480469,
|
| 1044 |
+
"loss": 0.2067,
|
| 1045 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1046 |
+
"rewards/chosen": 0.3441086411476135,
|
| 1047 |
+
"rewards/margins": 0.4712149500846863,
|
| 1048 |
+
"rewards/rejected": -0.12710630893707275,
|
| 1049 |
+
"step": 640
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 1.69,
|
| 1053 |
+
"grad_norm": 27.875,
|
| 1054 |
+
"learning_rate": 3.579601087369492e-06,
|
| 1055 |
+
"logits/chosen": -2.03548264503479,
|
| 1056 |
+
"logits/rejected": -2.0378546714782715,
|
| 1057 |
+
"logps/chosen": -33.151695251464844,
|
| 1058 |
+
"logps/rejected": -33.8459587097168,
|
| 1059 |
+
"loss": 0.3123,
|
| 1060 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1061 |
+
"rewards/chosen": 0.3514065742492676,
|
| 1062 |
+
"rewards/margins": 0.465198814868927,
|
| 1063 |
+
"rewards/rejected": -0.11379221826791763,
|
| 1064 |
+
"step": 650
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 1.71,
|
| 1068 |
+
"grad_norm": 47.5,
|
| 1069 |
+
"learning_rate": 3.5282177578265295e-06,
|
| 1070 |
+
"logits/chosen": -1.9493051767349243,
|
| 1071 |
+
"logits/rejected": -1.9498169422149658,
|
| 1072 |
+
"logps/chosen": -33.52117156982422,
|
| 1073 |
+
"logps/rejected": -31.683589935302734,
|
| 1074 |
+
"loss": 0.3393,
|
| 1075 |
+
"rewards/accuracies": 0.8125,
|
| 1076 |
+
"rewards/chosen": 0.2979589104652405,
|
| 1077 |
+
"rewards/margins": 0.35139504075050354,
|
| 1078 |
+
"rewards/rejected": -0.053436122834682465,
|
| 1079 |
+
"step": 660
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 1.74,
|
| 1083 |
+
"grad_norm": 60.25,
|
| 1084 |
+
"learning_rate": 3.476306177936961e-06,
|
| 1085 |
+
"logits/chosen": -1.9889039993286133,
|
| 1086 |
+
"logits/rejected": -1.9790817499160767,
|
| 1087 |
+
"logps/chosen": -33.58037567138672,
|
| 1088 |
+
"logps/rejected": -32.429481506347656,
|
| 1089 |
+
"loss": 0.2419,
|
| 1090 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1091 |
+
"rewards/chosen": 0.2946351170539856,
|
| 1092 |
+
"rewards/margins": 0.37981870770454407,
|
| 1093 |
+
"rewards/rejected": -0.08518362790346146,
|
| 1094 |
+
"step": 670
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 1.77,
|
| 1098 |
+
"grad_norm": 58.75,
|
| 1099 |
+
"learning_rate": 3.423893017450324e-06,
|
| 1100 |
+
"logits/chosen": -1.8874107599258423,
|
| 1101 |
+
"logits/rejected": -1.8842309713363647,
|
| 1102 |
+
"logps/chosen": -30.960819244384766,
|
| 1103 |
+
"logps/rejected": -34.40943145751953,
|
| 1104 |
+
"loss": 0.2598,
|
| 1105 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1106 |
+
"rewards/chosen": 0.35009804368019104,
|
| 1107 |
+
"rewards/margins": 0.41369539499282837,
|
| 1108 |
+
"rewards/rejected": -0.06359735876321793,
|
| 1109 |
+
"step": 680
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.79,
|
| 1113 |
+
"grad_norm": 114.5,
|
| 1114 |
+
"learning_rate": 3.3710052038048794e-06,
|
| 1115 |
+
"logits/chosen": -1.944596290588379,
|
| 1116 |
+
"logits/rejected": -1.9437839984893799,
|
| 1117 |
+
"logps/chosen": -34.76592254638672,
|
| 1118 |
+
"logps/rejected": -35.450462341308594,
|
| 1119 |
+
"loss": 0.3155,
|
| 1120 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1121 |
+
"rewards/chosen": 0.3530573844909668,
|
| 1122 |
+
"rewards/margins": 0.42070499062538147,
|
| 1123 |
+
"rewards/rejected": -0.06764759868383408,
|
| 1124 |
+
"step": 690
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 1.82,
|
| 1128 |
+
"grad_norm": 35.25,
|
| 1129 |
+
"learning_rate": 3.3176699082935546e-06,
|
| 1130 |
+
"logits/chosen": -1.9258012771606445,
|
| 1131 |
+
"logits/rejected": -1.928860068321228,
|
| 1132 |
+
"logps/chosen": -32.108642578125,
|
| 1133 |
+
"logps/rejected": -35.734214782714844,
|
| 1134 |
+
"loss": 0.3298,
|
| 1135 |
+
"rewards/accuracies": 0.875,
|
| 1136 |
+
"rewards/chosen": 0.3459445536136627,
|
| 1137 |
+
"rewards/margins": 0.4107227325439453,
|
| 1138 |
+
"rewards/rejected": -0.06477821618318558,
|
| 1139 |
+
"step": 700
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 1.82,
|
| 1143 |
+
"eval_logits/chosen": -2.244976043701172,
|
| 1144 |
+
"eval_logits/rejected": -2.240100622177124,
|
| 1145 |
+
"eval_logps/chosen": -33.94416046142578,
|
| 1146 |
+
"eval_logps/rejected": -37.445308685302734,
|
| 1147 |
+
"eval_loss": 0.579022228717804,
|
| 1148 |
+
"eval_rewards/accuracies": 0.5195183157920837,
|
| 1149 |
+
"eval_rewards/chosen": 0.06327207386493683,
|
| 1150 |
+
"eval_rewards/margins": 0.013353652320802212,
|
| 1151 |
+
"eval_rewards/rejected": 0.049918416887521744,
|
| 1152 |
+
"eval_runtime": 145.2212,
|
| 1153 |
+
"eval_samples_per_second": 2.362,
|
| 1154 |
+
"eval_steps_per_second": 0.296,
|
| 1155 |
+
"step": 700
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 1.84,
|
| 1159 |
+
"grad_norm": 43.5,
|
| 1160 |
+
"learning_rate": 3.2639145321045933e-06,
|
| 1161 |
+
"logits/chosen": -2.0379347801208496,
|
| 1162 |
+
"logits/rejected": -2.0407185554504395,
|
| 1163 |
+
"logps/chosen": -34.414249420166016,
|
| 1164 |
+
"logps/rejected": -34.340797424316406,
|
| 1165 |
+
"loss": 0.4057,
|
| 1166 |
+
"rewards/accuracies": 0.8125,
|
| 1167 |
+
"rewards/chosen": 0.3369380831718445,
|
| 1168 |
+
"rewards/margins": 0.38638633489608765,
|
| 1169 |
+
"rewards/rejected": -0.04944825917482376,
|
| 1170 |
+
"step": 710
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 1.87,
|
| 1174 |
+
"grad_norm": 105.5,
|
| 1175 |
+
"learning_rate": 3.2097666922441107e-06,
|
| 1176 |
+
"logits/chosen": -1.8936197757720947,
|
| 1177 |
+
"logits/rejected": -1.8877111673355103,
|
| 1178 |
+
"logps/chosen": -34.38829040527344,
|
| 1179 |
+
"logps/rejected": -32.45317459106445,
|
| 1180 |
+
"loss": 0.3661,
|
| 1181 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1182 |
+
"rewards/chosen": 0.31145918369293213,
|
| 1183 |
+
"rewards/margins": 0.3823915123939514,
|
| 1184 |
+
"rewards/rejected": -0.0709322914481163,
|
| 1185 |
+
"step": 720
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 1.9,
|
| 1189 |
+
"grad_norm": 34.75,
|
| 1190 |
+
"learning_rate": 3.1552542073477554e-06,
|
| 1191 |
+
"logits/chosen": -2.068476438522339,
|
| 1192 |
+
"logits/rejected": -2.065448045730591,
|
| 1193 |
+
"logps/chosen": -30.17218017578125,
|
| 1194 |
+
"logps/rejected": -31.612863540649414,
|
| 1195 |
+
"loss": 0.3787,
|
| 1196 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1197 |
+
"rewards/chosen": 0.3365642726421356,
|
| 1198 |
+
"rewards/margins": 0.45183300971984863,
|
| 1199 |
+
"rewards/rejected": -0.11526870727539062,
|
| 1200 |
+
"step": 730
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.92,
|
| 1204 |
+
"grad_norm": 77.0,
|
| 1205 |
+
"learning_rate": 3.100405083388799e-06,
|
| 1206 |
+
"logits/chosen": -1.911435842514038,
|
| 1207 |
+
"logits/rejected": -1.9112478494644165,
|
| 1208 |
+
"logps/chosen": -32.99172592163086,
|
| 1209 |
+
"logps/rejected": -37.709041595458984,
|
| 1210 |
+
"loss": 0.3501,
|
| 1211 |
+
"rewards/accuracies": 0.8125,
|
| 1212 |
+
"rewards/chosen": 0.3300129473209381,
|
| 1213 |
+
"rewards/margins": 0.3868759870529175,
|
| 1214 |
+
"rewards/rejected": -0.05686303228139877,
|
| 1215 |
+
"step": 740
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 1.95,
|
| 1219 |
+
"grad_norm": 93.0,
|
| 1220 |
+
"learning_rate": 3.0452474992899645e-06,
|
| 1221 |
+
"logits/chosen": -1.7933385372161865,
|
| 1222 |
+
"logits/rejected": -1.7985591888427734,
|
| 1223 |
+
"logps/chosen": -36.41147994995117,
|
| 1224 |
+
"logps/rejected": -33.99577713012695,
|
| 1225 |
+
"loss": 0.3683,
|
| 1226 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 1227 |
+
"rewards/chosen": 0.364685595035553,
|
| 1228 |
+
"rewards/margins": 0.38109368085861206,
|
| 1229 |
+
"rewards/rejected": -0.016408095136284828,
|
| 1230 |
+
"step": 750
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 1.97,
|
| 1234 |
+
"grad_norm": 111.5,
|
| 1235 |
+
"learning_rate": 2.989809792446417e-06,
|
| 1236 |
+
"logits/chosen": -2.0044448375701904,
|
| 1237 |
+
"logits/rejected": -2.0057899951934814,
|
| 1238 |
+
"logps/chosen": -32.149024963378906,
|
| 1239 |
+
"logps/rejected": -32.881797790527344,
|
| 1240 |
+
"loss": 0.3432,
|
| 1241 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1242 |
+
"rewards/chosen": 0.3196606934070587,
|
| 1243 |
+
"rewards/margins": 0.39146822690963745,
|
| 1244 |
+
"rewards/rejected": -0.07180755585432053,
|
| 1245 |
+
"step": 760
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 2.0,
|
| 1249 |
+
"grad_norm": 71.0,
|
| 1250 |
+
"learning_rate": 2.9341204441673267e-06,
|
| 1251 |
+
"logits/chosen": -1.9735772609710693,
|
| 1252 |
+
"logits/rejected": -1.9727319478988647,
|
| 1253 |
+
"logps/chosen": -31.752365112304688,
|
| 1254 |
+
"logps/rejected": -35.02272033691406,
|
| 1255 |
+
"loss": 0.3104,
|
| 1256 |
+
"rewards/accuracies": 0.8125,
|
| 1257 |
+
"rewards/chosen": 0.2962704002857208,
|
| 1258 |
+
"rewards/margins": 0.29246559739112854,
|
| 1259 |
+
"rewards/rejected": 0.0038047723937779665,
|
| 1260 |
+
"step": 770
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 2.03,
|
| 1264 |
+
"grad_norm": 27.0,
|
| 1265 |
+
"learning_rate": 2.878208065043501e-06,
|
| 1266 |
+
"logits/chosen": -1.9740774631500244,
|
| 1267 |
+
"logits/rejected": -1.9732996225357056,
|
| 1268 |
+
"logps/chosen": -34.172630310058594,
|
| 1269 |
+
"logps/rejected": -31.525737762451172,
|
| 1270 |
+
"loss": 0.2392,
|
| 1271 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1272 |
+
"rewards/chosen": 0.33308273553848267,
|
| 1273 |
+
"rewards/margins": 0.40743428468704224,
|
| 1274 |
+
"rewards/rejected": -0.07435150444507599,
|
| 1275 |
+
"step": 780
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 2.05,
|
| 1279 |
+
"grad_norm": 32.0,
|
| 1280 |
+
"learning_rate": 2.8221013802485974e-06,
|
| 1281 |
+
"logits/chosen": -2.00411057472229,
|
| 1282 |
+
"logits/rejected": -2.0026121139526367,
|
| 1283 |
+
"logps/chosen": -29.090576171875,
|
| 1284 |
+
"logps/rejected": -32.80150604248047,
|
| 1285 |
+
"loss": 0.166,
|
| 1286 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1287 |
+
"rewards/chosen": 0.3308223783969879,
|
| 1288 |
+
"rewards/margins": 0.4464591145515442,
|
| 1289 |
+
"rewards/rejected": -0.11563672870397568,
|
| 1290 |
+
"step": 790
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 2.08,
|
| 1294 |
+
"grad_norm": 28.625,
|
| 1295 |
+
"learning_rate": 2.76582921478147e-06,
|
| 1296 |
+
"logits/chosen": -2.0544497966766357,
|
| 1297 |
+
"logits/rejected": -2.05083966255188,
|
| 1298 |
+
"logps/chosen": -31.90934181213379,
|
| 1299 |
+
"logps/rejected": -34.358177185058594,
|
| 1300 |
+
"loss": 0.14,
|
| 1301 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1302 |
+
"rewards/chosen": 0.36250725388526917,
|
| 1303 |
+
"rewards/margins": 0.4739067554473877,
|
| 1304 |
+
"rewards/rejected": -0.11139953136444092,
|
| 1305 |
+
"step": 800
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 2.08,
|
| 1309 |
+
"eval_logits/chosen": -2.247178077697754,
|
| 1310 |
+
"eval_logits/rejected": -2.2423007488250732,
|
| 1311 |
+
"eval_logps/chosen": -33.95090103149414,
|
| 1312 |
+
"eval_logps/rejected": -37.43887710571289,
|
| 1313 |
+
"eval_loss": 0.5903980731964111,
|
| 1314 |
+
"eval_rewards/accuracies": 0.5161960124969482,
|
| 1315 |
+
"eval_rewards/chosen": 0.05855472758412361,
|
| 1316 |
+
"eval_rewards/margins": 0.0041350084356963634,
|
| 1317 |
+
"eval_rewards/rejected": 0.05441971868276596,
|
| 1318 |
+
"eval_runtime": 145.1787,
|
| 1319 |
+
"eval_samples_per_second": 2.363,
|
| 1320 |
+
"eval_steps_per_second": 0.296,
|
| 1321 |
+
"step": 800
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 2.1,
|
| 1325 |
+
"grad_norm": 104.0,
|
| 1326 |
+
"learning_rate": 2.7094204786572254e-06,
|
| 1327 |
+
"logits/chosen": -1.8908456563949585,
|
| 1328 |
+
"logits/rejected": -1.8829033374786377,
|
| 1329 |
+
"logps/chosen": -33.954002380371094,
|
| 1330 |
+
"logps/rejected": -34.61323165893555,
|
| 1331 |
+
"loss": 0.3154,
|
| 1332 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 1333 |
+
"rewards/chosen": 0.3515668511390686,
|
| 1334 |
+
"rewards/margins": 0.38875049352645874,
|
| 1335 |
+
"rewards/rejected": -0.03718366473913193,
|
| 1336 |
+
"step": 810
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.13,
|
| 1340 |
+
"grad_norm": 43.75,
|
| 1341 |
+
"learning_rate": 2.6529041520546072e-06,
|
| 1342 |
+
"logits/chosen": -1.9656623601913452,
|
| 1343 |
+
"logits/rejected": -1.976292371749878,
|
| 1344 |
+
"logps/chosen": -35.201332092285156,
|
| 1345 |
+
"logps/rejected": -32.56475067138672,
|
| 1346 |
+
"loss": 0.2191,
|
| 1347 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1348 |
+
"rewards/chosen": 0.3460533618927002,
|
| 1349 |
+
"rewards/margins": 0.4393738806247711,
|
| 1350 |
+
"rewards/rejected": -0.09332051128149033,
|
| 1351 |
+
"step": 820
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 2.16,
|
| 1355 |
+
"grad_norm": 58.0,
|
| 1356 |
+
"learning_rate": 2.5963092704273302e-06,
|
| 1357 |
+
"logits/chosen": -2.0145046710968018,
|
| 1358 |
+
"logits/rejected": -2.018724203109741,
|
| 1359 |
+
"logps/chosen": -34.39170455932617,
|
| 1360 |
+
"logps/rejected": -29.336902618408203,
|
| 1361 |
+
"loss": 0.196,
|
| 1362 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1363 |
+
"rewards/chosen": 0.357754647731781,
|
| 1364 |
+
"rewards/margins": 0.4058148264884949,
|
| 1365 |
+
"rewards/rejected": -0.04806024581193924,
|
| 1366 |
+
"step": 830
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 2.18,
|
| 1370 |
+
"grad_norm": 28.5,
|
| 1371 |
+
"learning_rate": 2.53966490958702e-06,
|
| 1372 |
+
"logits/chosen": -2.012115240097046,
|
| 1373 |
+
"logits/rejected": -2.0198771953582764,
|
| 1374 |
+
"logps/chosen": -34.340667724609375,
|
| 1375 |
+
"logps/rejected": -29.972625732421875,
|
| 1376 |
+
"loss": 0.375,
|
| 1377 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1378 |
+
"rewards/chosen": 0.3131526708602905,
|
| 1379 |
+
"rewards/margins": 0.40060001611709595,
|
| 1380 |
+
"rewards/rejected": -0.08744733780622482,
|
| 1381 |
+
"step": 840
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 2.21,
|
| 1385 |
+
"grad_norm": 48.75,
|
| 1386 |
+
"learning_rate": 2.4830001707654135e-06,
|
| 1387 |
+
"logits/chosen": -1.942496657371521,
|
| 1388 |
+
"logits/rejected": -1.932924509048462,
|
| 1389 |
+
"logps/chosen": -31.318111419677734,
|
| 1390 |
+
"logps/rejected": -32.194252014160156,
|
| 1391 |
+
"loss": 0.2437,
|
| 1392 |
+
"rewards/accuracies": 0.9375,
|
| 1393 |
+
"rewards/chosen": 0.3325595259666443,
|
| 1394 |
+
"rewards/margins": 0.4588742256164551,
|
| 1395 |
+
"rewards/rejected": -0.1263146996498108,
|
| 1396 |
+
"step": 850
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 2.23,
|
| 1400 |
+
"grad_norm": 57.0,
|
| 1401 |
+
"learning_rate": 2.4263441656635054e-06,
|
| 1402 |
+
"logits/chosen": -2.0852370262145996,
|
| 1403 |
+
"logits/rejected": -2.075453281402588,
|
| 1404 |
+
"logps/chosen": -25.809616088867188,
|
| 1405 |
+
"logps/rejected": -30.046661376953125,
|
| 1406 |
+
"loss": 0.2,
|
| 1407 |
+
"rewards/accuracies": 0.875,
|
| 1408 |
+
"rewards/chosen": 0.29011884331703186,
|
| 1409 |
+
"rewards/margins": 0.4121415615081787,
|
| 1410 |
+
"rewards/rejected": -0.12202271074056625,
|
| 1411 |
+
"step": 860
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 2.26,
|
| 1415 |
+
"grad_norm": 62.0,
|
| 1416 |
+
"learning_rate": 2.3697260014953107e-06,
|
| 1417 |
+
"logits/chosen": -1.9381517171859741,
|
| 1418 |
+
"logits/rejected": -1.9393552541732788,
|
| 1419 |
+
"logps/chosen": -33.79442596435547,
|
| 1420 |
+
"logps/rejected": -30.08365821838379,
|
| 1421 |
+
"loss": 0.3179,
|
| 1422 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1423 |
+
"rewards/chosen": 0.37700217962265015,
|
| 1424 |
+
"rewards/margins": 0.4478107988834381,
|
| 1425 |
+
"rewards/rejected": -0.07080861181020737,
|
| 1426 |
+
"step": 870
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 2.29,
|
| 1430 |
+
"grad_norm": 25.25,
|
| 1431 |
+
"learning_rate": 2.3131747660339396e-06,
|
| 1432 |
+
"logits/chosen": -1.9556992053985596,
|
| 1433 |
+
"logits/rejected": -1.9564440250396729,
|
| 1434 |
+
"logps/chosen": -31.94332504272461,
|
| 1435 |
+
"logps/rejected": -33.166770935058594,
|
| 1436 |
+
"loss": 0.2801,
|
| 1437 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1438 |
+
"rewards/chosen": 0.3121483623981476,
|
| 1439 |
+
"rewards/margins": 0.3932458460330963,
|
| 1440 |
+
"rewards/rejected": -0.08109745383262634,
|
| 1441 |
+
"step": 880
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 2.31,
|
| 1445 |
+
"grad_norm": 42.25,
|
| 1446 |
+
"learning_rate": 2.256719512667651e-06,
|
| 1447 |
+
"logits/chosen": -1.858231782913208,
|
| 1448 |
+
"logits/rejected": -1.856436014175415,
|
| 1449 |
+
"logps/chosen": -35.18943405151367,
|
| 1450 |
+
"logps/rejected": -36.062679290771484,
|
| 1451 |
+
"loss": 0.2634,
|
| 1452 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1453 |
+
"rewards/chosen": 0.3328378200531006,
|
| 1454 |
+
"rewards/margins": 0.46265101432800293,
|
| 1455 |
+
"rewards/rejected": -0.12981316447257996,
|
| 1456 |
+
"step": 890
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 2.34,
|
| 1460 |
+
"grad_norm": 79.0,
|
| 1461 |
+
"learning_rate": 2.2003892454735786e-06,
|
| 1462 |
+
"logits/chosen": -2.0039381980895996,
|
| 1463 |
+
"logits/rejected": -1.9971100091934204,
|
| 1464 |
+
"logps/chosen": -31.39545249938965,
|
| 1465 |
+
"logps/rejected": -32.833343505859375,
|
| 1466 |
+
"loss": 0.2302,
|
| 1467 |
+
"rewards/accuracies": 0.9375,
|
| 1468 |
+
"rewards/chosen": 0.4250277578830719,
|
| 1469 |
+
"rewards/margins": 0.47772127389907837,
|
| 1470 |
+
"rewards/rejected": -0.05269356444478035,
|
| 1471 |
+
"step": 900
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 2.34,
|
| 1475 |
+
"eval_logits/chosen": -2.2496516704559326,
|
| 1476 |
+
"eval_logits/rejected": -2.2447705268859863,
|
| 1477 |
+
"eval_logps/chosen": -33.913028717041016,
|
| 1478 |
+
"eval_logps/rejected": -37.41090393066406,
|
| 1479 |
+
"eval_loss": 0.5758183598518372,
|
| 1480 |
+
"eval_rewards/accuracies": 0.5544019937515259,
|
| 1481 |
+
"eval_rewards/chosen": 0.08506669104099274,
|
| 1482 |
+
"eval_rewards/margins": 0.011066065169870853,
|
| 1483 |
+
"eval_rewards/rejected": 0.07400061935186386,
|
| 1484 |
+
"eval_runtime": 145.1902,
|
| 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": 21.875,
|
| 1492 |
+
"learning_rate": 2.1442129043167877e-06,
|
| 1493 |
+
"logits/chosen": -2.0508811473846436,
|
| 1494 |
+
"logits/rejected": -2.0461061000823975,
|
| 1495 |
+
"logps/chosen": -33.02284240722656,
|
| 1496 |
+
"logps/rejected": -34.64262390136719,
|
| 1497 |
+
"loss": 0.2017,
|
| 1498 |
+
"rewards/accuracies": 0.9375,
|
| 1499 |
+
"rewards/chosen": 0.3743865489959717,
|
| 1500 |
+
"rewards/margins": 0.45418882369995117,
|
| 1501 |
+
"rewards/rejected": -0.0798022598028183,
|
| 1502 |
+
"step": 910
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 2.39,
|
| 1506 |
+
"grad_norm": 31.875,
|
| 1507 |
+
"learning_rate": 2.088219349982323e-06,
|
| 1508 |
+
"logits/chosen": -1.9676059484481812,
|
| 1509 |
+
"logits/rejected": -1.972545862197876,
|
| 1510 |
+
"logps/chosen": -34.80748748779297,
|
| 1511 |
+
"logps/rejected": -32.547122955322266,
|
| 1512 |
+
"loss": 0.297,
|
| 1513 |
+
"rewards/accuracies": 0.875,
|
| 1514 |
+
"rewards/chosen": 0.3787701427936554,
|
| 1515 |
+
"rewards/margins": 0.41182270646095276,
|
| 1516 |
+
"rewards/rejected": -0.03305256366729736,
|
| 1517 |
+
"step": 920
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 2.42,
|
| 1521 |
+
"grad_norm": 45.0,
|
| 1522 |
+
"learning_rate": 2.0324373493478803e-06,
|
| 1523 |
+
"logits/chosen": -2.0639491081237793,
|
| 1524 |
+
"logits/rejected": -2.054579257965088,
|
| 1525 |
+
"logps/chosen": -31.739765167236328,
|
| 1526 |
+
"logps/rejected": -34.55535125732422,
|
| 1527 |
+
"loss": 0.1863,
|
| 1528 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1529 |
+
"rewards/chosen": 0.3816167712211609,
|
| 1530 |
+
"rewards/margins": 0.4604073166847229,
|
| 1531 |
+
"rewards/rejected": -0.07879054546356201,
|
| 1532 |
+
"step": 930
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 2.44,
|
| 1536 |
+
"grad_norm": 63.5,
|
| 1537 |
+
"learning_rate": 1.976895560604729e-06,
|
| 1538 |
+
"logits/chosen": -1.9907861948013306,
|
| 1539 |
+
"logits/rejected": -1.987722635269165,
|
| 1540 |
+
"logps/chosen": -30.749156951904297,
|
| 1541 |
+
"logps/rejected": -32.56981658935547,
|
| 1542 |
+
"loss": 0.2258,
|
| 1543 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1544 |
+
"rewards/chosen": 0.3775084614753723,
|
| 1545 |
+
"rewards/margins": 0.4506160318851471,
|
| 1546 |
+
"rewards/rejected": -0.07310755550861359,
|
| 1547 |
+
"step": 940
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 2.47,
|
| 1551 |
+
"grad_norm": 29.875,
|
| 1552 |
+
"learning_rate": 1.921622518534466e-06,
|
| 1553 |
+
"logits/chosen": -1.9317022562026978,
|
| 1554 |
+
"logits/rejected": -1.9392095804214478,
|
| 1555 |
+
"logps/chosen": -32.34041976928711,
|
| 1556 |
+
"logps/rejected": -35.695228576660156,
|
| 1557 |
+
"loss": 0.2412,
|
| 1558 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1559 |
+
"rewards/chosen": 0.3714412748813629,
|
| 1560 |
+
"rewards/margins": 0.47314995527267456,
|
| 1561 |
+
"rewards/rejected": -0.10170867294073105,
|
| 1562 |
+
"step": 950
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 2.49,
|
| 1566 |
+
"grad_norm": 114.5,
|
| 1567 |
+
"learning_rate": 1.8666466198491794e-06,
|
| 1568 |
+
"logits/chosen": -1.9403190612792969,
|
| 1569 |
+
"logits/rejected": -1.9334567785263062,
|
| 1570 |
+
"logps/chosen": -33.3146858215332,
|
| 1571 |
+
"logps/rejected": -36.74799346923828,
|
| 1572 |
+
"loss": 0.3338,
|
| 1573 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1574 |
+
"rewards/chosen": 0.3596806824207306,
|
| 1575 |
+
"rewards/margins": 0.46411022543907166,
|
| 1576 |
+
"rewards/rejected": -0.1044294610619545,
|
| 1577 |
+
"step": 960
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 2.52,
|
| 1581 |
+
"grad_norm": 30.375,
|
| 1582 |
+
"learning_rate": 1.8119961086025376e-06,
|
| 1583 |
+
"logits/chosen": -1.9422130584716797,
|
| 1584 |
+
"logits/rejected": -1.942911148071289,
|
| 1585 |
+
"logps/chosen": -30.72060203552246,
|
| 1586 |
+
"logps/rejected": -33.139339447021484,
|
| 1587 |
+
"loss": 0.2958,
|
| 1588 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1589 |
+
"rewards/chosen": 0.3610595762729645,
|
| 1590 |
+
"rewards/margins": 0.4717164933681488,
|
| 1591 |
+
"rewards/rejected": -0.11065696179866791,
|
| 1592 |
+
"step": 970
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 2.55,
|
| 1596 |
+
"grad_norm": 49.75,
|
| 1597 |
+
"learning_rate": 1.7576990616793139e-06,
|
| 1598 |
+
"logits/chosen": -1.9468282461166382,
|
| 1599 |
+
"logits/rejected": -1.9580564498901367,
|
| 1600 |
+
"logps/chosen": -32.50783157348633,
|
| 1601 |
+
"logps/rejected": -34.23973083496094,
|
| 1602 |
+
"loss": 0.2541,
|
| 1603 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1604 |
+
"rewards/chosen": 0.3981999158859253,
|
| 1605 |
+
"rewards/margins": 0.47416257858276367,
|
| 1606 |
+
"rewards/rejected": -0.07596264779567719,
|
| 1607 |
+
"step": 980
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 2.57,
|
| 1611 |
+
"grad_norm": 83.5,
|
| 1612 |
+
"learning_rate": 1.7037833743707892e-06,
|
| 1613 |
+
"logits/chosen": -2.0208706855773926,
|
| 1614 |
+
"logits/rejected": -2.018265962600708,
|
| 1615 |
+
"logps/chosen": -35.62385940551758,
|
| 1616 |
+
"logps/rejected": -32.450279235839844,
|
| 1617 |
+
"loss": 0.3659,
|
| 1618 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 1619 |
+
"rewards/chosen": 0.34803569316864014,
|
| 1620 |
+
"rewards/margins": 0.38377222418785095,
|
| 1621 |
+
"rewards/rejected": -0.03573652729392052,
|
| 1622 |
+
"step": 990
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 2.6,
|
| 1626 |
+
"grad_norm": 24.375,
|
| 1627 |
+
"learning_rate": 1.6502767460434588e-06,
|
| 1628 |
+
"logits/chosen": -1.9889123439788818,
|
| 1629 |
+
"logits/rejected": -1.9926494359970093,
|
| 1630 |
+
"logps/chosen": -34.00056457519531,
|
| 1631 |
+
"logps/rejected": -34.569400787353516,
|
| 1632 |
+
"loss": 0.2296,
|
| 1633 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1634 |
+
"rewards/chosen": 0.32390862703323364,
|
| 1635 |
+
"rewards/margins": 0.38225558400154114,
|
| 1636 |
+
"rewards/rejected": -0.05834692716598511,
|
| 1637 |
+
"step": 1000
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 2.6,
|
| 1641 |
+
"eval_logits/chosen": -2.248878240585327,
|
| 1642 |
+
"eval_logits/rejected": -2.2439959049224854,
|
| 1643 |
+
"eval_logps/chosen": -33.944374084472656,
|
| 1644 |
+
"eval_logps/rejected": -37.437828063964844,
|
| 1645 |
+
"eval_loss": 0.5750337243080139,
|
| 1646 |
+
"eval_rewards/accuracies": 0.5074750781059265,
|
| 1647 |
+
"eval_rewards/chosen": 0.06312327831983566,
|
| 1648 |
+
"eval_rewards/margins": 0.007969892583787441,
|
| 1649 |
+
"eval_rewards/rejected": 0.05515338107943535,
|
| 1650 |
+
"eval_runtime": 145.3511,
|
| 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": 75.0,
|
| 1658 |
+
"learning_rate": 1.5972066659083796e-06,
|
| 1659 |
+
"logits/chosen": -1.9976009130477905,
|
| 1660 |
+
"logits/rejected": -2.0026943683624268,
|
| 1661 |
+
"logps/chosen": -32.26424789428711,
|
| 1662 |
+
"logps/rejected": -32.64023208618164,
|
| 1663 |
+
"loss": 0.204,
|
| 1664 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1665 |
+
"rewards/chosen": 0.3753224313259125,
|
| 1666 |
+
"rewards/margins": 0.45860418677330017,
|
| 1667 |
+
"rewards/rejected": -0.08328177034854889,
|
| 1668 |
+
"step": 1010
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 2.65,
|
| 1672 |
+
"grad_norm": 58.5,
|
| 1673 |
+
"learning_rate": 1.5446003988985041e-06,
|
| 1674 |
+
"logits/chosen": -1.9610779285430908,
|
| 1675 |
+
"logits/rejected": -1.9637727737426758,
|
| 1676 |
+
"logps/chosen": -29.9860897064209,
|
| 1677 |
+
"logps/rejected": -31.74285316467285,
|
| 1678 |
+
"loss": 0.2786,
|
| 1679 |
+
"rewards/accuracies": 0.875,
|
| 1680 |
+
"rewards/chosen": 0.32168638706207275,
|
| 1681 |
+
"rewards/margins": 0.3522700071334839,
|
| 1682 |
+
"rewards/rejected": -0.03058364987373352,
|
| 1683 |
+
"step": 1020
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 2.68,
|
| 1687 |
+
"grad_norm": 24.0,
|
| 1688 |
+
"learning_rate": 1.4924849716612211e-06,
|
| 1689 |
+
"logits/chosen": -2.0177597999572754,
|
| 1690 |
+
"logits/rejected": -2.011871814727783,
|
| 1691 |
+
"logps/chosen": -33.44524383544922,
|
| 1692 |
+
"logps/rejected": -33.608463287353516,
|
| 1693 |
+
"loss": 0.1846,
|
| 1694 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1695 |
+
"rewards/chosen": 0.38070768117904663,
|
| 1696 |
+
"rewards/margins": 0.44403451681137085,
|
| 1697 |
+
"rewards/rejected": -0.06332685798406601,
|
| 1698 |
+
"step": 1030
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 2.7,
|
| 1702 |
+
"grad_norm": 46.25,
|
| 1703 |
+
"learning_rate": 1.440887158673332e-06,
|
| 1704 |
+
"logits/chosen": -1.945654273033142,
|
| 1705 |
+
"logits/rejected": -1.9493560791015625,
|
| 1706 |
+
"logps/chosen": -35.2274284362793,
|
| 1707 |
+
"logps/rejected": -34.58420181274414,
|
| 1708 |
+
"loss": 0.2056,
|
| 1709 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1710 |
+
"rewards/chosen": 0.36841046810150146,
|
| 1711 |
+
"rewards/margins": 0.483646959066391,
|
| 1712 |
+
"rewards/rejected": -0.11523648351430893,
|
| 1713 |
+
"step": 1040
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 2.73,
|
| 1717 |
+
"grad_norm": 17.375,
|
| 1718 |
+
"learning_rate": 1.3898334684855647e-06,
|
| 1719 |
+
"logits/chosen": -1.9452226161956787,
|
| 1720 |
+
"logits/rejected": -1.958367109298706,
|
| 1721 |
+
"logps/chosen": -31.376358032226562,
|
| 1722 |
+
"logps/rejected": -33.11796951293945,
|
| 1723 |
+
"loss": 0.2082,
|
| 1724 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1725 |
+
"rewards/chosen": 0.34057068824768066,
|
| 1726 |
+
"rewards/margins": 0.40912023186683655,
|
| 1727 |
+
"rewards/rejected": -0.06854955106973648,
|
| 1728 |
+
"step": 1050
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 2.75,
|
| 1732 |
+
"grad_norm": 49.5,
|
| 1733 |
+
"learning_rate": 1.3393501301037245e-06,
|
| 1734 |
+
"logits/chosen": -1.9365898370742798,
|
| 1735 |
+
"logits/rejected": -1.9300063848495483,
|
| 1736 |
+
"logps/chosen": -31.5865478515625,
|
| 1737 |
+
"logps/rejected": -33.410743713378906,
|
| 1738 |
+
"loss": 0.1983,
|
| 1739 |
+
"rewards/accuracies": 0.9375,
|
| 1740 |
+
"rewards/chosen": 0.4009482264518738,
|
| 1741 |
+
"rewards/margins": 0.46589574217796326,
|
| 1742 |
+
"rewards/rejected": -0.06494755297899246,
|
| 1743 |
+
"step": 1060
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 2.78,
|
| 1747 |
+
"grad_norm": 30.125,
|
| 1748 |
+
"learning_rate": 1.2894630795134454e-06,
|
| 1749 |
+
"logits/chosen": -2.062471866607666,
|
| 1750 |
+
"logits/rejected": -2.063298225402832,
|
| 1751 |
+
"logps/chosen": -33.060279846191406,
|
| 1752 |
+
"logps/rejected": -33.027076721191406,
|
| 1753 |
+
"loss": 0.1608,
|
| 1754 |
+
"rewards/accuracies": 0.9375,
|
| 1755 |
+
"rewards/chosen": 0.3407577872276306,
|
| 1756 |
+
"rewards/margins": 0.46842724084854126,
|
| 1757 |
+
"rewards/rejected": -0.12766942381858826,
|
| 1758 |
+
"step": 1070
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 2.81,
|
| 1762 |
+
"grad_norm": 19.125,
|
| 1763 |
+
"learning_rate": 1.2401979463554984e-06,
|
| 1764 |
+
"logits/chosen": -1.9935426712036133,
|
| 1765 |
+
"logits/rejected": -1.9923900365829468,
|
| 1766 |
+
"logps/chosen": -33.55720901489258,
|
| 1767 |
+
"logps/rejected": -33.266090393066406,
|
| 1768 |
+
"loss": 0.2082,
|
| 1769 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1770 |
+
"rewards/chosen": 0.372835636138916,
|
| 1771 |
+
"rewards/margins": 0.42481860518455505,
|
| 1772 |
+
"rewards/rejected": -0.051982950419187546,
|
| 1773 |
+
"step": 1080
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 2.83,
|
| 1777 |
+
"grad_norm": 37.0,
|
| 1778 |
+
"learning_rate": 1.1915800407584705e-06,
|
| 1779 |
+
"logits/chosen": -2.021810531616211,
|
| 1780 |
+
"logits/rejected": -2.01427960395813,
|
| 1781 |
+
"logps/chosen": -33.71491622924805,
|
| 1782 |
+
"logps/rejected": -31.135467529296875,
|
| 1783 |
+
"loss": 0.1958,
|
| 1784 |
+
"rewards/accuracies": 0.9375,
|
| 1785 |
+
"rewards/chosen": 0.3576733469963074,
|
| 1786 |
+
"rewards/margins": 0.42342647910118103,
|
| 1787 |
+
"rewards/rejected": -0.06575315445661545,
|
| 1788 |
+
"step": 1090
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 2.86,
|
| 1792 |
+
"grad_norm": 67.5,
|
| 1793 |
+
"learning_rate": 1.1436343403356019e-06,
|
| 1794 |
+
"logits/chosen": -1.9883337020874023,
|
| 1795 |
+
"logits/rejected": -1.9875253438949585,
|
| 1796 |
+
"logps/chosen": -34.86940002441406,
|
| 1797 |
+
"logps/rejected": -36.43136978149414,
|
| 1798 |
+
"loss": 0.2798,
|
| 1799 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1800 |
+
"rewards/chosen": 0.3843076229095459,
|
| 1801 |
+
"rewards/margins": 0.4393698275089264,
|
| 1802 |
+
"rewards/rejected": -0.055062174797058105,
|
| 1803 |
+
"step": 1100
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 2.86,
|
| 1807 |
+
"eval_logits/chosen": -2.2467923164367676,
|
| 1808 |
+
"eval_logits/rejected": -2.2419159412384033,
|
| 1809 |
+
"eval_logps/chosen": -33.93034744262695,
|
| 1810 |
+
"eval_logps/rejected": -37.438724517822266,
|
| 1811 |
+
"eval_loss": 0.5483447909355164,
|
| 1812 |
+
"eval_rewards/accuracies": 0.5427741408348083,
|
| 1813 |
+
"eval_rewards/chosen": 0.07294297218322754,
|
| 1814 |
+
"eval_rewards/margins": 0.018412673845887184,
|
| 1815 |
+
"eval_rewards/rejected": 0.05453029274940491,
|
| 1816 |
+
"eval_runtime": 145.1205,
|
| 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": 54.25,
|
| 1824 |
+
"learning_rate": 1.0963854773524548e-06,
|
| 1825 |
+
"logits/chosen": -2.0355682373046875,
|
| 1826 |
+
"logits/rejected": -2.0409066677093506,
|
| 1827 |
+
"logps/chosen": -34.98146057128906,
|
| 1828 |
+
"logps/rejected": -35.63863754272461,
|
| 1829 |
+
"loss": 0.2279,
|
| 1830 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1831 |
+
"rewards/chosen": 0.3399600386619568,
|
| 1832 |
+
"rewards/margins": 0.40649813413619995,
|
| 1833 |
+
"rewards/rejected": -0.06653807312250137,
|
| 1834 |
+
"step": 1110
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 2.91,
|
| 1838 |
+
"grad_norm": 71.5,
|
| 1839 |
+
"learning_rate": 1.049857726072005e-06,
|
| 1840 |
+
"logits/chosen": -2.0084500312805176,
|
| 1841 |
+
"logits/rejected": -2.006545066833496,
|
| 1842 |
+
"logps/chosen": -32.13376235961914,
|
| 1843 |
+
"logps/rejected": -33.14947509765625,
|
| 1844 |
+
"loss": 0.2249,
|
| 1845 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1846 |
+
"rewards/chosen": 0.33771127462387085,
|
| 1847 |
+
"rewards/margins": 0.45441800355911255,
|
| 1848 |
+
"rewards/rejected": -0.11670677363872528,
|
| 1849 |
+
"step": 1120
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 2.94,
|
| 1853 |
+
"grad_norm": 17.875,
|
| 1854 |
+
"learning_rate": 1.0040749902836508e-06,
|
| 1855 |
+
"logits/chosen": -1.9112991094589233,
|
| 1856 |
+
"logits/rejected": -1.9141557216644287,
|
| 1857 |
+
"logps/chosen": -29.258800506591797,
|
| 1858 |
+
"logps/rejected": -30.369312286376953,
|
| 1859 |
+
"loss": 0.2421,
|
| 1860 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1861 |
+
"rewards/chosen": 0.3428861200809479,
|
| 1862 |
+
"rewards/margins": 0.429875910282135,
|
| 1863 |
+
"rewards/rejected": -0.08698974549770355,
|
| 1864 |
+
"step": 1130
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 2.96,
|
| 1868 |
+
"grad_norm": 42.25,
|
| 1869 |
+
"learning_rate": 9.59060791022566e-07,
|
| 1870 |
+
"logits/chosen": -2.0088257789611816,
|
| 1871 |
+
"logits/rejected": -2.005708694458008,
|
| 1872 |
+
"logps/chosen": -32.668514251708984,
|
| 1873 |
+
"logps/rejected": -32.6502799987793,
|
| 1874 |
+
"loss": 0.2007,
|
| 1875 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1876 |
+
"rewards/chosen": 0.37466686964035034,
|
| 1877 |
+
"rewards/margins": 0.461048424243927,
|
| 1878 |
+
"rewards/rejected": -0.08638156205415726,
|
| 1879 |
+
"step": 1140
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 2.99,
|
| 1883 |
+
"grad_norm": 29.25,
|
| 1884 |
+
"learning_rate": 9.148382544856885e-07,
|
| 1885 |
+
"logits/chosen": -1.930190086364746,
|
| 1886 |
+
"logits/rejected": -1.9238803386688232,
|
| 1887 |
+
"logps/chosen": -27.67626953125,
|
| 1888 |
+
"logps/rejected": -32.34447479248047,
|
| 1889 |
+
"loss": 0.1585,
|
| 1890 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1891 |
+
"rewards/chosen": 0.32320261001586914,
|
| 1892 |
+
"rewards/margins": 0.4605270028114319,
|
| 1893 |
+
"rewards/rejected": -0.13732439279556274,
|
| 1894 |
+
"step": 1150
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 3.01,
|
| 1898 |
+
"grad_norm": 35.75,
|
| 1899 |
+
"learning_rate": 8.714301001505568e-07,
|
| 1900 |
+
"logits/chosen": -2.04903507232666,
|
| 1901 |
+
"logits/rejected": -2.0465400218963623,
|
| 1902 |
+
"logps/chosen": -32.619815826416016,
|
| 1903 |
+
"logps/rejected": -35.1733283996582,
|
| 1904 |
+
"loss": 0.1505,
|
| 1905 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1906 |
+
"rewards/chosen": 0.3333774507045746,
|
| 1907 |
+
"rewards/margins": 0.39063215255737305,
|
| 1908 |
+
"rewards/rejected": -0.05725465342402458,
|
| 1909 |
+
"step": 1160
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 3.04,
|
| 1913 |
+
"grad_norm": 15.625,
|
| 1914 |
+
"learning_rate": 8.288586291031025e-07,
|
| 1915 |
+
"logits/chosen": -2.0296478271484375,
|
| 1916 |
+
"logits/rejected": -2.0277647972106934,
|
| 1917 |
+
"logps/chosen": -31.087743759155273,
|
| 1918 |
+
"logps/rejected": -32.51514434814453,
|
| 1919 |
+
"loss": 0.0912,
|
| 1920 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1921 |
+
"rewards/chosen": 0.36866381764411926,
|
| 1922 |
+
"rewards/margins": 0.4744475483894348,
|
| 1923 |
+
"rewards/rejected": -0.10578374564647675,
|
| 1924 |
+
"step": 1170
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 3.06,
|
| 1928 |
+
"grad_norm": 35.75,
|
| 1929 |
+
"learning_rate": 7.871457125803897e-07,
|
| 1930 |
+
"logits/chosen": -2.0225229263305664,
|
| 1931 |
+
"logits/rejected": -2.0101773738861084,
|
| 1932 |
+
"logps/chosen": -35.537513732910156,
|
| 1933 |
+
"logps/rejected": -34.152286529541016,
|
| 1934 |
+
"loss": 0.183,
|
| 1935 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1936 |
+
"rewards/chosen": 0.35545071959495544,
|
| 1937 |
+
"rewards/margins": 0.4120173454284668,
|
| 1938 |
+
"rewards/rejected": -0.056566618382930756,
|
| 1939 |
+
"step": 1180
|
| 1940 |
+
},
|
| 1941 |
+
{
|
| 1942 |
+
"epoch": 3.09,
|
| 1943 |
+
"grad_norm": 30.5,
|
| 1944 |
+
"learning_rate": 7.463127807341966e-07,
|
| 1945 |
+
"logits/chosen": -1.9124174118041992,
|
| 1946 |
+
"logits/rejected": -1.907052755355835,
|
| 1947 |
+
"logps/chosen": -34.80583953857422,
|
| 1948 |
+
"logps/rejected": -34.315452575683594,
|
| 1949 |
+
"loss": 0.1388,
|
| 1950 |
+
"rewards/accuracies": 0.9375,
|
| 1951 |
+
"rewards/chosen": 0.3806923031806946,
|
| 1952 |
+
"rewards/margins": 0.43503332138061523,
|
| 1953 |
+
"rewards/rejected": -0.05434093996882439,
|
| 1954 |
+
"step": 1190
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 3.12,
|
| 1958 |
+
"grad_norm": 30.75,
|
| 1959 |
+
"learning_rate": 7.063808116212021e-07,
|
| 1960 |
+
"logits/chosen": -1.9541196823120117,
|
| 1961 |
+
"logits/rejected": -1.9571189880371094,
|
| 1962 |
+
"logps/chosen": -32.08814239501953,
|
| 1963 |
+
"logps/rejected": -31.591243743896484,
|
| 1964 |
+
"loss": 0.1195,
|
| 1965 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1966 |
+
"rewards/chosen": 0.3688638210296631,
|
| 1967 |
+
"rewards/margins": 0.4424096941947937,
|
| 1968 |
+
"rewards/rejected": -0.07354583591222763,
|
| 1969 |
+
"step": 1200
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 3.12,
|
| 1973 |
+
"eval_logits/chosen": -2.247305393218994,
|
| 1974 |
+
"eval_logits/rejected": -2.242413282394409,
|
| 1975 |
+
"eval_logps/chosen": -33.93861770629883,
|
| 1976 |
+
"eval_logps/rejected": -37.42906188964844,
|
| 1977 |
+
"eval_loss": 0.5758681297302246,
|
| 1978 |
+
"eval_rewards/accuracies": 0.5137043595314026,
|
| 1979 |
+
"eval_rewards/chosen": 0.06715311110019684,
|
| 1980 |
+
"eval_rewards/margins": 0.005864266771823168,
|
| 1981 |
+
"eval_rewards/rejected": 0.06128884479403496,
|
| 1982 |
+
"eval_runtime": 145.3379,
|
| 1983 |
+
"eval_samples_per_second": 2.36,
|
| 1984 |
+
"eval_steps_per_second": 0.296,
|
| 1985 |
+
"step": 1200
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 3.14,
|
| 1989 |
+
"grad_norm": 32.75,
|
| 1990 |
+
"learning_rate": 6.673703204254348e-07,
|
| 1991 |
+
"logits/chosen": -1.986507773399353,
|
| 1992 |
+
"logits/rejected": -1.9819329977035522,
|
| 1993 |
+
"logps/chosen": -30.773900985717773,
|
| 1994 |
+
"logps/rejected": -31.090429306030273,
|
| 1995 |
+
"loss": 0.117,
|
| 1996 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1997 |
+
"rewards/chosen": 0.349838525056839,
|
| 1998 |
+
"rewards/margins": 0.4774731695652008,
|
| 1999 |
+
"rewards/rejected": -0.12763459980487823,
|
| 2000 |
+
"step": 1210
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 3.17,
|
| 2004 |
+
"grad_norm": 33.75,
|
| 2005 |
+
"learning_rate": 6.293013489185315e-07,
|
| 2006 |
+
"logits/chosen": -1.9981073141098022,
|
| 2007 |
+
"logits/rejected": -1.9925674200057983,
|
| 2008 |
+
"logps/chosen": -34.77349090576172,
|
| 2009 |
+
"logps/rejected": -34.63701248168945,
|
| 2010 |
+
"loss": 0.1218,
|
| 2011 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2012 |
+
"rewards/chosen": 0.35031452775001526,
|
| 2013 |
+
"rewards/margins": 0.3973073959350586,
|
| 2014 |
+
"rewards/rejected": -0.046992845833301544,
|
| 2015 |
+
"step": 1220
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 3.19,
|
| 2019 |
+
"grad_norm": 172.0,
|
| 2020 |
+
"learning_rate": 5.921934551632086e-07,
|
| 2021 |
+
"logits/chosen": -1.9999059438705444,
|
| 2022 |
+
"logits/rejected": -1.9867637157440186,
|
| 2023 |
+
"logps/chosen": -32.57126998901367,
|
| 2024 |
+
"logps/rejected": -34.21943283081055,
|
| 2025 |
+
"loss": 0.1846,
|
| 2026 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2027 |
+
"rewards/chosen": 0.3659234941005707,
|
| 2028 |
+
"rewards/margins": 0.4356323182582855,
|
| 2029 |
+
"rewards/rejected": -0.06970882415771484,
|
| 2030 |
+
"step": 1230
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 3.22,
|
| 2034 |
+
"grad_norm": 36.25,
|
| 2035 |
+
"learning_rate": 5.560657034652405e-07,
|
| 2036 |
+
"logits/chosen": -2.0442328453063965,
|
| 2037 |
+
"logits/rejected": -2.041350841522217,
|
| 2038 |
+
"logps/chosen": -34.520362854003906,
|
| 2039 |
+
"logps/rejected": -31.955829620361328,
|
| 2040 |
+
"loss": 0.1368,
|
| 2041 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2042 |
+
"rewards/chosen": 0.3587776720523834,
|
| 2043 |
+
"rewards/margins": 0.4399491250514984,
|
| 2044 |
+
"rewards/rejected": -0.08117148280143738,
|
| 2045 |
+
"step": 1240
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 3.25,
|
| 2049 |
+
"grad_norm": 95.5,
|
| 2050 |
+
"learning_rate": 5.2093665457911e-07,
|
| 2051 |
+
"logits/chosen": -1.97548508644104,
|
| 2052 |
+
"logits/rejected": -1.9726585149765015,
|
| 2053 |
+
"logps/chosen": -33.8781623840332,
|
| 2054 |
+
"logps/rejected": -34.95212173461914,
|
| 2055 |
+
"loss": 0.1618,
|
| 2056 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2057 |
+
"rewards/chosen": 0.3625948429107666,
|
| 2058 |
+
"rewards/margins": 0.47109127044677734,
|
| 2059 |
+
"rewards/rejected": -0.10849642753601074,
|
| 2060 |
+
"step": 1250
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 3.27,
|
| 2064 |
+
"grad_norm": 26.25,
|
| 2065 |
+
"learning_rate": 4.868243561723535e-07,
|
| 2066 |
+
"logits/chosen": -2.038349151611328,
|
| 2067 |
+
"logits/rejected": -2.033210277557373,
|
| 2068 |
+
"logps/chosen": -30.266530990600586,
|
| 2069 |
+
"logps/rejected": -32.47600555419922,
|
| 2070 |
+
"loss": 0.1155,
|
| 2071 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2072 |
+
"rewards/chosen": 0.3459872007369995,
|
| 2073 |
+
"rewards/margins": 0.4114041328430176,
|
| 2074 |
+
"rewards/rejected": -0.06541696935892105,
|
| 2075 |
+
"step": 1260
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 3.3,
|
| 2079 |
+
"grad_norm": 17.75,
|
| 2080 |
+
"learning_rate": 4.537463335535161e-07,
|
| 2081 |
+
"logits/chosen": -2.0647287368774414,
|
| 2082 |
+
"logits/rejected": -2.069408416748047,
|
| 2083 |
+
"logps/chosen": -32.192989349365234,
|
| 2084 |
+
"logps/rejected": -31.8017578125,
|
| 2085 |
+
"loss": 0.126,
|
| 2086 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2087 |
+
"rewards/chosen": 0.3626050055027008,
|
| 2088 |
+
"rewards/margins": 0.446880578994751,
|
| 2089 |
+
"rewards/rejected": -0.08427558094263077,
|
| 2090 |
+
"step": 1270
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 3.32,
|
| 2094 |
+
"grad_norm": 42.0,
|
| 2095 |
+
"learning_rate": 4.217195806684629e-07,
|
| 2096 |
+
"logits/chosen": -1.9595832824707031,
|
| 2097 |
+
"logits/rejected": -1.966700792312622,
|
| 2098 |
+
"logps/chosen": -34.40987777709961,
|
| 2099 |
+
"logps/rejected": -32.77952575683594,
|
| 2100 |
+
"loss": 0.1119,
|
| 2101 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2102 |
+
"rewards/chosen": 0.3481543958187103,
|
| 2103 |
+
"rewards/margins": 0.4750334620475769,
|
| 2104 |
+
"rewards/rejected": -0.126879021525383,
|
| 2105 |
+
"step": 1280
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 3.35,
|
| 2109 |
+
"grad_norm": 30.375,
|
| 2110 |
+
"learning_rate": 3.907605513696808e-07,
|
| 2111 |
+
"logits/chosen": -1.8622219562530518,
|
| 2112 |
+
"logits/rejected": -1.8643817901611328,
|
| 2113 |
+
"logps/chosen": -32.59587860107422,
|
| 2114 |
+
"logps/rejected": -36.642635345458984,
|
| 2115 |
+
"loss": 0.1373,
|
| 2116 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2117 |
+
"rewards/chosen": 0.3752126097679138,
|
| 2118 |
+
"rewards/margins": 0.4370684027671814,
|
| 2119 |
+
"rewards/rejected": -0.06185578554868698,
|
| 2120 |
+
"step": 1290
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 3.38,
|
| 2124 |
+
"grad_norm": 89.5,
|
| 2125 |
+
"learning_rate": 3.6088515096305675e-07,
|
| 2126 |
+
"logits/chosen": -1.9182116985321045,
|
| 2127 |
+
"logits/rejected": -1.9216159582138062,
|
| 2128 |
+
"logps/chosen": -32.71792221069336,
|
| 2129 |
+
"logps/rejected": -32.585205078125,
|
| 2130 |
+
"loss": 0.1371,
|
| 2131 |
+
"rewards/accuracies": 0.9375,
|
| 2132 |
+
"rewards/chosen": 0.3491230309009552,
|
| 2133 |
+
"rewards/margins": 0.4338512420654297,
|
| 2134 |
+
"rewards/rejected": -0.08472825586795807,
|
| 2135 |
+
"step": 1300
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 3.38,
|
| 2139 |
+
"eval_logits/chosen": -2.2483434677124023,
|
| 2140 |
+
"eval_logits/rejected": -2.243445634841919,
|
| 2141 |
+
"eval_logps/chosen": -33.929874420166016,
|
| 2142 |
+
"eval_logps/rejected": -37.43461990356445,
|
| 2143 |
+
"eval_loss": 0.5592050552368164,
|
| 2144 |
+
"eval_rewards/accuracies": 0.5494186282157898,
|
| 2145 |
+
"eval_rewards/chosen": 0.0732741430401802,
|
| 2146 |
+
"eval_rewards/margins": 0.01587507873773575,
|
| 2147 |
+
"eval_rewards/rejected": 0.05739906057715416,
|
| 2148 |
+
"eval_runtime": 145.2128,
|
| 2149 |
+
"eval_samples_per_second": 2.362,
|
| 2150 |
+
"eval_steps_per_second": 0.296,
|
| 2151 |
+
"step": 1300
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 3.4,
|
| 2155 |
+
"grad_norm": 22.125,
|
| 2156 |
+
"learning_rate": 3.321087280364757e-07,
|
| 2157 |
+
"logits/chosen": -2.0270915031433105,
|
| 2158 |
+
"logits/rejected": -2.006920337677002,
|
| 2159 |
+
"logps/chosen": -30.607580184936523,
|
| 2160 |
+
"logps/rejected": -35.26466369628906,
|
| 2161 |
+
"loss": 0.14,
|
| 2162 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2163 |
+
"rewards/chosen": 0.35407310724258423,
|
| 2164 |
+
"rewards/margins": 0.42285603284835815,
|
| 2165 |
+
"rewards/rejected": -0.0687829852104187,
|
| 2166 |
+
"step": 1310
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 3.43,
|
| 2170 |
+
"grad_norm": 22.875,
|
| 2171 |
+
"learning_rate": 3.044460665744284e-07,
|
| 2172 |
+
"logits/chosen": -2.006091594696045,
|
| 2173 |
+
"logits/rejected": -2.0118560791015625,
|
| 2174 |
+
"logps/chosen": -32.49515914916992,
|
| 2175 |
+
"logps/rejected": -32.58808517456055,
|
| 2176 |
+
"loss": 0.1516,
|
| 2177 |
+
"rewards/accuracies": 0.9375,
|
| 2178 |
+
"rewards/chosen": 0.37704116106033325,
|
| 2179 |
+
"rewards/margins": 0.4451761245727539,
|
| 2180 |
+
"rewards/rejected": -0.06813497841358185,
|
| 2181 |
+
"step": 1320
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 3.45,
|
| 2185 |
+
"grad_norm": 58.25,
|
| 2186 |
+
"learning_rate": 2.779113783626916e-07,
|
| 2187 |
+
"logits/chosen": -1.9745457172393799,
|
| 2188 |
+
"logits/rejected": -1.969630479812622,
|
| 2189 |
+
"logps/chosen": -32.551475524902344,
|
| 2190 |
+
"logps/rejected": -34.6931266784668,
|
| 2191 |
+
"loss": 0.1414,
|
| 2192 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2193 |
+
"rewards/chosen": 0.3297252357006073,
|
| 2194 |
+
"rewards/margins": 0.4143126606941223,
|
| 2195 |
+
"rewards/rejected": -0.08458739519119263,
|
| 2196 |
+
"step": 1330
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 3.48,
|
| 2200 |
+
"grad_norm": 38.5,
|
| 2201 |
+
"learning_rate": 2.5251829568697204e-07,
|
| 2202 |
+
"logits/chosen": -1.840356469154358,
|
| 2203 |
+
"logits/rejected": -1.8500515222549438,
|
| 2204 |
+
"logps/chosen": -33.240806579589844,
|
| 2205 |
+
"logps/rejected": -31.120624542236328,
|
| 2206 |
+
"loss": 0.1088,
|
| 2207 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2208 |
+
"rewards/chosen": 0.391147643327713,
|
| 2209 |
+
"rewards/margins": 0.4517039358615875,
|
| 2210 |
+
"rewards/rejected": -0.0605563148856163,
|
| 2211 |
+
"step": 1340
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 3.51,
|
| 2215 |
+
"grad_norm": 25.75,
|
| 2216 |
+
"learning_rate": 2.2827986432927774e-07,
|
| 2217 |
+
"logits/chosen": -1.9017702341079712,
|
| 2218 |
+
"logits/rejected": -1.906445860862732,
|
| 2219 |
+
"logps/chosen": -33.377037048339844,
|
| 2220 |
+
"logps/rejected": -33.292083740234375,
|
| 2221 |
+
"loss": 0.0844,
|
| 2222 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2223 |
+
"rewards/chosen": 0.4017297625541687,
|
| 2224 |
+
"rewards/margins": 0.4628194272518158,
|
| 2225 |
+
"rewards/rejected": -0.06108973175287247,
|
| 2226 |
+
"step": 1350
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 3.53,
|
| 2230 |
+
"grad_norm": 70.5,
|
| 2231 |
+
"learning_rate": 2.0520853686560177e-07,
|
| 2232 |
+
"logits/chosen": -1.9277992248535156,
|
| 2233 |
+
"logits/rejected": -1.9333375692367554,
|
| 2234 |
+
"logps/chosen": -33.33270263671875,
|
| 2235 |
+
"logps/rejected": -34.68195343017578,
|
| 2236 |
+
"loss": 0.13,
|
| 2237 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2238 |
+
"rewards/chosen": 0.3492715060710907,
|
| 2239 |
+
"rewards/margins": 0.43249478936195374,
|
| 2240 |
+
"rewards/rejected": -0.08322323858737946,
|
| 2241 |
+
"step": 1360
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 3.56,
|
| 2245 |
+
"grad_norm": 45.0,
|
| 2246 |
+
"learning_rate": 1.833161662683672e-07,
|
| 2247 |
+
"logits/chosen": -2.017637252807617,
|
| 2248 |
+
"logits/rejected": -2.0131232738494873,
|
| 2249 |
+
"logps/chosen": -32.15598678588867,
|
| 2250 |
+
"logps/rejected": -30.987085342407227,
|
| 2251 |
+
"loss": 0.1109,
|
| 2252 |
+
"rewards/accuracies": 0.9375,
|
| 2253 |
+
"rewards/chosen": 0.335787296295166,
|
| 2254 |
+
"rewards/margins": 0.38841742277145386,
|
| 2255 |
+
"rewards/rejected": -0.052630096673965454,
|
| 2256 |
+
"step": 1370
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 3.58,
|
| 2260 |
+
"grad_norm": 23.75,
|
| 2261 |
+
"learning_rate": 1.626139998169246e-07,
|
| 2262 |
+
"logits/chosen": -1.9399824142456055,
|
| 2263 |
+
"logits/rejected": -1.9417117834091187,
|
| 2264 |
+
"logps/chosen": -29.205978393554688,
|
| 2265 |
+
"logps/rejected": -30.845836639404297,
|
| 2266 |
+
"loss": 0.1135,
|
| 2267 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2268 |
+
"rewards/chosen": 0.3173268437385559,
|
| 2269 |
+
"rewards/margins": 0.3884861469268799,
|
| 2270 |
+
"rewards/rejected": -0.07115931063890457,
|
| 2271 |
+
"step": 1380
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 3.61,
|
| 2275 |
+
"grad_norm": 36.5,
|
| 2276 |
+
"learning_rate": 1.4311267331922535e-07,
|
| 2277 |
+
"logits/chosen": -1.9257915019989014,
|
| 2278 |
+
"logits/rejected": -1.9303070306777954,
|
| 2279 |
+
"logps/chosen": -31.455486297607422,
|
| 2280 |
+
"logps/rejected": -33.11508560180664,
|
| 2281 |
+
"loss": 0.0911,
|
| 2282 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2283 |
+
"rewards/chosen": 0.39480385184288025,
|
| 2284 |
+
"rewards/margins": 0.4837164878845215,
|
| 2285 |
+
"rewards/rejected": -0.08891263604164124,
|
| 2286 |
+
"step": 1390
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 3.64,
|
| 2290 |
+
"grad_norm": 60.0,
|
| 2291 |
+
"learning_rate": 1.2482220564763669e-07,
|
| 2292 |
+
"logits/chosen": -1.99787175655365,
|
| 2293 |
+
"logits/rejected": -2.0011351108551025,
|
| 2294 |
+
"logps/chosen": -34.36697006225586,
|
| 2295 |
+
"logps/rejected": -33.57072448730469,
|
| 2296 |
+
"loss": 0.0993,
|
| 2297 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2298 |
+
"rewards/chosen": 0.3945205807685852,
|
| 2299 |
+
"rewards/margins": 0.4975336194038391,
|
| 2300 |
+
"rewards/rejected": -0.10301297903060913,
|
| 2301 |
+
"step": 1400
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 3.64,
|
| 2305 |
+
"eval_logits/chosen": -2.247082471847534,
|
| 2306 |
+
"eval_logits/rejected": -2.242189884185791,
|
| 2307 |
+
"eval_logps/chosen": -33.95657730102539,
|
| 2308 |
+
"eval_logps/rejected": -37.431129455566406,
|
| 2309 |
+
"eval_loss": 0.6129570007324219,
|
| 2310 |
+
"eval_rewards/accuracies": 0.4871262311935425,
|
| 2311 |
+
"eval_rewards/chosen": 0.05458100512623787,
|
| 2312 |
+
"eval_rewards/margins": -0.005263464525341988,
|
| 2313 |
+
"eval_rewards/rejected": 0.05984446778893471,
|
| 2314 |
+
"eval_runtime": 145.3553,
|
| 2315 |
+
"eval_samples_per_second": 2.36,
|
| 2316 |
+
"eval_steps_per_second": 0.296,
|
| 2317 |
+
"step": 1400
|
| 2318 |
+
},
|
| 2319 |
+
{
|
| 2320 |
+
"epoch": 3.66,
|
| 2321 |
+
"grad_norm": 64.5,
|
| 2322 |
+
"learning_rate": 1.0775199359171346e-07,
|
| 2323 |
+
"logits/chosen": -2.0669209957122803,
|
| 2324 |
+
"logits/rejected": -2.0600974559783936,
|
| 2325 |
+
"logps/chosen": -33.28879928588867,
|
| 2326 |
+
"logps/rejected": -34.330623626708984,
|
| 2327 |
+
"loss": 0.14,
|
| 2328 |
+
"rewards/accuracies": 0.9375,
|
| 2329 |
+
"rewards/chosen": 0.3596227765083313,
|
| 2330 |
+
"rewards/margins": 0.4198623299598694,
|
| 2331 |
+
"rewards/rejected": -0.06023954600095749,
|
| 2332 |
+
"step": 1410
|
| 2333 |
+
},
|
| 2334 |
+
{
|
| 2335 |
+
"epoch": 3.69,
|
| 2336 |
+
"grad_norm": 25.625,
|
| 2337 |
+
"learning_rate": 9.191080703056604e-08,
|
| 2338 |
+
"logits/chosen": -1.9519445896148682,
|
| 2339 |
+
"logits/rejected": -1.9626754522323608,
|
| 2340 |
+
"logps/chosen": -33.65924072265625,
|
| 2341 |
+
"logps/rejected": -33.730491638183594,
|
| 2342 |
+
"loss": 0.0849,
|
| 2343 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2344 |
+
"rewards/chosen": 0.34304356575012207,
|
| 2345 |
+
"rewards/margins": 0.4406338632106781,
|
| 2346 |
+
"rewards/rejected": -0.09759029000997543,
|
| 2347 |
+
"step": 1420
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 3.71,
|
| 2351 |
+
"grad_norm": 58.0,
|
| 2352 |
+
"learning_rate": 7.730678442730539e-08,
|
| 2353 |
+
"logits/chosen": -2.0385050773620605,
|
| 2354 |
+
"logits/rejected": -2.0503756999969482,
|
| 2355 |
+
"logps/chosen": -34.54597473144531,
|
| 2356 |
+
"logps/rejected": -33.77598571777344,
|
| 2357 |
+
"loss": 0.1741,
|
| 2358 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2359 |
+
"rewards/chosen": 0.36038076877593994,
|
| 2360 |
+
"rewards/margins": 0.4567357897758484,
|
| 2361 |
+
"rewards/rejected": -0.09635499119758606,
|
| 2362 |
+
"step": 1430
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 3.74,
|
| 2366 |
+
"grad_norm": 18.875,
|
| 2367 |
+
"learning_rate": 6.394742864787806e-08,
|
| 2368 |
+
"logits/chosen": -2.0281505584716797,
|
| 2369 |
+
"logits/rejected": -2.030089855194092,
|
| 2370 |
+
"logps/chosen": -32.70575714111328,
|
| 2371 |
+
"logps/rejected": -33.9802360534668,
|
| 2372 |
+
"loss": 0.1205,
|
| 2373 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2374 |
+
"rewards/chosen": 0.35183992981910706,
|
| 2375 |
+
"rewards/margins": 0.4360295832157135,
|
| 2376 |
+
"rewards/rejected": -0.08418963849544525,
|
| 2377 |
+
"step": 1440
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 3.77,
|
| 2381 |
+
"grad_norm": 20.5,
|
| 2382 |
+
"learning_rate": 5.183960310644748e-08,
|
| 2383 |
+
"logits/chosen": -2.0094611644744873,
|
| 2384 |
+
"logits/rejected": -2.0022339820861816,
|
| 2385 |
+
"logps/chosen": -35.01443099975586,
|
| 2386 |
+
"logps/rejected": -33.88014221191406,
|
| 2387 |
+
"loss": 0.1175,
|
| 2388 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2389 |
+
"rewards/chosen": 0.38772544264793396,
|
| 2390 |
+
"rewards/margins": 0.464539110660553,
|
| 2391 |
+
"rewards/rejected": -0.07681362330913544,
|
| 2392 |
+
"step": 1450
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 3.79,
|
| 2396 |
+
"grad_norm": 25.625,
|
| 2397 |
+
"learning_rate": 4.098952823928693e-08,
|
| 2398 |
+
"logits/chosen": -1.9764325618743896,
|
| 2399 |
+
"logits/rejected": -1.9825315475463867,
|
| 2400 |
+
"logps/chosen": -30.76479148864746,
|
| 2401 |
+
"logps/rejected": -33.833831787109375,
|
| 2402 |
+
"loss": 0.1383,
|
| 2403 |
+
"rewards/accuracies": 0.9375,
|
| 2404 |
+
"rewards/chosen": 0.31386876106262207,
|
| 2405 |
+
"rewards/margins": 0.4285518229007721,
|
| 2406 |
+
"rewards/rejected": -0.11468305438756943,
|
| 2407 |
+
"step": 1460
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 3.82,
|
| 2411 |
+
"grad_norm": 21.125,
|
| 2412 |
+
"learning_rate": 3.1402778309014284e-08,
|
| 2413 |
+
"logits/chosen": -1.923027753829956,
|
| 2414 |
+
"logits/rejected": -1.9207102060317993,
|
| 2415 |
+
"logps/chosen": -30.280017852783203,
|
| 2416 |
+
"logps/rejected": -30.681386947631836,
|
| 2417 |
+
"loss": 0.1347,
|
| 2418 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2419 |
+
"rewards/chosen": 0.3958771824836731,
|
| 2420 |
+
"rewards/margins": 0.4568881392478943,
|
| 2421 |
+
"rewards/rejected": -0.061010997742414474,
|
| 2422 |
+
"step": 1470
|
| 2423 |
+
},
|
| 2424 |
+
{
|
| 2425 |
+
"epoch": 3.84,
|
| 2426 |
+
"grad_norm": 18.25,
|
| 2427 |
+
"learning_rate": 2.3084278540791427e-08,
|
| 2428 |
+
"logits/chosen": -2.0964748859405518,
|
| 2429 |
+
"logits/rejected": -2.090756893157959,
|
| 2430 |
+
"logps/chosen": -34.767372131347656,
|
| 2431 |
+
"logps/rejected": -32.20579528808594,
|
| 2432 |
+
"loss": 0.1015,
|
| 2433 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2434 |
+
"rewards/chosen": 0.3625258803367615,
|
| 2435 |
+
"rewards/margins": 0.4668874740600586,
|
| 2436 |
+
"rewards/rejected": -0.1043616384267807,
|
| 2437 |
+
"step": 1480
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 3.87,
|
| 2441 |
+
"grad_norm": 21.375,
|
| 2442 |
+
"learning_rate": 1.6038302591975807e-08,
|
| 2443 |
+
"logits/chosen": -1.9879157543182373,
|
| 2444 |
+
"logits/rejected": -1.9899780750274658,
|
| 2445 |
+
"logps/chosen": -27.98676109313965,
|
| 2446 |
+
"logps/rejected": -28.119998931884766,
|
| 2447 |
+
"loss": 0.1002,
|
| 2448 |
+
"rewards/accuracies": 0.9375,
|
| 2449 |
+
"rewards/chosen": 0.34273290634155273,
|
| 2450 |
+
"rewards/margins": 0.41113558411598206,
|
| 2451 |
+
"rewards/rejected": -0.06840268522500992,
|
| 2452 |
+
"step": 1490
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 3.9,
|
| 2456 |
+
"grad_norm": 59.75,
|
| 2457 |
+
"learning_rate": 1.0268470356514237e-08,
|
| 2458 |
+
"logits/chosen": -1.9974746704101562,
|
| 2459 |
+
"logits/rejected": -1.9917857646942139,
|
| 2460 |
+
"logps/chosen": -32.35822677612305,
|
| 2461 |
+
"logps/rejected": -32.09123611450195,
|
| 2462 |
+
"loss": 0.18,
|
| 2463 |
+
"rewards/accuracies": 0.9375,
|
| 2464 |
+
"rewards/chosen": 0.3097951412200928,
|
| 2465 |
+
"rewards/margins": 0.3953918516635895,
|
| 2466 |
+
"rewards/rejected": -0.0855967253446579,
|
| 2467 |
+
"step": 1500
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 3.9,
|
| 2471 |
+
"eval_logits/chosen": -2.2471930980682373,
|
| 2472 |
+
"eval_logits/rejected": -2.242299795150757,
|
| 2473 |
+
"eval_logps/chosen": -33.923362731933594,
|
| 2474 |
+
"eval_logps/rejected": -37.430606842041016,
|
| 2475 |
+
"eval_loss": 0.556613028049469,
|
| 2476 |
+
"eval_rewards/accuracies": 0.5049834251403809,
|
| 2477 |
+
"eval_rewards/chosen": 0.07783377170562744,
|
| 2478 |
+
"eval_rewards/margins": 0.01762736216187477,
|
| 2479 |
+
"eval_rewards/rejected": 0.06020640954375267,
|
| 2480 |
+
"eval_runtime": 145.2496,
|
| 2481 |
+
"eval_samples_per_second": 2.361,
|
| 2482 |
+
"eval_steps_per_second": 0.296,
|
| 2483 |
+
"step": 1500
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 3.92,
|
| 2487 |
+
"grad_norm": 31.25,
|
| 2488 |
+
"learning_rate": 5.777746105209147e-09,
|
| 2489 |
+
"logits/chosen": -1.9237276315689087,
|
| 2490 |
+
"logits/rejected": -1.9281084537506104,
|
| 2491 |
+
"logps/chosen": -33.68144989013672,
|
| 2492 |
+
"logps/rejected": -34.5217399597168,
|
| 2493 |
+
"loss": 0.1599,
|
| 2494 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2495 |
+
"rewards/chosen": 0.34268397092819214,
|
| 2496 |
+
"rewards/margins": 0.4206327497959137,
|
| 2497 |
+
"rewards/rejected": -0.07794882357120514,
|
| 2498 |
+
"step": 1510
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 3.95,
|
| 2502 |
+
"grad_norm": 44.25,
|
| 2503 |
+
"learning_rate": 2.5684369628148352e-09,
|
| 2504 |
+
"logits/chosen": -1.9838800430297852,
|
| 2505 |
+
"logits/rejected": -1.9821369647979736,
|
| 2506 |
+
"logps/chosen": -30.222864151000977,
|
| 2507 |
+
"logps/rejected": -32.720550537109375,
|
| 2508 |
+
"loss": 0.1456,
|
| 2509 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2510 |
+
"rewards/chosen": 0.3579188287258148,
|
| 2511 |
+
"rewards/margins": 0.44949913024902344,
|
| 2512 |
+
"rewards/rejected": -0.09158027172088623,
|
| 2513 |
+
"step": 1520
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 3.97,
|
| 2517 |
+
"grad_norm": 40.75,
|
| 2518 |
+
"learning_rate": 6.421917227455999e-10,
|
| 2519 |
+
"logits/chosen": -2.0813286304473877,
|
| 2520 |
+
"logits/rejected": -2.0737102031707764,
|
| 2521 |
+
"logps/chosen": -27.602453231811523,
|
| 2522 |
+
"logps/rejected": -28.810745239257812,
|
| 2523 |
+
"loss": 0.091,
|
| 2524 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2525 |
+
"rewards/chosen": 0.32234111428260803,
|
| 2526 |
+
"rewards/margins": 0.41973695158958435,
|
| 2527 |
+
"rewards/rejected": -0.09739579260349274,
|
| 2528 |
+
"step": 1530
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 4.0,
|
| 2532 |
+
"grad_norm": 19.0,
|
| 2533 |
+
"learning_rate": 0.0,
|
| 2534 |
+
"logits/chosen": -1.979434609413147,
|
| 2535 |
+
"logits/rejected": -1.9692825078964233,
|
| 2536 |
+
"logps/chosen": -32.49375534057617,
|
| 2537 |
+
"logps/rejected": -35.5590705871582,
|
| 2538 |
+
"loss": 0.1005,
|
| 2539 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2540 |
+
"rewards/chosen": 0.40070921182632446,
|
| 2541 |
+
"rewards/margins": 0.5042387247085571,
|
| 2542 |
+
"rewards/rejected": -0.10352955758571625,
|
| 2543 |
+
"step": 1540
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 4.0,
|
| 2547 |
+
"step": 1540,
|
| 2548 |
"total_flos": 0.0,
|
| 2549 |
+
"train_loss": 0.20226844080856868,
|
| 2550 |
+
"train_runtime": 10768.956,
|
| 2551 |
+
"train_samples_per_second": 1.144,
|
| 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,
|