Model save
Browse files
README.md
ADDED
|
@@ -0,0 +1,83 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: mistralai/Mistral-7B-v0.1
|
| 3 |
+
library_name: peft
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- dpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: zephyr-7b-dpo-qlora
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# zephyr-7b-dpo-qlora
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.4952
|
| 22 |
+
- Rewards/chosen: -2.8107
|
| 23 |
+
- Rewards/rejected: -3.8708
|
| 24 |
+
- Rewards/accuracies: 0.7718
|
| 25 |
+
- Rewards/margins: 1.0601
|
| 26 |
+
- Logps/rejected: -631.7385
|
| 27 |
+
- Logps/chosen: -545.9743
|
| 28 |
+
- Logits/rejected: -1.0385
|
| 29 |
+
- Logits/chosen: -1.1509
|
| 30 |
+
|
| 31 |
+
## Model description
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Intended uses & limitations
|
| 36 |
+
|
| 37 |
+
More information needed
|
| 38 |
+
|
| 39 |
+
## Training and evaluation data
|
| 40 |
+
|
| 41 |
+
More information needed
|
| 42 |
+
|
| 43 |
+
## Training procedure
|
| 44 |
+
|
| 45 |
+
### Training hyperparameters
|
| 46 |
+
|
| 47 |
+
The following hyperparameters were used during training:
|
| 48 |
+
- learning_rate: 5e-06
|
| 49 |
+
- train_batch_size: 4
|
| 50 |
+
- eval_batch_size: 8
|
| 51 |
+
- seed: 42
|
| 52 |
+
- distributed_type: multi-GPU
|
| 53 |
+
- num_devices: 4
|
| 54 |
+
- gradient_accumulation_steps: 4
|
| 55 |
+
- total_train_batch_size: 64
|
| 56 |
+
- total_eval_batch_size: 32
|
| 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: 1
|
| 61 |
+
|
| 62 |
+
### Training results
|
| 63 |
+
|
| 64 |
+
| Training Loss | Epoch | Step | Logits/chosen | Logits/rejected | Logps/chosen | Logps/rejected | Validation Loss | Rewards/accuracies | Rewards/chosen | Rewards/margins | Rewards/rejected |
|
| 65 |
+
|:-------------:|:------:|:----:|:-------------:|:---------------:|:------------:|:--------------:|:---------------:|:------------------:|:--------------:|:---------------:|:----------------:|
|
| 66 |
+
| 0.6163 | 0.1047 | 100 | -2.1006 | -2.0162 | -303.8351 | -310.3097 | 0.6178 | 0.6806 | -0.3893 | 0.2672 | -0.6565 |
|
| 67 |
+
| 0.5679 | 0.2094 | 200 | -1.8227 | -1.7394 | -352.2879 | -389.6575 | 0.5567 | 0.7401 | -0.8739 | 0.5761 | -1.4500 |
|
| 68 |
+
| 0.5412 | 0.3141 | 300 | -1.3111 | -1.2181 | -421.3257 | -483.0423 | 0.5305 | 0.7460 | -1.5642 | 0.8196 | -2.3838 |
|
| 69 |
+
| 0.5364 | 0.4187 | 400 | -1.2334 | -1.1332 | -416.6979 | -476.3458 | 0.5143 | 0.7579 | -1.5180 | 0.7989 | -2.3169 |
|
| 70 |
+
| 0.5046 | 0.5234 | 500 | -1.1373 | -1.0302 | -529.9542 | -605.2977 | 0.5062 | 0.7579 | -2.6505 | 0.9559 | -3.6064 |
|
| 71 |
+
| 0.4736 | 0.6281 | 600 | 0.5059 | -2.7244 | -3.7650 | 0.7639 | 1.0406 | -621.1549 | -537.3406 | -1.0135 | -1.1253 |
|
| 72 |
+
| 0.4619 | 0.7328 | 700 | 0.4994 | -2.9240 | -3.9991 | 0.7619 | 1.0750 | -644.5651 | -557.3041 | -1.0064 | -1.1194 |
|
| 73 |
+
| 0.4926 | 0.8375 | 800 | 0.4962 | -2.7247 | -3.7455 | 0.7659 | 1.0207 | -619.2051 | -537.3770 | -1.0516 | -1.1641 |
|
| 74 |
+
| 0.4856 | 0.9422 | 900 | 0.4952 | -2.8107 | -3.8708 | 0.7718 | 1.0601 | -631.7385 | -545.9743 | -1.0385 | -1.1509 |
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
### Framework versions
|
| 78 |
+
|
| 79 |
+
- PEFT 0.12.0
|
| 80 |
+
- Transformers 4.44.2
|
| 81 |
+
- Pytorch 2.4.0
|
| 82 |
+
- Datasets 2.21.0
|
| 83 |
+
- Tokenizers 0.19.1
|
adapter_model.safetensors
CHANGED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:16eb89e252e4ccfc437718d2f729474d0f3b93de224dbf67bc157b3f99bbe96c
|
| 3 |
+
size 671150064
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9997382884061764,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.2358125359600127,
|
| 5 |
+
"train_runtime": 19082.7985,
|
| 6 |
+
"train_samples": 61134,
|
| 7 |
+
"train_samples_per_second": 3.204,
|
| 8 |
+
"train_steps_per_second": 0.05
|
| 9 |
+
}
|
runs/Sep12_12-22-03_qa-a40-005.crc.nd.edu/events.out.tfevents.1726158593.qa-a40-005.crc.nd.edu.408111.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a98a894bc09c2b34903e8ab4c2119f67724d8810ecf18d5e86e5f236863456af
|
| 3 |
+
size 40527
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9997382884061764,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.2358125359600127,
|
| 5 |
+
"train_runtime": 19082.7985,
|
| 6 |
+
"train_samples": 61134,
|
| 7 |
+
"train_samples_per_second": 3.204,
|
| 8 |
+
"train_steps_per_second": 0.05
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1626 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9997382884061764,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 955,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0010468463752944255,
|
| 13 |
+
"grad_norm": 1.1940648296392757,
|
| 14 |
+
"learning_rate": 5.208333333333333e-08,
|
| 15 |
+
"logits/chosen": -2.519019603729248,
|
| 16 |
+
"logits/rejected": -2.354379177093506,
|
| 17 |
+
"logps/chosen": -297.6008605957031,
|
| 18 |
+
"logps/rejected": -252.44248962402344,
|
| 19 |
+
"loss": 0.693,
|
| 20 |
+
"rewards/accuracies": 0.5625,
|
| 21 |
+
"rewards/chosen": 0.0007321774610318244,
|
| 22 |
+
"rewards/margins": 6.297111394815147e-05,
|
| 23 |
+
"rewards/rejected": 0.0006692063761875033,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.010468463752944255,
|
| 28 |
+
"grad_norm": 1.106408968948395,
|
| 29 |
+
"learning_rate": 5.208333333333334e-07,
|
| 30 |
+
"logits/chosen": -2.2454488277435303,
|
| 31 |
+
"logits/rejected": -2.215104818344116,
|
| 32 |
+
"logps/chosen": -275.67840576171875,
|
| 33 |
+
"logps/rejected": -254.77935791015625,
|
| 34 |
+
"loss": 0.6927,
|
| 35 |
+
"rewards/accuracies": 0.5694444179534912,
|
| 36 |
+
"rewards/chosen": 0.004385500214993954,
|
| 37 |
+
"rewards/margins": 0.0008544763550162315,
|
| 38 |
+
"rewards/rejected": 0.0035310229286551476,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.02093692750588851,
|
| 43 |
+
"grad_norm": 1.1723169233645143,
|
| 44 |
+
"learning_rate": 1.0416666666666667e-06,
|
| 45 |
+
"logits/chosen": -2.231245517730713,
|
| 46 |
+
"logits/rejected": -2.1146271228790283,
|
| 47 |
+
"logps/chosen": -277.5929260253906,
|
| 48 |
+
"logps/rejected": -255.2114715576172,
|
| 49 |
+
"loss": 0.6907,
|
| 50 |
+
"rewards/accuracies": 0.643750011920929,
|
| 51 |
+
"rewards/chosen": 0.026119515299797058,
|
| 52 |
+
"rewards/margins": 0.005599636118859053,
|
| 53 |
+
"rewards/rejected": 0.020519878715276718,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.031405391258832765,
|
| 58 |
+
"grad_norm": 1.1930628386131166,
|
| 59 |
+
"learning_rate": 1.5625e-06,
|
| 60 |
+
"logits/chosen": -2.314044952392578,
|
| 61 |
+
"logits/rejected": -2.211050510406494,
|
| 62 |
+
"logps/chosen": -281.39801025390625,
|
| 63 |
+
"logps/rejected": -262.43841552734375,
|
| 64 |
+
"loss": 0.6858,
|
| 65 |
+
"rewards/accuracies": 0.643750011920929,
|
| 66 |
+
"rewards/chosen": 0.042405955493450165,
|
| 67 |
+
"rewards/margins": 0.015279242768883705,
|
| 68 |
+
"rewards/rejected": 0.027126718312501907,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.04187385501177702,
|
| 73 |
+
"grad_norm": 1.1771194457412584,
|
| 74 |
+
"learning_rate": 2.0833333333333334e-06,
|
| 75 |
+
"logits/chosen": -2.3085246086120605,
|
| 76 |
+
"logits/rejected": -2.2145581245422363,
|
| 77 |
+
"logps/chosen": -268.41571044921875,
|
| 78 |
+
"logps/rejected": -255.4882354736328,
|
| 79 |
+
"loss": 0.6814,
|
| 80 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 81 |
+
"rewards/chosen": 0.05249170586466789,
|
| 82 |
+
"rewards/margins": 0.027045782655477524,
|
| 83 |
+
"rewards/rejected": 0.025445926934480667,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.05234231876472128,
|
| 88 |
+
"grad_norm": 1.1817503689886124,
|
| 89 |
+
"learning_rate": 2.604166666666667e-06,
|
| 90 |
+
"logits/chosen": -2.279360294342041,
|
| 91 |
+
"logits/rejected": -2.1755104064941406,
|
| 92 |
+
"logps/chosen": -227.83438110351562,
|
| 93 |
+
"logps/rejected": -206.697265625,
|
| 94 |
+
"loss": 0.6761,
|
| 95 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 96 |
+
"rewards/chosen": 0.05746235325932503,
|
| 97 |
+
"rewards/margins": 0.04355225712060928,
|
| 98 |
+
"rewards/rejected": 0.013910098001360893,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.06281078251766553,
|
| 103 |
+
"grad_norm": 1.318174585891366,
|
| 104 |
+
"learning_rate": 3.125e-06,
|
| 105 |
+
"logits/chosen": -2.287376880645752,
|
| 106 |
+
"logits/rejected": -2.186383008956909,
|
| 107 |
+
"logps/chosen": -264.8748474121094,
|
| 108 |
+
"logps/rejected": -229.34848022460938,
|
| 109 |
+
"loss": 0.6681,
|
| 110 |
+
"rewards/accuracies": 0.6875,
|
| 111 |
+
"rewards/chosen": 0.045027900487184525,
|
| 112 |
+
"rewards/margins": 0.06142578274011612,
|
| 113 |
+
"rewards/rejected": -0.016397882252931595,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.07327924627060979,
|
| 118 |
+
"grad_norm": 1.5287003725547763,
|
| 119 |
+
"learning_rate": 3.6458333333333333e-06,
|
| 120 |
+
"logits/chosen": -2.129021406173706,
|
| 121 |
+
"logits/rejected": -2.0812830924987793,
|
| 122 |
+
"logps/chosen": -257.1100158691406,
|
| 123 |
+
"logps/rejected": -263.87359619140625,
|
| 124 |
+
"loss": 0.6544,
|
| 125 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 126 |
+
"rewards/chosen": -0.00797443836927414,
|
| 127 |
+
"rewards/margins": 0.10772331804037094,
|
| 128 |
+
"rewards/rejected": -0.11569775640964508,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.08374771002355404,
|
| 133 |
+
"grad_norm": 2.951307924811214,
|
| 134 |
+
"learning_rate": 4.166666666666667e-06,
|
| 135 |
+
"logits/chosen": -2.2603583335876465,
|
| 136 |
+
"logits/rejected": -2.1030356884002686,
|
| 137 |
+
"logps/chosen": -265.6451110839844,
|
| 138 |
+
"logps/rejected": -258.9366455078125,
|
| 139 |
+
"loss": 0.6386,
|
| 140 |
+
"rewards/accuracies": 0.699999988079071,
|
| 141 |
+
"rewards/chosen": -0.1463477909564972,
|
| 142 |
+
"rewards/margins": 0.12257362902164459,
|
| 143 |
+
"rewards/rejected": -0.268921434879303,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.0942161737764983,
|
| 148 |
+
"grad_norm": 2.228941966023797,
|
| 149 |
+
"learning_rate": 4.6875000000000004e-06,
|
| 150 |
+
"logits/chosen": -2.1493210792541504,
|
| 151 |
+
"logits/rejected": -2.081740140914917,
|
| 152 |
+
"logps/chosen": -267.15631103515625,
|
| 153 |
+
"logps/rejected": -286.8982238769531,
|
| 154 |
+
"loss": 0.6314,
|
| 155 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 156 |
+
"rewards/chosen": -0.1616009771823883,
|
| 157 |
+
"rewards/margins": 0.18432030081748962,
|
| 158 |
+
"rewards/rejected": -0.34592124819755554,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.10468463752944256,
|
| 163 |
+
"grad_norm": 2.3171658298596007,
|
| 164 |
+
"learning_rate": 4.9997324926814375e-06,
|
| 165 |
+
"logits/chosen": -2.198456287384033,
|
| 166 |
+
"logits/rejected": -2.08889102935791,
|
| 167 |
+
"logps/chosen": -310.09893798828125,
|
| 168 |
+
"logps/rejected": -328.76763916015625,
|
| 169 |
+
"loss": 0.6163,
|
| 170 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 171 |
+
"rewards/chosen": -0.37811416387557983,
|
| 172 |
+
"rewards/margins": 0.23748056590557098,
|
| 173 |
+
"rewards/rejected": -0.615594744682312,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.10468463752944256,
|
| 178 |
+
"eval_logits/chosen": -2.1005911827087402,
|
| 179 |
+
"eval_logits/rejected": -2.0161635875701904,
|
| 180 |
+
"eval_logps/chosen": -303.8350524902344,
|
| 181 |
+
"eval_logps/rejected": -310.3097229003906,
|
| 182 |
+
"eval_loss": 0.6178256869316101,
|
| 183 |
+
"eval_rewards/accuracies": 0.6805555820465088,
|
| 184 |
+
"eval_rewards/chosen": -0.38932088017463684,
|
| 185 |
+
"eval_rewards/margins": 0.26718538999557495,
|
| 186 |
+
"eval_rewards/rejected": -0.6565062999725342,
|
| 187 |
+
"eval_runtime": 497.4053,
|
| 188 |
+
"eval_samples_per_second": 4.021,
|
| 189 |
+
"eval_steps_per_second": 0.127,
|
| 190 |
+
"step": 100
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.11515310128238682,
|
| 194 |
+
"grad_norm": 3.2280840758041305,
|
| 195 |
+
"learning_rate": 4.996723692767927e-06,
|
| 196 |
+
"logits/chosen": -2.185472011566162,
|
| 197 |
+
"logits/rejected": -2.0565757751464844,
|
| 198 |
+
"logps/chosen": -283.398681640625,
|
| 199 |
+
"logps/rejected": -271.3462219238281,
|
| 200 |
+
"loss": 0.6159,
|
| 201 |
+
"rewards/accuracies": 0.737500011920929,
|
| 202 |
+
"rewards/chosen": -0.35291650891304016,
|
| 203 |
+
"rewards/margins": 0.27926865220069885,
|
| 204 |
+
"rewards/rejected": -0.632185161113739,
|
| 205 |
+
"step": 110
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.12562156503533106,
|
| 209 |
+
"grad_norm": 4.321662600064664,
|
| 210 |
+
"learning_rate": 4.9903757462135984e-06,
|
| 211 |
+
"logits/chosen": -2.1495065689086914,
|
| 212 |
+
"logits/rejected": -2.0793585777282715,
|
| 213 |
+
"logps/chosen": -293.3641662597656,
|
| 214 |
+
"logps/rejected": -339.5681457519531,
|
| 215 |
+
"loss": 0.5751,
|
| 216 |
+
"rewards/accuracies": 0.75,
|
| 217 |
+
"rewards/chosen": -0.4733458161354065,
|
| 218 |
+
"rewards/margins": 0.3658001124858856,
|
| 219 |
+
"rewards/rejected": -0.8391459584236145,
|
| 220 |
+
"step": 120
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.1360900287882753,
|
| 224 |
+
"grad_norm": 5.281496669320805,
|
| 225 |
+
"learning_rate": 4.980697142834315e-06,
|
| 226 |
+
"logits/chosen": -2.118959426879883,
|
| 227 |
+
"logits/rejected": -2.0077736377716064,
|
| 228 |
+
"logps/chosen": -376.8563232421875,
|
| 229 |
+
"logps/rejected": -348.50531005859375,
|
| 230 |
+
"loss": 0.5843,
|
| 231 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 232 |
+
"rewards/chosen": -0.770808756351471,
|
| 233 |
+
"rewards/margins": 0.335581511259079,
|
| 234 |
+
"rewards/rejected": -1.1063902378082275,
|
| 235 |
+
"step": 130
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.14655849254121958,
|
| 239 |
+
"grad_norm": 5.5524041616811255,
|
| 240 |
+
"learning_rate": 4.967700826904229e-06,
|
| 241 |
+
"logits/chosen": -2.0535736083984375,
|
| 242 |
+
"logits/rejected": -1.9956505298614502,
|
| 243 |
+
"logps/chosen": -353.10333251953125,
|
| 244 |
+
"logps/rejected": -393.17108154296875,
|
| 245 |
+
"loss": 0.5757,
|
| 246 |
+
"rewards/accuracies": 0.762499988079071,
|
| 247 |
+
"rewards/chosen": -0.8411790132522583,
|
| 248 |
+
"rewards/margins": 0.44028061628341675,
|
| 249 |
+
"rewards/rejected": -1.2814596891403198,
|
| 250 |
+
"step": 140
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.15702695629416383,
|
| 254 |
+
"grad_norm": 4.851579207894888,
|
| 255 |
+
"learning_rate": 4.951404179843963e-06,
|
| 256 |
+
"logits/chosen": -2.1413304805755615,
|
| 257 |
+
"logits/rejected": -2.0062034130096436,
|
| 258 |
+
"logps/chosen": -385.7857360839844,
|
| 259 |
+
"logps/rejected": -377.45819091796875,
|
| 260 |
+
"loss": 0.5887,
|
| 261 |
+
"rewards/accuracies": 0.75,
|
| 262 |
+
"rewards/chosen": -0.9732216000556946,
|
| 263 |
+
"rewards/margins": 0.39891117811203003,
|
| 264 |
+
"rewards/rejected": -1.3721327781677246,
|
| 265 |
+
"step": 150
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.16749542004710807,
|
| 269 |
+
"grad_norm": 7.670497596403814,
|
| 270 |
+
"learning_rate": 4.931828996974498e-06,
|
| 271 |
+
"logits/chosen": -2.0889430046081543,
|
| 272 |
+
"logits/rejected": -1.922328233718872,
|
| 273 |
+
"logps/chosen": -347.14544677734375,
|
| 274 |
+
"logps/rejected": -342.8069763183594,
|
| 275 |
+
"loss": 0.5572,
|
| 276 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 277 |
+
"rewards/chosen": -0.606709361076355,
|
| 278 |
+
"rewards/margins": 0.5376572608947754,
|
| 279 |
+
"rewards/rejected": -1.1443665027618408,
|
| 280 |
+
"step": 160
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.17796388380005235,
|
| 284 |
+
"grad_norm": 5.79406761477676,
|
| 285 |
+
"learning_rate": 4.909001458367867e-06,
|
| 286 |
+
"logits/chosen": -1.9932262897491455,
|
| 287 |
+
"logits/rejected": -1.8520256280899048,
|
| 288 |
+
"logps/chosen": -393.05499267578125,
|
| 289 |
+
"logps/rejected": -439.379150390625,
|
| 290 |
+
"loss": 0.5534,
|
| 291 |
+
"rewards/accuracies": 0.65625,
|
| 292 |
+
"rewards/chosen": -1.3342100381851196,
|
| 293 |
+
"rewards/margins": 0.6197064518928528,
|
| 294 |
+
"rewards/rejected": -1.9539167881011963,
|
| 295 |
+
"step": 170
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.1884323475529966,
|
| 299 |
+
"grad_norm": 5.16143784148925,
|
| 300 |
+
"learning_rate": 4.882952093833628e-06,
|
| 301 |
+
"logits/chosen": -1.9763168096542358,
|
| 302 |
+
"logits/rejected": -1.9191144704818726,
|
| 303 |
+
"logps/chosen": -370.9029235839844,
|
| 304 |
+
"logps/rejected": -421.9246520996094,
|
| 305 |
+
"loss": 0.5354,
|
| 306 |
+
"rewards/accuracies": 0.699999988079071,
|
| 307 |
+
"rewards/chosen": -1.0804250240325928,
|
| 308 |
+
"rewards/margins": 0.5884348154067993,
|
| 309 |
+
"rewards/rejected": -1.668859839439392,
|
| 310 |
+
"step": 180
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.19890081130594087,
|
| 314 |
+
"grad_norm": 4.841208626728555,
|
| 315 |
+
"learning_rate": 4.853715742087947e-06,
|
| 316 |
+
"logits/chosen": -1.8281657695770264,
|
| 317 |
+
"logits/rejected": -1.785474419593811,
|
| 318 |
+
"logps/chosen": -358.50970458984375,
|
| 319 |
+
"logps/rejected": -449.71478271484375,
|
| 320 |
+
"loss": 0.5327,
|
| 321 |
+
"rewards/accuracies": 0.8125,
|
| 322 |
+
"rewards/chosen": -1.1242244243621826,
|
| 323 |
+
"rewards/margins": 0.7143529653549194,
|
| 324 |
+
"rewards/rejected": -1.8385772705078125,
|
| 325 |
+
"step": 190
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.2093692750588851,
|
| 329 |
+
"grad_norm": 4.320034048961276,
|
| 330 |
+
"learning_rate": 4.821331504159906e-06,
|
| 331 |
+
"logits/chosen": -1.9889659881591797,
|
| 332 |
+
"logits/rejected": -1.8328838348388672,
|
| 333 |
+
"logps/chosen": -417.7860412597656,
|
| 334 |
+
"logps/rejected": -414.47528076171875,
|
| 335 |
+
"loss": 0.5679,
|
| 336 |
+
"rewards/accuracies": 0.71875,
|
| 337 |
+
"rewards/chosen": -1.052539587020874,
|
| 338 |
+
"rewards/margins": 0.5909037590026855,
|
| 339 |
+
"rewards/rejected": -1.6434433460235596,
|
| 340 |
+
"step": 200
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.2093692750588851,
|
| 344 |
+
"eval_logits/chosen": -1.8226656913757324,
|
| 345 |
+
"eval_logits/rejected": -1.73935866355896,
|
| 346 |
+
"eval_logps/chosen": -352.2879333496094,
|
| 347 |
+
"eval_logps/rejected": -389.65753173828125,
|
| 348 |
+
"eval_loss": 0.5566642880439758,
|
| 349 |
+
"eval_rewards/accuracies": 0.7400793433189392,
|
| 350 |
+
"eval_rewards/chosen": -0.8738502264022827,
|
| 351 |
+
"eval_rewards/margins": 0.5761341452598572,
|
| 352 |
+
"eval_rewards/rejected": -1.4499843120574951,
|
| 353 |
+
"eval_runtime": 495.1318,
|
| 354 |
+
"eval_samples_per_second": 4.039,
|
| 355 |
+
"eval_steps_per_second": 0.127,
|
| 356 |
+
"step": 200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.21983773881182936,
|
| 360 |
+
"grad_norm": 4.600314274109389,
|
| 361 |
+
"learning_rate": 4.7858426910973435e-06,
|
| 362 |
+
"logits/chosen": -1.884235143661499,
|
| 363 |
+
"logits/rejected": -1.8107131719589233,
|
| 364 |
+
"logps/chosen": -395.2080993652344,
|
| 365 |
+
"logps/rejected": -432.21160888671875,
|
| 366 |
+
"loss": 0.5779,
|
| 367 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 368 |
+
"rewards/chosen": -1.1761962175369263,
|
| 369 |
+
"rewards/margins": 0.46132412552833557,
|
| 370 |
+
"rewards/rejected": -1.637520432472229,
|
| 371 |
+
"step": 210
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.23030620256477363,
|
| 375 |
+
"grad_norm": 5.190566663970971,
|
| 376 |
+
"learning_rate": 4.747296766042161e-06,
|
| 377 |
+
"logits/chosen": -1.8367938995361328,
|
| 378 |
+
"logits/rejected": -1.7323997020721436,
|
| 379 |
+
"logps/chosen": -448.40631103515625,
|
| 380 |
+
"logps/rejected": -463.5052185058594,
|
| 381 |
+
"loss": 0.5486,
|
| 382 |
+
"rewards/accuracies": 0.731249988079071,
|
| 383 |
+
"rewards/chosen": -1.5517961978912354,
|
| 384 |
+
"rewards/margins": 0.5995836853981018,
|
| 385 |
+
"rewards/rejected": -2.1513800621032715,
|
| 386 |
+
"step": 220
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.24077466631771788,
|
| 390 |
+
"grad_norm": 6.503627910465957,
|
| 391 |
+
"learning_rate": 4.705745280752586e-06,
|
| 392 |
+
"logits/chosen": -1.6796023845672607,
|
| 393 |
+
"logits/rejected": -1.613417387008667,
|
| 394 |
+
"logps/chosen": -363.4635009765625,
|
| 395 |
+
"logps/rejected": -397.3682861328125,
|
| 396 |
+
"loss": 0.5564,
|
| 397 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 398 |
+
"rewards/chosen": -1.1600974798202515,
|
| 399 |
+
"rewards/margins": 0.502653956413269,
|
| 400 |
+
"rewards/rejected": -1.66275155544281,
|
| 401 |
+
"step": 230
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.2512431300706621,
|
| 405 |
+
"grad_norm": 5.303084146530296,
|
| 406 |
+
"learning_rate": 4.661243806657256e-06,
|
| 407 |
+
"logits/chosen": -1.7421451807022095,
|
| 408 |
+
"logits/rejected": -1.6884396076202393,
|
| 409 |
+
"logps/chosen": -386.07489013671875,
|
| 410 |
+
"logps/rejected": -414.7723693847656,
|
| 411 |
+
"loss": 0.5355,
|
| 412 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 413 |
+
"rewards/chosen": -1.2963042259216309,
|
| 414 |
+
"rewards/margins": 0.5929707884788513,
|
| 415 |
+
"rewards/rejected": -1.8892749547958374,
|
| 416 |
+
"step": 240
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.26171159382360637,
|
| 420 |
+
"grad_norm": 4.582341616936657,
|
| 421 |
+
"learning_rate": 4.613851860533367e-06,
|
| 422 |
+
"logits/chosen": -1.766645073890686,
|
| 423 |
+
"logits/rejected": -1.6948877573013306,
|
| 424 |
+
"logps/chosen": -388.61328125,
|
| 425 |
+
"logps/rejected": -415.88690185546875,
|
| 426 |
+
"loss": 0.5156,
|
| 427 |
+
"rewards/accuracies": 0.793749988079071,
|
| 428 |
+
"rewards/chosen": -1.220664620399475,
|
| 429 |
+
"rewards/margins": 0.6846394538879395,
|
| 430 |
+
"rewards/rejected": -1.905303955078125,
|
| 431 |
+
"step": 250
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.2721800575765506,
|
| 435 |
+
"grad_norm": 5.258036005468907,
|
| 436 |
+
"learning_rate": 4.563632824908252e-06,
|
| 437 |
+
"logits/chosen": -1.777336835861206,
|
| 438 |
+
"logits/rejected": -1.651208519935608,
|
| 439 |
+
"logps/chosen": -376.238525390625,
|
| 440 |
+
"logps/rejected": -421.6551818847656,
|
| 441 |
+
"loss": 0.5197,
|
| 442 |
+
"rewards/accuracies": 0.668749988079071,
|
| 443 |
+
"rewards/chosen": -1.0568403005599976,
|
| 444 |
+
"rewards/margins": 0.6248779296875,
|
| 445 |
+
"rewards/rejected": -1.6817182302474976,
|
| 446 |
+
"step": 260
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.2826485213294949,
|
| 450 |
+
"grad_norm": 4.95113233440466,
|
| 451 |
+
"learning_rate": 4.510653863290871e-06,
|
| 452 |
+
"logits/chosen": -1.6241214275360107,
|
| 453 |
+
"logits/rejected": -1.5427122116088867,
|
| 454 |
+
"logps/chosen": -393.2299499511719,
|
| 455 |
+
"logps/rejected": -455.7496032714844,
|
| 456 |
+
"loss": 0.5353,
|
| 457 |
+
"rewards/accuracies": 0.731249988079071,
|
| 458 |
+
"rewards/chosen": -1.1928220987319946,
|
| 459 |
+
"rewards/margins": 0.8355104327201843,
|
| 460 |
+
"rewards/rejected": -2.028332471847534,
|
| 461 |
+
"step": 270
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.29311698508243916,
|
| 465 |
+
"grad_norm": 4.316588759606841,
|
| 466 |
+
"learning_rate": 4.454985830346574e-06,
|
| 467 |
+
"logits/chosen": -1.6576063632965088,
|
| 468 |
+
"logits/rejected": -1.546891212463379,
|
| 469 |
+
"logps/chosen": -402.0686340332031,
|
| 470 |
+
"logps/rejected": -449.80694580078125,
|
| 471 |
+
"loss": 0.5469,
|
| 472 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 473 |
+
"rewards/chosen": -1.1402246952056885,
|
| 474 |
+
"rewards/margins": 0.8147087097167969,
|
| 475 |
+
"rewards/rejected": -1.9549334049224854,
|
| 476 |
+
"step": 280
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.3035854488353834,
|
| 480 |
+
"grad_norm": 4.334299410781991,
|
| 481 |
+
"learning_rate": 4.396703177135262e-06,
|
| 482 |
+
"logits/chosen": -1.7125823497772217,
|
| 483 |
+
"logits/rejected": -1.547957181930542,
|
| 484 |
+
"logps/chosen": -367.20501708984375,
|
| 485 |
+
"logps/rejected": -389.2141418457031,
|
| 486 |
+
"loss": 0.5292,
|
| 487 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 488 |
+
"rewards/chosen": -0.8416954278945923,
|
| 489 |
+
"rewards/margins": 0.5679039359092712,
|
| 490 |
+
"rewards/rejected": -1.4095993041992188,
|
| 491 |
+
"step": 290
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.31405391258832765,
|
| 495 |
+
"grad_norm": 5.689892772495371,
|
| 496 |
+
"learning_rate": 4.335883851539693e-06,
|
| 497 |
+
"logits/chosen": -1.4305754899978638,
|
| 498 |
+
"logits/rejected": -1.4178600311279297,
|
| 499 |
+
"logps/chosen": -366.5882263183594,
|
| 500 |
+
"logps/rejected": -427.017333984375,
|
| 501 |
+
"loss": 0.5412,
|
| 502 |
+
"rewards/accuracies": 0.699999988079071,
|
| 503 |
+
"rewards/chosen": -1.239133596420288,
|
| 504 |
+
"rewards/margins": 0.6136574149131775,
|
| 505 |
+
"rewards/rejected": -1.8527911901474,
|
| 506 |
+
"step": 300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.31405391258832765,
|
| 510 |
+
"eval_logits/chosen": -1.3111425638198853,
|
| 511 |
+
"eval_logits/rejected": -1.218103051185608,
|
| 512 |
+
"eval_logps/chosen": -421.32574462890625,
|
| 513 |
+
"eval_logps/rejected": -483.0422668457031,
|
| 514 |
+
"eval_loss": 0.5305107831954956,
|
| 515 |
+
"eval_rewards/accuracies": 0.7460317611694336,
|
| 516 |
+
"eval_rewards/chosen": -1.5642281770706177,
|
| 517 |
+
"eval_rewards/margins": 0.819603681564331,
|
| 518 |
+
"eval_rewards/rejected": -2.3838319778442383,
|
| 519 |
+
"eval_runtime": 495.2081,
|
| 520 |
+
"eval_samples_per_second": 4.039,
|
| 521 |
+
"eval_steps_per_second": 0.127,
|
| 522 |
+
"step": 300
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.3245223763412719,
|
| 526 |
+
"grad_norm": 5.39322804571776,
|
| 527 |
+
"learning_rate": 4.2726091940171055e-06,
|
| 528 |
+
"logits/chosen": -1.4802881479263306,
|
| 529 |
+
"logits/rejected": -1.3236931562423706,
|
| 530 |
+
"logps/chosen": -424.968505859375,
|
| 531 |
+
"logps/rejected": -472.8589782714844,
|
| 532 |
+
"loss": 0.5441,
|
| 533 |
+
"rewards/accuracies": 0.731249988079071,
|
| 534 |
+
"rewards/chosen": -1.5254794359207153,
|
| 535 |
+
"rewards/margins": 0.8688099980354309,
|
| 536 |
+
"rewards/rejected": -2.394289493560791,
|
| 537 |
+
"step": 310
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.33499084009421615,
|
| 541 |
+
"grad_norm": 5.55708768926168,
|
| 542 |
+
"learning_rate": 4.206963828813555e-06,
|
| 543 |
+
"logits/chosen": -1.5147395133972168,
|
| 544 |
+
"logits/rejected": -1.4260919094085693,
|
| 545 |
+
"logps/chosen": -356.251953125,
|
| 546 |
+
"logps/rejected": -424.2235412597656,
|
| 547 |
+
"loss": 0.5537,
|
| 548 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 549 |
+
"rewards/chosen": -1.1914732456207275,
|
| 550 |
+
"rewards/margins": 0.6315604448318481,
|
| 551 |
+
"rewards/rejected": -1.8230335712432861,
|
| 552 |
+
"step": 320
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 0.34545930384716045,
|
| 556 |
+
"grad_norm": 5.004482442410713,
|
| 557 |
+
"learning_rate": 4.139035550786495e-06,
|
| 558 |
+
"logits/chosen": -1.504990577697754,
|
| 559 |
+
"logits/rejected": -1.4216346740722656,
|
| 560 |
+
"logps/chosen": -406.04473876953125,
|
| 561 |
+
"logps/rejected": -461.72283935546875,
|
| 562 |
+
"loss": 0.5321,
|
| 563 |
+
"rewards/accuracies": 0.699999988079071,
|
| 564 |
+
"rewards/chosen": -1.5744965076446533,
|
| 565 |
+
"rewards/margins": 0.6061158776283264,
|
| 566 |
+
"rewards/rejected": -2.180612325668335,
|
| 567 |
+
"step": 330
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.3559277676001047,
|
| 571 |
+
"grad_norm": 5.1331172845370165,
|
| 572 |
+
"learning_rate": 4.068915207986931e-06,
|
| 573 |
+
"logits/chosen": -1.2623932361602783,
|
| 574 |
+
"logits/rejected": -1.1655136346817017,
|
| 575 |
+
"logps/chosen": -411.1018981933594,
|
| 576 |
+
"logps/rejected": -456.156982421875,
|
| 577 |
+
"loss": 0.5393,
|
| 578 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 579 |
+
"rewards/chosen": -1.6048648357391357,
|
| 580 |
+
"rewards/margins": 0.7221436500549316,
|
| 581 |
+
"rewards/rejected": -2.3270087242126465,
|
| 582 |
+
"step": 340
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 0.36639623135304894,
|
| 586 |
+
"grad_norm": 6.204494142283613,
|
| 587 |
+
"learning_rate": 3.996696580158211e-06,
|
| 588 |
+
"logits/chosen": -1.4823843240737915,
|
| 589 |
+
"logits/rejected": -1.3702657222747803,
|
| 590 |
+
"logps/chosen": -387.9534606933594,
|
| 591 |
+
"logps/rejected": -448.01373291015625,
|
| 592 |
+
"loss": 0.5072,
|
| 593 |
+
"rewards/accuracies": 0.699999988079071,
|
| 594 |
+
"rewards/chosen": -1.3335731029510498,
|
| 595 |
+
"rewards/margins": 0.7507054805755615,
|
| 596 |
+
"rewards/rejected": -2.0842788219451904,
|
| 597 |
+
"step": 350
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.3768646951059932,
|
| 601 |
+
"grad_norm": 8.049482757420606,
|
| 602 |
+
"learning_rate": 3.922476253313921e-06,
|
| 603 |
+
"logits/chosen": -1.307775855064392,
|
| 604 |
+
"logits/rejected": -1.24771249294281,
|
| 605 |
+
"logps/chosen": -400.76910400390625,
|
| 606 |
+
"logps/rejected": -469.23651123046875,
|
| 607 |
+
"loss": 0.4704,
|
| 608 |
+
"rewards/accuracies": 0.768750011920929,
|
| 609 |
+
"rewards/chosen": -1.4676017761230469,
|
| 610 |
+
"rewards/margins": 0.8654597997665405,
|
| 611 |
+
"rewards/rejected": -2.333061456680298,
|
| 612 |
+
"step": 360
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.38733315885893743,
|
| 616 |
+
"grad_norm": 5.519216288621626,
|
| 617 |
+
"learning_rate": 3.846353490562664e-06,
|
| 618 |
+
"logits/chosen": -1.2273913621902466,
|
| 619 |
+
"logits/rejected": -1.1667524576187134,
|
| 620 |
+
"logps/chosen": -445.7137756347656,
|
| 621 |
+
"logps/rejected": -534.3553466796875,
|
| 622 |
+
"loss": 0.5271,
|
| 623 |
+
"rewards/accuracies": 0.762499988079071,
|
| 624 |
+
"rewards/chosen": -2.012014865875244,
|
| 625 |
+
"rewards/margins": 1.0135688781738281,
|
| 626 |
+
"rewards/rejected": -3.0255837440490723,
|
| 627 |
+
"step": 370
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.39780162261188173,
|
| 631 |
+
"grad_norm": 6.046235636435534,
|
| 632 |
+
"learning_rate": 3.768430099352445e-06,
|
| 633 |
+
"logits/chosen": -1.2952733039855957,
|
| 634 |
+
"logits/rejected": -1.2378871440887451,
|
| 635 |
+
"logps/chosen": -393.07000732421875,
|
| 636 |
+
"logps/rejected": -465.0540466308594,
|
| 637 |
+
"loss": 0.5257,
|
| 638 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 639 |
+
"rewards/chosen": -1.3947498798370361,
|
| 640 |
+
"rewards/margins": 0.7023818492889404,
|
| 641 |
+
"rewards/rejected": -2.0971317291259766,
|
| 642 |
+
"step": 380
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 0.408270086364826,
|
| 646 |
+
"grad_norm": 3.5923613903948133,
|
| 647 |
+
"learning_rate": 3.6888102953122307e-06,
|
| 648 |
+
"logits/chosen": -1.3613311052322388,
|
| 649 |
+
"logits/rejected": -1.2925641536712646,
|
| 650 |
+
"logps/chosen": -355.1478576660156,
|
| 651 |
+
"logps/rejected": -389.9283752441406,
|
| 652 |
+
"loss": 0.5452,
|
| 653 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 654 |
+
"rewards/chosen": -0.9670610427856445,
|
| 655 |
+
"rewards/margins": 0.5416972637176514,
|
| 656 |
+
"rewards/rejected": -1.5087581872940063,
|
| 657 |
+
"step": 390
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.4187385501177702,
|
| 661 |
+
"grad_norm": 4.7487137488013484,
|
| 662 |
+
"learning_rate": 3.607600562872785e-06,
|
| 663 |
+
"logits/chosen": -1.3237230777740479,
|
| 664 |
+
"logits/rejected": -1.2487279176712036,
|
| 665 |
+
"logps/chosen": -384.8719177246094,
|
| 666 |
+
"logps/rejected": -442.74957275390625,
|
| 667 |
+
"loss": 0.5364,
|
| 668 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 669 |
+
"rewards/chosen": -1.4063746929168701,
|
| 670 |
+
"rewards/margins": 0.5768125057220459,
|
| 671 |
+
"rewards/rejected": -1.9831870794296265,
|
| 672 |
+
"step": 400
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 0.4187385501177702,
|
| 676 |
+
"eval_logits/chosen": -1.2333532571792603,
|
| 677 |
+
"eval_logits/rejected": -1.1332385540008545,
|
| 678 |
+
"eval_logps/chosen": -416.69793701171875,
|
| 679 |
+
"eval_logps/rejected": -476.3457946777344,
|
| 680 |
+
"eval_loss": 0.5142984986305237,
|
| 681 |
+
"eval_rewards/accuracies": 0.7579365372657776,
|
| 682 |
+
"eval_rewards/chosen": -1.5179502964019775,
|
| 683 |
+
"eval_rewards/margins": 0.7989169359207153,
|
| 684 |
+
"eval_rewards/rejected": -2.3168673515319824,
|
| 685 |
+
"eval_runtime": 495.3694,
|
| 686 |
+
"eval_samples_per_second": 4.037,
|
| 687 |
+
"eval_steps_per_second": 0.127,
|
| 688 |
+
"step": 400
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.42920701387071447,
|
| 692 |
+
"grad_norm": 4.606627296879372,
|
| 693 |
+
"learning_rate": 3.5249095128531863e-06,
|
| 694 |
+
"logits/chosen": -1.2254225015640259,
|
| 695 |
+
"logits/rejected": -1.1056945323944092,
|
| 696 |
+
"logps/chosen": -440.85369873046875,
|
| 697 |
+
"logps/rejected": -507.37774658203125,
|
| 698 |
+
"loss": 0.5273,
|
| 699 |
+
"rewards/accuracies": 0.75,
|
| 700 |
+
"rewards/chosen": -1.7427661418914795,
|
| 701 |
+
"rewards/margins": 0.9440004229545593,
|
| 702 |
+
"rewards/rejected": -2.6867663860321045,
|
| 703 |
+
"step": 410
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.4396754776236587,
|
| 707 |
+
"grad_norm": 7.32723857385592,
|
| 708 |
+
"learning_rate": 3.4408477372034743e-06,
|
| 709 |
+
"logits/chosen": -1.3876991271972656,
|
| 710 |
+
"logits/rejected": -1.1844953298568726,
|
| 711 |
+
"logps/chosen": -440.2545471191406,
|
| 712 |
+
"logps/rejected": -458.7481384277344,
|
| 713 |
+
"loss": 0.5242,
|
| 714 |
+
"rewards/accuracies": 0.706250011920929,
|
| 715 |
+
"rewards/chosen": -1.5724732875823975,
|
| 716 |
+
"rewards/margins": 0.7581513524055481,
|
| 717 |
+
"rewards/rejected": -2.330624580383301,
|
| 718 |
+
"step": 420
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.45014394137660296,
|
| 722 |
+
"grad_norm": 4.83507866233375,
|
| 723 |
+
"learning_rate": 3.355527661097728e-06,
|
| 724 |
+
"logits/chosen": -1.3145965337753296,
|
| 725 |
+
"logits/rejected": -1.2404569387435913,
|
| 726 |
+
"logps/chosen": -391.62640380859375,
|
| 727 |
+
"logps/rejected": -483.40020751953125,
|
| 728 |
+
"loss": 0.5203,
|
| 729 |
+
"rewards/accuracies": 0.75,
|
| 730 |
+
"rewards/chosen": -1.6367257833480835,
|
| 731 |
+
"rewards/margins": 0.8808576464653015,
|
| 732 |
+
"rewards/rejected": -2.5175833702087402,
|
| 733 |
+
"step": 430
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 0.46061240512954726,
|
| 737 |
+
"grad_norm": 6.295429995521184,
|
| 738 |
+
"learning_rate": 3.269063392575352e-06,
|
| 739 |
+
"logits/chosen": -1.246955156326294,
|
| 740 |
+
"logits/rejected": -1.2044976949691772,
|
| 741 |
+
"logps/chosen": -412.8096618652344,
|
| 742 |
+
"logps/rejected": -462.73175048828125,
|
| 743 |
+
"loss": 0.5309,
|
| 744 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 745 |
+
"rewards/chosen": -1.6489166021347046,
|
| 746 |
+
"rewards/margins": 0.686037540435791,
|
| 747 |
+
"rewards/rejected": -2.334954261779785,
|
| 748 |
+
"step": 440
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 0.4710808688824915,
|
| 752 |
+
"grad_norm": 8.155653541939053,
|
| 753 |
+
"learning_rate": 3.181570569931697e-06,
|
| 754 |
+
"logits/chosen": -1.3565101623535156,
|
| 755 |
+
"logits/rejected": -1.2913014888763428,
|
| 756 |
+
"logps/chosen": -404.30938720703125,
|
| 757 |
+
"logps/rejected": -522.0363159179688,
|
| 758 |
+
"loss": 0.5151,
|
| 759 |
+
"rewards/accuracies": 0.768750011920929,
|
| 760 |
+
"rewards/chosen": -1.5972925424575806,
|
| 761 |
+
"rewards/margins": 1.02085280418396,
|
| 762 |
+
"rewards/rejected": -2.61814546585083,
|
| 763 |
+
"step": 450
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 0.48154933263543576,
|
| 767 |
+
"grad_norm": 5.153431243466889,
|
| 768 |
+
"learning_rate": 3.09316620706208e-06,
|
| 769 |
+
"logits/chosen": -1.3433706760406494,
|
| 770 |
+
"logits/rejected": -1.2827776670455933,
|
| 771 |
+
"logps/chosen": -448.34283447265625,
|
| 772 |
+
"logps/rejected": -517.4175415039062,
|
| 773 |
+
"loss": 0.5305,
|
| 774 |
+
"rewards/accuracies": 0.71875,
|
| 775 |
+
"rewards/chosen": -1.879359245300293,
|
| 776 |
+
"rewards/margins": 0.8195317983627319,
|
| 777 |
+
"rewards/rejected": -2.6988909244537354,
|
| 778 |
+
"step": 460
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 0.49201779638838,
|
| 782 |
+
"grad_norm": 5.610862161934833,
|
| 783 |
+
"learning_rate": 3.0039685369660785e-06,
|
| 784 |
+
"logits/chosen": -1.3113436698913574,
|
| 785 |
+
"logits/rejected": -1.173460602760315,
|
| 786 |
+
"logps/chosen": -466.1058044433594,
|
| 787 |
+
"logps/rejected": -511.7659606933594,
|
| 788 |
+
"loss": 0.4856,
|
| 789 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 790 |
+
"rewards/chosen": -2.0702614784240723,
|
| 791 |
+
"rewards/margins": 0.7728853821754456,
|
| 792 |
+
"rewards/rejected": -2.843147039413452,
|
| 793 |
+
"step": 470
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.5024862601413242,
|
| 797 |
+
"grad_norm": 6.26181575439223,
|
| 798 |
+
"learning_rate": 2.91409685362137e-06,
|
| 799 |
+
"logits/chosen": -1.2794568538665771,
|
| 800 |
+
"logits/rejected": -1.123780608177185,
|
| 801 |
+
"logps/chosen": -523.4374389648438,
|
| 802 |
+
"logps/rejected": -596.9427490234375,
|
| 803 |
+
"loss": 0.5046,
|
| 804 |
+
"rewards/accuracies": 0.762499988079071,
|
| 805 |
+
"rewards/chosen": -2.514650583267212,
|
| 806 |
+
"rewards/margins": 1.0325826406478882,
|
| 807 |
+
"rewards/rejected": -3.5472328662872314,
|
| 808 |
+
"step": 480
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.5129547238942685,
|
| 812 |
+
"grad_norm": 4.885030756434318,
|
| 813 |
+
"learning_rate": 2.8236713524386085e-06,
|
| 814 |
+
"logits/chosen": -1.2807716131210327,
|
| 815 |
+
"logits/rejected": -1.107301950454712,
|
| 816 |
+
"logps/chosen": -547.37890625,
|
| 817 |
+
"logps/rejected": -605.0728149414062,
|
| 818 |
+
"loss": 0.4989,
|
| 819 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 820 |
+
"rewards/chosen": -2.634721040725708,
|
| 821 |
+
"rewards/margins": 0.92457515001297,
|
| 822 |
+
"rewards/rejected": -3.5592963695526123,
|
| 823 |
+
"step": 490
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.5234231876472127,
|
| 827 |
+
"grad_norm": 4.812586396006598,
|
| 828 |
+
"learning_rate": 2.7328129695107205e-06,
|
| 829 |
+
"logits/chosen": -1.2112505435943604,
|
| 830 |
+
"logits/rejected": -1.1074721813201904,
|
| 831 |
+
"logps/chosen": -523.0760498046875,
|
| 832 |
+
"logps/rejected": -595.048095703125,
|
| 833 |
+
"loss": 0.5046,
|
| 834 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 835 |
+
"rewards/chosen": -2.574467420578003,
|
| 836 |
+
"rewards/margins": 0.9888777732849121,
|
| 837 |
+
"rewards/rejected": -3.563344955444336,
|
| 838 |
+
"step": 500
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.5234231876472127,
|
| 842 |
+
"eval_logits/chosen": -1.1373296976089478,
|
| 843 |
+
"eval_logits/rejected": -1.0301768779754639,
|
| 844 |
+
"eval_logps/chosen": -529.9542236328125,
|
| 845 |
+
"eval_logps/rejected": -605.2976684570312,
|
| 846 |
+
"eval_loss": 0.5062148571014404,
|
| 847 |
+
"eval_rewards/accuracies": 0.7579365372657776,
|
| 848 |
+
"eval_rewards/chosen": -2.650512456893921,
|
| 849 |
+
"eval_rewards/margins": 0.9558730721473694,
|
| 850 |
+
"eval_rewards/rejected": -3.6063857078552246,
|
| 851 |
+
"eval_runtime": 494.8767,
|
| 852 |
+
"eval_samples_per_second": 4.041,
|
| 853 |
+
"eval_steps_per_second": 0.127,
|
| 854 |
+
"step": 500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.533891651400157,
|
| 858 |
+
"grad_norm": 4.595490575025387,
|
| 859 |
+
"learning_rate": 2.641643219871597e-06,
|
| 860 |
+
"logits/chosen": -1.2453548908233643,
|
| 861 |
+
"logits/rejected": -1.1635602712631226,
|
| 862 |
+
"logps/chosen": -500.3102111816406,
|
| 863 |
+
"logps/rejected": -559.625244140625,
|
| 864 |
+
"loss": 0.4867,
|
| 865 |
+
"rewards/accuracies": 0.737500011920929,
|
| 866 |
+
"rewards/chosen": -2.409621000289917,
|
| 867 |
+
"rewards/margins": 0.8319376707077026,
|
| 868 |
+
"rewards/rejected": -3.24155855178833,
|
| 869 |
+
"step": 510
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 0.5443601151531012,
|
| 873 |
+
"grad_norm": 7.198268767989711,
|
| 874 |
+
"learning_rate": 2.5502840349805074e-06,
|
| 875 |
+
"logits/chosen": -1.2010407447814941,
|
| 876 |
+
"logits/rejected": -1.0801866054534912,
|
| 877 |
+
"logps/chosen": -578.692138671875,
|
| 878 |
+
"logps/rejected": -627.0946044921875,
|
| 879 |
+
"loss": 0.5249,
|
| 880 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 881 |
+
"rewards/chosen": -3.033255100250244,
|
| 882 |
+
"rewards/margins": 0.9397614598274231,
|
| 883 |
+
"rewards/rejected": -3.9730167388916016,
|
| 884 |
+
"step": 520
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.5548285789060455,
|
| 888 |
+
"grad_norm": 6.519693692708999,
|
| 889 |
+
"learning_rate": 2.4588575996495797e-06,
|
| 890 |
+
"logits/chosen": -1.169585943222046,
|
| 891 |
+
"logits/rejected": -1.0577750205993652,
|
| 892 |
+
"logps/chosen": -533.5675048828125,
|
| 893 |
+
"logps/rejected": -615.3276977539062,
|
| 894 |
+
"loss": 0.4947,
|
| 895 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 896 |
+
"rewards/chosen": -2.606905460357666,
|
| 897 |
+
"rewards/margins": 1.0898813009262085,
|
| 898 |
+
"rewards/rejected": -3.696786880493164,
|
| 899 |
+
"step": 530
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 0.5652970426589898,
|
| 903 |
+
"grad_norm": 4.681072379969835,
|
| 904 |
+
"learning_rate": 2.367486188632446e-06,
|
| 905 |
+
"logits/chosen": -1.2149909734725952,
|
| 906 |
+
"logits/rejected": -1.0861246585845947,
|
| 907 |
+
"logps/chosen": -495.14093017578125,
|
| 908 |
+
"logps/rejected": -530.830810546875,
|
| 909 |
+
"loss": 0.5288,
|
| 910 |
+
"rewards/accuracies": 0.706250011920929,
|
| 911 |
+
"rewards/chosen": -2.406106472015381,
|
| 912 |
+
"rewards/margins": 0.7864419221878052,
|
| 913 |
+
"rewards/rejected": -3.1925485134124756,
|
| 914 |
+
"step": 540
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.575765506411934,
|
| 918 |
+
"grad_norm": 4.605359920599186,
|
| 919 |
+
"learning_rate": 2.276292003092593e-06,
|
| 920 |
+
"logits/chosen": -1.25673508644104,
|
| 921 |
+
"logits/rejected": -1.0914690494537354,
|
| 922 |
+
"logps/chosen": -509.0771484375,
|
| 923 |
+
"logps/rejected": -543.2958374023438,
|
| 924 |
+
"loss": 0.512,
|
| 925 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 926 |
+
"rewards/chosen": -2.4712395668029785,
|
| 927 |
+
"rewards/margins": 0.7063032388687134,
|
| 928 |
+
"rewards/rejected": -3.1775424480438232,
|
| 929 |
+
"step": 550
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.5862339701648783,
|
| 933 |
+
"grad_norm": 6.363842169113803,
|
| 934 |
+
"learning_rate": 2.1853970071701415e-06,
|
| 935 |
+
"logits/chosen": -1.153988242149353,
|
| 936 |
+
"logits/rejected": -1.0725597143173218,
|
| 937 |
+
"logps/chosen": -513.7321166992188,
|
| 938 |
+
"logps/rejected": -567.1982421875,
|
| 939 |
+
"loss": 0.5248,
|
| 940 |
+
"rewards/accuracies": 0.768750011920929,
|
| 941 |
+
"rewards/chosen": -2.496264934539795,
|
| 942 |
+
"rewards/margins": 0.8739057779312134,
|
| 943 |
+
"rewards/rejected": -3.370171070098877,
|
| 944 |
+
"step": 560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.5967024339178225,
|
| 948 |
+
"grad_norm": 6.958473109368914,
|
| 949 |
+
"learning_rate": 2.0949227648656194e-06,
|
| 950 |
+
"logits/chosen": -1.243544578552246,
|
| 951 |
+
"logits/rejected": -1.138840913772583,
|
| 952 |
+
"logps/chosen": -514.3547973632812,
|
| 953 |
+
"logps/rejected": -589.1139526367188,
|
| 954 |
+
"loss": 0.5081,
|
| 955 |
+
"rewards/accuracies": 0.737500011920929,
|
| 956 |
+
"rewards/chosen": -2.536358594894409,
|
| 957 |
+
"rewards/margins": 0.8745514154434204,
|
| 958 |
+
"rewards/rejected": -3.410910129547119,
|
| 959 |
+
"step": 570
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 0.6071708976707668,
|
| 963 |
+
"grad_norm": 5.30154815627735,
|
| 964 |
+
"learning_rate": 2.00499027745888e-06,
|
| 965 |
+
"logits/chosen": -1.1577163934707642,
|
| 966 |
+
"logits/rejected": -1.0448474884033203,
|
| 967 |
+
"logps/chosen": -500.58599853515625,
|
| 968 |
+
"logps/rejected": -561.0791015625,
|
| 969 |
+
"loss": 0.5112,
|
| 970 |
+
"rewards/accuracies": 0.706250011920929,
|
| 971 |
+
"rewards/chosen": -2.6586689949035645,
|
| 972 |
+
"rewards/margins": 0.8494758605957031,
|
| 973 |
+
"rewards/rejected": -3.5081450939178467,
|
| 974 |
+
"step": 580
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.6176393614237111,
|
| 978 |
+
"grad_norm": 5.957700119060022,
|
| 979 |
+
"learning_rate": 1.915719821680624e-06,
|
| 980 |
+
"logits/chosen": -1.342710256576538,
|
| 981 |
+
"logits/rejected": -1.2864643335342407,
|
| 982 |
+
"logps/chosen": -480.73272705078125,
|
| 983 |
+
"logps/rejected": -578.4927978515625,
|
| 984 |
+
"loss": 0.4987,
|
| 985 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 986 |
+
"rewards/chosen": -2.3478407859802246,
|
| 987 |
+
"rewards/margins": 0.940704345703125,
|
| 988 |
+
"rewards/rejected": -3.2885451316833496,
|
| 989 |
+
"step": 590
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.6281078251766553,
|
| 993 |
+
"grad_norm": 5.805746259122187,
|
| 994 |
+
"learning_rate": 1.8272307888529276e-06,
|
| 995 |
+
"logits/chosen": -1.1955945491790771,
|
| 996 |
+
"logits/rejected": -1.0686155557632446,
|
| 997 |
+
"logps/chosen": -551.1502685546875,
|
| 998 |
+
"logps/rejected": -601.6345825195312,
|
| 999 |
+
"loss": 0.4736,
|
| 1000 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1001 |
+
"rewards/chosen": -2.542466640472412,
|
| 1002 |
+
"rewards/margins": 0.8464245796203613,
|
| 1003 |
+
"rewards/rejected": -3.3888916969299316,
|
| 1004 |
+
"step": 600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.6281078251766553,
|
| 1008 |
+
"eval_logits/chosen": -1.1252615451812744,
|
| 1009 |
+
"eval_logits/rejected": -1.0135337114334106,
|
| 1010 |
+
"eval_logps/chosen": -537.3406372070312,
|
| 1011 |
+
"eval_logps/rejected": -621.1549072265625,
|
| 1012 |
+
"eval_loss": 0.5059433579444885,
|
| 1013 |
+
"eval_rewards/accuracies": 0.7638888955116272,
|
| 1014 |
+
"eval_rewards/chosen": -2.724376678466797,
|
| 1015 |
+
"eval_rewards/margins": 1.0405809879302979,
|
| 1016 |
+
"eval_rewards/rejected": -3.7649576663970947,
|
| 1017 |
+
"eval_runtime": 496.2742,
|
| 1018 |
+
"eval_samples_per_second": 4.03,
|
| 1019 |
+
"eval_steps_per_second": 0.127,
|
| 1020 |
+
"step": 600
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 0.6385762889295996,
|
| 1024 |
+
"grad_norm": 5.793634680662145,
|
| 1025 |
+
"learning_rate": 1.739641525213929e-06,
|
| 1026 |
+
"logits/chosen": -1.216796636581421,
|
| 1027 |
+
"logits/rejected": -1.1526730060577393,
|
| 1028 |
+
"logps/chosen": -538.6376953125,
|
| 1029 |
+
"logps/rejected": -615.1124877929688,
|
| 1030 |
+
"loss": 0.5134,
|
| 1031 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1032 |
+
"rewards/chosen": -2.6652350425720215,
|
| 1033 |
+
"rewards/margins": 1.1263973712921143,
|
| 1034 |
+
"rewards/rejected": -3.7916324138641357,
|
| 1035 |
+
"step": 610
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.6490447526825438,
|
| 1039 |
+
"grad_norm": 5.8551862397886,
|
| 1040 |
+
"learning_rate": 1.6530691736402317e-06,
|
| 1041 |
+
"logits/chosen": -1.3594751358032227,
|
| 1042 |
+
"logits/rejected": -1.1856592893600464,
|
| 1043 |
+
"logps/chosen": -541.5281372070312,
|
| 1044 |
+
"logps/rejected": -591.7969970703125,
|
| 1045 |
+
"loss": 0.4667,
|
| 1046 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1047 |
+
"rewards/chosen": -2.4107346534729004,
|
| 1048 |
+
"rewards/margins": 1.0584288835525513,
|
| 1049 |
+
"rewards/rejected": -3.4691638946533203,
|
| 1050 |
+
"step": 620
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 0.6595132164354881,
|
| 1054 |
+
"grad_norm": 7.40407247870995,
|
| 1055 |
+
"learning_rate": 1.5676295169786864e-06,
|
| 1056 |
+
"logits/chosen": -1.2886357307434082,
|
| 1057 |
+
"logits/rejected": -1.160259485244751,
|
| 1058 |
+
"logps/chosen": -528.3416137695312,
|
| 1059 |
+
"logps/rejected": -582.4341430664062,
|
| 1060 |
+
"loss": 0.4807,
|
| 1061 |
+
"rewards/accuracies": 0.6875,
|
| 1062 |
+
"rewards/chosen": -2.508863925933838,
|
| 1063 |
+
"rewards/margins": 0.9682027697563171,
|
| 1064 |
+
"rewards/rejected": -3.477067232131958,
|
| 1065 |
+
"step": 630
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 0.6699816801884323,
|
| 1069 |
+
"grad_norm": 6.664520102647163,
|
| 1070 |
+
"learning_rate": 1.4834368231970922e-06,
|
| 1071 |
+
"logits/chosen": -1.1566966772079468,
|
| 1072 |
+
"logits/rejected": -1.0316822528839111,
|
| 1073 |
+
"logps/chosen": -519.1185913085938,
|
| 1074 |
+
"logps/rejected": -613.6243896484375,
|
| 1075 |
+
"loss": 0.4678,
|
| 1076 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 1077 |
+
"rewards/chosen": -2.8396458625793457,
|
| 1078 |
+
"rewards/margins": 1.1082502603530884,
|
| 1079 |
+
"rewards/rejected": -3.9478962421417236,
|
| 1080 |
+
"step": 640
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.6804501439413766,
|
| 1084 |
+
"grad_norm": 6.202576057234435,
|
| 1085 |
+
"learning_rate": 1.4006036925609245e-06,
|
| 1086 |
+
"logits/chosen": -1.2698938846588135,
|
| 1087 |
+
"logits/rejected": -1.2037663459777832,
|
| 1088 |
+
"logps/chosen": -512.4635009765625,
|
| 1089 |
+
"logps/rejected": -599.8485107421875,
|
| 1090 |
+
"loss": 0.5124,
|
| 1091 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1092 |
+
"rewards/chosen": -2.5361390113830566,
|
| 1093 |
+
"rewards/margins": 0.8807609677314758,
|
| 1094 |
+
"rewards/rejected": -3.416900157928467,
|
| 1095 |
+
"step": 650
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.6909186076943209,
|
| 1099 |
+
"grad_norm": 5.411195685464347,
|
| 1100 |
+
"learning_rate": 1.3192409070404582e-06,
|
| 1101 |
+
"logits/chosen": -1.33925461769104,
|
| 1102 |
+
"logits/rejected": -1.2339675426483154,
|
| 1103 |
+
"logps/chosen": -507.4242248535156,
|
| 1104 |
+
"logps/rejected": -586.1668701171875,
|
| 1105 |
+
"loss": 0.4971,
|
| 1106 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1107 |
+
"rewards/chosen": -2.554523468017578,
|
| 1108 |
+
"rewards/margins": 0.9770146608352661,
|
| 1109 |
+
"rewards/rejected": -3.5315380096435547,
|
| 1110 |
+
"step": 660
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 0.7013870714472651,
|
| 1114 |
+
"grad_norm": 7.306078973174853,
|
| 1115 |
+
"learning_rate": 1.2394572821496953e-06,
|
| 1116 |
+
"logits/chosen": -1.1949710845947266,
|
| 1117 |
+
"logits/rejected": -1.0358936786651611,
|
| 1118 |
+
"logps/chosen": -556.1329956054688,
|
| 1119 |
+
"logps/rejected": -644.8235473632812,
|
| 1120 |
+
"loss": 0.4783,
|
| 1121 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1122 |
+
"rewards/chosen": -3.012876510620117,
|
| 1123 |
+
"rewards/margins": 1.1485122442245483,
|
| 1124 |
+
"rewards/rejected": -4.161388397216797,
|
| 1125 |
+
"step": 670
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.7118555352002094,
|
| 1129 |
+
"grad_norm": 6.569348565798864,
|
| 1130 |
+
"learning_rate": 1.1613595214152713e-06,
|
| 1131 |
+
"logits/chosen": -1.1749083995819092,
|
| 1132 |
+
"logits/rejected": -1.0992681980133057,
|
| 1133 |
+
"logps/chosen": -572.3611450195312,
|
| 1134 |
+
"logps/rejected": -654.3881225585938,
|
| 1135 |
+
"loss": 0.5193,
|
| 1136 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1137 |
+
"rewards/chosen": -3.059669017791748,
|
| 1138 |
+
"rewards/margins": 0.9747347831726074,
|
| 1139 |
+
"rewards/rejected": -4.0344038009643555,
|
| 1140 |
+
"step": 680
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 0.7223239989531536,
|
| 1144 |
+
"grad_norm": 4.877247655613693,
|
| 1145 |
+
"learning_rate": 1.0850520736699362e-06,
|
| 1146 |
+
"logits/chosen": -1.3139179944992065,
|
| 1147 |
+
"logits/rejected": -1.1570199728012085,
|
| 1148 |
+
"logps/chosen": -550.0722045898438,
|
| 1149 |
+
"logps/rejected": -594.5676879882812,
|
| 1150 |
+
"loss": 0.4496,
|
| 1151 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1152 |
+
"rewards/chosen": -2.778721570968628,
|
| 1153 |
+
"rewards/margins": 0.9407702684402466,
|
| 1154 |
+
"rewards/rejected": -3.719491958618164,
|
| 1155 |
+
"step": 690
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.7327924627060979,
|
| 1159 |
+
"grad_norm": 5.850084839342618,
|
| 1160 |
+
"learning_rate": 1.0106369933615043e-06,
|
| 1161 |
+
"logits/chosen": -1.1720659732818604,
|
| 1162 |
+
"logits/rejected": -1.1006288528442383,
|
| 1163 |
+
"logps/chosen": -543.8446044921875,
|
| 1164 |
+
"logps/rejected": -637.2833862304688,
|
| 1165 |
+
"loss": 0.4619,
|
| 1166 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1167 |
+
"rewards/chosen": -2.941668748855591,
|
| 1168 |
+
"rewards/margins": 1.034071922302246,
|
| 1169 |
+
"rewards/rejected": -3.975741147994995,
|
| 1170 |
+
"step": 700
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 0.7327924627060979,
|
| 1174 |
+
"eval_logits/chosen": -1.119376301765442,
|
| 1175 |
+
"eval_logits/rejected": -1.006381869316101,
|
| 1176 |
+
"eval_logps/chosen": -557.3041381835938,
|
| 1177 |
+
"eval_logps/rejected": -644.5651245117188,
|
| 1178 |
+
"eval_loss": 0.4993818998336792,
|
| 1179 |
+
"eval_rewards/accuracies": 0.761904776096344,
|
| 1180 |
+
"eval_rewards/chosen": -2.9240119457244873,
|
| 1181 |
+
"eval_rewards/margins": 1.0750477313995361,
|
| 1182 |
+
"eval_rewards/rejected": -3.9990594387054443,
|
| 1183 |
+
"eval_runtime": 494.8985,
|
| 1184 |
+
"eval_samples_per_second": 4.041,
|
| 1185 |
+
"eval_steps_per_second": 0.127,
|
| 1186 |
+
"step": 700
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.7432609264590422,
|
| 1190 |
+
"grad_norm": 5.940990677339228,
|
| 1191 |
+
"learning_rate": 9.382138040640714e-07,
|
| 1192 |
+
"logits/chosen": -1.175462245941162,
|
| 1193 |
+
"logits/rejected": -1.0119060277938843,
|
| 1194 |
+
"logps/chosen": -566.7076416015625,
|
| 1195 |
+
"logps/rejected": -634.9506225585938,
|
| 1196 |
+
"loss": 0.5486,
|
| 1197 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1198 |
+
"rewards/chosen": -3.01143217086792,
|
| 1199 |
+
"rewards/margins": 0.9881938695907593,
|
| 1200 |
+
"rewards/rejected": -3.9996261596679688,
|
| 1201 |
+
"step": 710
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 0.7537293902119864,
|
| 1205 |
+
"grad_norm": 5.5111666858195125,
|
| 1206 |
+
"learning_rate": 8.678793653740633e-07,
|
| 1207 |
+
"logits/chosen": -1.3316354751586914,
|
| 1208 |
+
"logits/rejected": -1.171438455581665,
|
| 1209 |
+
"logps/chosen": -583.8369140625,
|
| 1210 |
+
"logps/rejected": -640.4646606445312,
|
| 1211 |
+
"loss": 0.4747,
|
| 1212 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1213 |
+
"rewards/chosen": -2.6381282806396484,
|
| 1214 |
+
"rewards/margins": 1.0775749683380127,
|
| 1215 |
+
"rewards/rejected": -3.715702772140503,
|
| 1216 |
+
"step": 720
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 0.7641978539649307,
|
| 1220 |
+
"grad_norm": 5.993566017988564,
|
| 1221 |
+
"learning_rate": 7.997277433690984e-07,
|
| 1222 |
+
"logits/chosen": -1.208320140838623,
|
| 1223 |
+
"logits/rejected": -1.071702480316162,
|
| 1224 |
+
"logps/chosen": -516.3609619140625,
|
| 1225 |
+
"logps/rejected": -598.6437377929688,
|
| 1226 |
+
"loss": 0.5075,
|
| 1227 |
+
"rewards/accuracies": 0.75,
|
| 1228 |
+
"rewards/chosen": -2.7794041633605957,
|
| 1229 |
+
"rewards/margins": 0.9944884181022644,
|
| 1230 |
+
"rewards/rejected": -3.773892879486084,
|
| 1231 |
+
"step": 730
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.7746663177178749,
|
| 1235 |
+
"grad_norm": 5.60538349675765,
|
| 1236 |
+
"learning_rate": 7.338500848029603e-07,
|
| 1237 |
+
"logits/chosen": -1.2300490140914917,
|
| 1238 |
+
"logits/rejected": -1.0854498147964478,
|
| 1239 |
+
"logps/chosen": -557.65234375,
|
| 1240 |
+
"logps/rejected": -595.0419921875,
|
| 1241 |
+
"loss": 0.5081,
|
| 1242 |
+
"rewards/accuracies": 0.71875,
|
| 1243 |
+
"rewards/chosen": -2.7690441608428955,
|
| 1244 |
+
"rewards/margins": 0.9237399101257324,
|
| 1245 |
+
"rewards/rejected": -3.692783832550049,
|
| 1246 |
+
"step": 740
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 0.7851347814708192,
|
| 1250 |
+
"grad_norm": 5.061641492930568,
|
| 1251 |
+
"learning_rate": 6.70334495204884e-07,
|
| 1252 |
+
"logits/chosen": -1.0873275995254517,
|
| 1253 |
+
"logits/rejected": -1.0135056972503662,
|
| 1254 |
+
"logps/chosen": -510.3736877441406,
|
| 1255 |
+
"logps/rejected": -621.0284423828125,
|
| 1256 |
+
"loss": 0.4814,
|
| 1257 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1258 |
+
"rewards/chosen": -2.822464942932129,
|
| 1259 |
+
"rewards/margins": 1.0254597663879395,
|
| 1260 |
+
"rewards/rejected": -3.8479247093200684,
|
| 1261 |
+
"step": 750
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 0.7956032452237635,
|
| 1265 |
+
"grad_norm": 4.889175666405806,
|
| 1266 |
+
"learning_rate": 6.092659210462232e-07,
|
| 1267 |
+
"logits/chosen": -1.2110772132873535,
|
| 1268 |
+
"logits/rejected": -1.0558052062988281,
|
| 1269 |
+
"logps/chosen": -540.3411254882812,
|
| 1270 |
+
"logps/rejected": -596.0307006835938,
|
| 1271 |
+
"loss": 0.5076,
|
| 1272 |
+
"rewards/accuracies": 0.75,
|
| 1273 |
+
"rewards/chosen": -2.8302786350250244,
|
| 1274 |
+
"rewards/margins": 0.9066953659057617,
|
| 1275 |
+
"rewards/rejected": -3.736973524093628,
|
| 1276 |
+
"step": 760
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.8060717089767077,
|
| 1280 |
+
"grad_norm": 4.4612176619325075,
|
| 1281 |
+
"learning_rate": 5.507260361320738e-07,
|
| 1282 |
+
"logits/chosen": -1.2954254150390625,
|
| 1283 |
+
"logits/rejected": -1.2630584239959717,
|
| 1284 |
+
"logps/chosen": -560.3121337890625,
|
| 1285 |
+
"logps/rejected": -663.4389038085938,
|
| 1286 |
+
"loss": 0.4681,
|
| 1287 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1288 |
+
"rewards/chosen": -2.743813991546631,
|
| 1289 |
+
"rewards/margins": 1.0250587463378906,
|
| 1290 |
+
"rewards/rejected": -3.7688724994659424,
|
| 1291 |
+
"step": 770
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.816540172729652,
|
| 1295 |
+
"grad_norm": 5.524604671684533,
|
| 1296 |
+
"learning_rate": 4.947931323697983e-07,
|
| 1297 |
+
"logits/chosen": -1.3126929998397827,
|
| 1298 |
+
"logits/rejected": -1.088275671005249,
|
| 1299 |
+
"logps/chosen": -570.6515502929688,
|
| 1300 |
+
"logps/rejected": -593.5574951171875,
|
| 1301 |
+
"loss": 0.4959,
|
| 1302 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 1303 |
+
"rewards/chosen": -2.6409595012664795,
|
| 1304 |
+
"rewards/margins": 0.848545253276825,
|
| 1305 |
+
"rewards/rejected": -3.48950457572937,
|
| 1306 |
+
"step": 780
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 0.8270086364825961,
|
| 1310 |
+
"grad_norm": 5.9693834124907585,
|
| 1311 |
+
"learning_rate": 4.4154201506053985e-07,
|
| 1312 |
+
"logits/chosen": -1.1836011409759521,
|
| 1313 |
+
"logits/rejected": -1.0795371532440186,
|
| 1314 |
+
"logps/chosen": -529.63134765625,
|
| 1315 |
+
"logps/rejected": -619.8529663085938,
|
| 1316 |
+
"loss": 0.523,
|
| 1317 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1318 |
+
"rewards/chosen": -2.8782217502593994,
|
| 1319 |
+
"rewards/margins": 0.950698971748352,
|
| 1320 |
+
"rewards/rejected": -3.828920841217041,
|
| 1321 |
+
"step": 790
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 0.8374771002355405,
|
| 1325 |
+
"grad_norm": 6.22817743126936,
|
| 1326 |
+
"learning_rate": 3.910439028537638e-07,
|
| 1327 |
+
"logits/chosen": -1.1832584142684937,
|
| 1328 |
+
"logits/rejected": -1.1257516145706177,
|
| 1329 |
+
"logps/chosen": -505.93316650390625,
|
| 1330 |
+
"logps/rejected": -627.0811157226562,
|
| 1331 |
+
"loss": 0.4926,
|
| 1332 |
+
"rewards/accuracies": 0.78125,
|
| 1333 |
+
"rewards/chosen": -2.7608401775360107,
|
| 1334 |
+
"rewards/margins": 1.1763994693756104,
|
| 1335 |
+
"rewards/rejected": -3.937239408493042,
|
| 1336 |
+
"step": 800
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 0.8374771002355405,
|
| 1340 |
+
"eval_logits/chosen": -1.1640794277191162,
|
| 1341 |
+
"eval_logits/rejected": -1.0516477823257446,
|
| 1342 |
+
"eval_logps/chosen": -537.3770141601562,
|
| 1343 |
+
"eval_logps/rejected": -619.2051391601562,
|
| 1344 |
+
"eval_loss": 0.49621155858039856,
|
| 1345 |
+
"eval_rewards/accuracies": 0.7658730149269104,
|
| 1346 |
+
"eval_rewards/chosen": -2.724740743637085,
|
| 1347 |
+
"eval_rewards/margins": 1.0207195281982422,
|
| 1348 |
+
"eval_rewards/rejected": -3.745460033416748,
|
| 1349 |
+
"eval_runtime": 494.3028,
|
| 1350 |
+
"eval_samples_per_second": 4.046,
|
| 1351 |
+
"eval_steps_per_second": 0.127,
|
| 1352 |
+
"step": 800
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 0.8479455639884846,
|
| 1356 |
+
"grad_norm": 4.972938912679825,
|
| 1357 |
+
"learning_rate": 3.4336633249862084e-07,
|
| 1358 |
+
"logits/chosen": -1.2425668239593506,
|
| 1359 |
+
"logits/rejected": -1.0636166334152222,
|
| 1360 |
+
"logps/chosen": -561.7000732421875,
|
| 1361 |
+
"logps/rejected": -619.0046997070312,
|
| 1362 |
+
"loss": 0.508,
|
| 1363 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1364 |
+
"rewards/chosen": -2.7175962924957275,
|
| 1365 |
+
"rewards/margins": 0.9634302854537964,
|
| 1366 |
+
"rewards/rejected": -3.6810269355773926,
|
| 1367 |
+
"step": 810
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.8584140277414289,
|
| 1371 |
+
"grad_norm": 5.09720991480253,
|
| 1372 |
+
"learning_rate": 2.98573068519539e-07,
|
| 1373 |
+
"logits/chosen": -1.2827835083007812,
|
| 1374 |
+
"logits/rejected": -1.2060225009918213,
|
| 1375 |
+
"logps/chosen": -539.7214965820312,
|
| 1376 |
+
"logps/rejected": -615.7735595703125,
|
| 1377 |
+
"loss": 0.4994,
|
| 1378 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1379 |
+
"rewards/chosen": -2.778557538986206,
|
| 1380 |
+
"rewards/margins": 0.8738244771957397,
|
| 1381 |
+
"rewards/rejected": -3.6523823738098145,
|
| 1382 |
+
"step": 820
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.8688824914943732,
|
| 1386 |
+
"grad_norm": 5.782887668669889,
|
| 1387 |
+
"learning_rate": 2.5672401793681854e-07,
|
| 1388 |
+
"logits/chosen": -1.2188454866409302,
|
| 1389 |
+
"logits/rejected": -1.1675770282745361,
|
| 1390 |
+
"logps/chosen": -514.8630981445312,
|
| 1391 |
+
"logps/rejected": -599.4465942382812,
|
| 1392 |
+
"loss": 0.5131,
|
| 1393 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1394 |
+
"rewards/chosen": -2.682121753692627,
|
| 1395 |
+
"rewards/margins": 0.837975025177002,
|
| 1396 |
+
"rewards/rejected": -3.520097017288208,
|
| 1397 |
+
"step": 830
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 0.8793509552473174,
|
| 1401 |
+
"grad_norm": 6.123094674224966,
|
| 1402 |
+
"learning_rate": 2.178751501463036e-07,
|
| 1403 |
+
"logits/chosen": -1.2655466794967651,
|
| 1404 |
+
"logits/rejected": -1.162320613861084,
|
| 1405 |
+
"logps/chosen": -548.4150390625,
|
| 1406 |
+
"logps/rejected": -654.3787231445312,
|
| 1407 |
+
"loss": 0.4696,
|
| 1408 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1409 |
+
"rewards/chosen": -2.7333908081054688,
|
| 1410 |
+
"rewards/margins": 1.0990099906921387,
|
| 1411 |
+
"rewards/rejected": -3.8324007987976074,
|
| 1412 |
+
"step": 840
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 0.8898194190002617,
|
| 1416 |
+
"grad_norm": 5.66840809357227,
|
| 1417 |
+
"learning_rate": 1.820784220652766e-07,
|
| 1418 |
+
"logits/chosen": -1.1958177089691162,
|
| 1419 |
+
"logits/rejected": -1.0371004343032837,
|
| 1420 |
+
"logps/chosen": -533.7821044921875,
|
| 1421 |
+
"logps/rejected": -595.7824096679688,
|
| 1422 |
+
"loss": 0.4643,
|
| 1423 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1424 |
+
"rewards/chosen": -2.6792378425598145,
|
| 1425 |
+
"rewards/margins": 0.9526662826538086,
|
| 1426 |
+
"rewards/rejected": -3.631904125213623,
|
| 1427 |
+
"step": 850
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.9002878827532059,
|
| 1431 |
+
"grad_norm": 4.99425900843833,
|
| 1432 |
+
"learning_rate": 1.4938170864468636e-07,
|
| 1433 |
+
"logits/chosen": -1.257932424545288,
|
| 1434 |
+
"logits/rejected": -1.1045788526535034,
|
| 1435 |
+
"logps/chosen": -551.285888671875,
|
| 1436 |
+
"logps/rejected": -612.8128662109375,
|
| 1437 |
+
"loss": 0.4681,
|
| 1438 |
+
"rewards/accuracies": 0.75,
|
| 1439 |
+
"rewards/chosen": -2.804962396621704,
|
| 1440 |
+
"rewards/margins": 0.8677828907966614,
|
| 1441 |
+
"rewards/rejected": -3.6727447509765625,
|
| 1442 |
+
"step": 860
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 0.9107563465061502,
|
| 1446 |
+
"grad_norm": 4.991076961519154,
|
| 1447 |
+
"learning_rate": 1.1982873884064466e-07,
|
| 1448 |
+
"logits/chosen": -1.294597864151001,
|
| 1449 |
+
"logits/rejected": -1.1019765138626099,
|
| 1450 |
+
"logps/chosen": -572.5621337890625,
|
| 1451 |
+
"logps/rejected": -625.5676879882812,
|
| 1452 |
+
"loss": 0.5032,
|
| 1453 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 1454 |
+
"rewards/chosen": -2.914654493331909,
|
| 1455 |
+
"rewards/margins": 1.0726633071899414,
|
| 1456 |
+
"rewards/rejected": -3.9873173236846924,
|
| 1457 |
+
"step": 870
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 0.9212248102590945,
|
| 1461 |
+
"grad_norm": 5.0189704749647746,
|
| 1462 |
+
"learning_rate": 9.345903713082305e-08,
|
| 1463 |
+
"logits/chosen": -1.2915667295455933,
|
| 1464 |
+
"logits/rejected": -1.1525356769561768,
|
| 1465 |
+
"logps/chosen": -541.0573120117188,
|
| 1466 |
+
"logps/rejected": -600.8675537109375,
|
| 1467 |
+
"loss": 0.4752,
|
| 1468 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1469 |
+
"rewards/chosen": -2.7068932056427,
|
| 1470 |
+
"rewards/margins": 0.9946308135986328,
|
| 1471 |
+
"rewards/rejected": -3.701524019241333,
|
| 1472 |
+
"step": 880
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.9316932740120387,
|
| 1476 |
+
"grad_norm": 5.260823538009292,
|
| 1477 |
+
"learning_rate": 7.030787065396866e-08,
|
| 1478 |
+
"logits/chosen": -1.188293218612671,
|
| 1479 |
+
"logits/rejected": -1.0763866901397705,
|
| 1480 |
+
"logps/chosen": -527.5496215820312,
|
| 1481 |
+
"logps/rejected": -644.0205078125,
|
| 1482 |
+
"loss": 0.4845,
|
| 1483 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 1484 |
+
"rewards/chosen": -2.790327548980713,
|
| 1485 |
+
"rewards/margins": 1.2539947032928467,
|
| 1486 |
+
"rewards/rejected": -4.0443220138549805,
|
| 1487 |
+
"step": 890
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.942161737764983,
|
| 1491 |
+
"grad_norm": 6.861506991670971,
|
| 1492 |
+
"learning_rate": 5.0406202043228604e-08,
|
| 1493 |
+
"logits/chosen": -1.0917075872421265,
|
| 1494 |
+
"logits/rejected": -1.0418908596038818,
|
| 1495 |
+
"logps/chosen": -544.6124267578125,
|
| 1496 |
+
"logps/rejected": -697.5197143554688,
|
| 1497 |
+
"loss": 0.4856,
|
| 1498 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1499 |
+
"rewards/chosen": -2.783902406692505,
|
| 1500 |
+
"rewards/margins": 1.2628685235977173,
|
| 1501 |
+
"rewards/rejected": -4.0467705726623535,
|
| 1502 |
+
"step": 900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.942161737764983,
|
| 1506 |
+
"eval_logits/chosen": -1.1508536338806152,
|
| 1507 |
+
"eval_logits/rejected": -1.03853178024292,
|
| 1508 |
+
"eval_logps/chosen": -545.9743041992188,
|
| 1509 |
+
"eval_logps/rejected": -631.738525390625,
|
| 1510 |
+
"eval_loss": 0.49515971541404724,
|
| 1511 |
+
"eval_rewards/accuracies": 0.77182537317276,
|
| 1512 |
+
"eval_rewards/chosen": -2.810713768005371,
|
| 1513 |
+
"eval_rewards/margins": 1.0600804090499878,
|
| 1514 |
+
"eval_rewards/rejected": -3.8707938194274902,
|
| 1515 |
+
"eval_runtime": 496.7331,
|
| 1516 |
+
"eval_samples_per_second": 4.026,
|
| 1517 |
+
"eval_steps_per_second": 0.127,
|
| 1518 |
+
"step": 900
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 0.9526302015179272,
|
| 1522 |
+
"grad_norm": 7.396195290341192,
|
| 1523 |
+
"learning_rate": 3.378064801637687e-08,
|
| 1524 |
+
"logits/chosen": -1.2156776189804077,
|
| 1525 |
+
"logits/rejected": -1.0312784910202026,
|
| 1526 |
+
"logps/chosen": -519.9219360351562,
|
| 1527 |
+
"logps/rejected": -579.5977172851562,
|
| 1528 |
+
"loss": 0.4987,
|
| 1529 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1530 |
+
"rewards/chosen": -2.731100559234619,
|
| 1531 |
+
"rewards/margins": 1.0545246601104736,
|
| 1532 |
+
"rewards/rejected": -3.785625457763672,
|
| 1533 |
+
"step": 910
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 0.9630986652708715,
|
| 1537 |
+
"grad_norm": 5.785542711133717,
|
| 1538 |
+
"learning_rate": 2.0453443778310766e-08,
|
| 1539 |
+
"logits/chosen": -1.2227718830108643,
|
| 1540 |
+
"logits/rejected": -1.0613749027252197,
|
| 1541 |
+
"logps/chosen": -563.939453125,
|
| 1542 |
+
"logps/rejected": -609.8938598632812,
|
| 1543 |
+
"loss": 0.4994,
|
| 1544 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1545 |
+
"rewards/chosen": -2.846004009246826,
|
| 1546 |
+
"rewards/margins": 0.9199835062026978,
|
| 1547 |
+
"rewards/rejected": -3.7659878730773926,
|
| 1548 |
+
"step": 920
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 0.9735671290238157,
|
| 1552 |
+
"grad_norm": 5.845655612266485,
|
| 1553 |
+
"learning_rate": 1.0442413283435759e-08,
|
| 1554 |
+
"logits/chosen": -1.1646806001663208,
|
| 1555 |
+
"logits/rejected": -0.9581974744796753,
|
| 1556 |
+
"logps/chosen": -586.3766479492188,
|
| 1557 |
+
"logps/rejected": -622.99072265625,
|
| 1558 |
+
"loss": 0.5017,
|
| 1559 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1560 |
+
"rewards/chosen": -2.876279354095459,
|
| 1561 |
+
"rewards/margins": 0.9776785969734192,
|
| 1562 |
+
"rewards/rejected": -3.8539581298828125,
|
| 1563 |
+
"step": 930
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.98403559277676,
|
| 1567 |
+
"grad_norm": 6.202963993103904,
|
| 1568 |
+
"learning_rate": 3.760945397705828e-09,
|
| 1569 |
+
"logits/chosen": -1.1803127527236938,
|
| 1570 |
+
"logits/rejected": -1.1215112209320068,
|
| 1571 |
+
"logps/chosen": -542.6771850585938,
|
| 1572 |
+
"logps/rejected": -630.9061279296875,
|
| 1573 |
+
"loss": 0.4831,
|
| 1574 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 1575 |
+
"rewards/chosen": -2.7995152473449707,
|
| 1576 |
+
"rewards/margins": 0.8906386494636536,
|
| 1577 |
+
"rewards/rejected": -3.6901535987854004,
|
| 1578 |
+
"step": 940
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 0.9945040565297043,
|
| 1582 |
+
"grad_norm": 6.026230259479486,
|
| 1583 |
+
"learning_rate": 4.1797599220405605e-10,
|
| 1584 |
+
"logits/chosen": -1.2184993028640747,
|
| 1585 |
+
"logits/rejected": -1.041572093963623,
|
| 1586 |
+
"logps/chosen": -546.837890625,
|
| 1587 |
+
"logps/rejected": -604.7627563476562,
|
| 1588 |
+
"loss": 0.489,
|
| 1589 |
+
"rewards/accuracies": 0.706250011920929,
|
| 1590 |
+
"rewards/chosen": -2.8025598526000977,
|
| 1591 |
+
"rewards/margins": 0.9029384851455688,
|
| 1592 |
+
"rewards/rejected": -3.705498218536377,
|
| 1593 |
+
"step": 950
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.9997382884061764,
|
| 1597 |
+
"step": 955,
|
| 1598 |
+
"total_flos": 0.0,
|
| 1599 |
+
"train_loss": 0.2358125359600127,
|
| 1600 |
+
"train_runtime": 19082.7985,
|
| 1601 |
+
"train_samples_per_second": 3.204,
|
| 1602 |
+
"train_steps_per_second": 0.05
|
| 1603 |
+
}
|
| 1604 |
+
],
|
| 1605 |
+
"logging_steps": 10,
|
| 1606 |
+
"max_steps": 955,
|
| 1607 |
+
"num_input_tokens_seen": 0,
|
| 1608 |
+
"num_train_epochs": 1,
|
| 1609 |
+
"save_steps": 100,
|
| 1610 |
+
"stateful_callbacks": {
|
| 1611 |
+
"TrainerControl": {
|
| 1612 |
+
"args": {
|
| 1613 |
+
"should_epoch_stop": false,
|
| 1614 |
+
"should_evaluate": false,
|
| 1615 |
+
"should_log": false,
|
| 1616 |
+
"should_save": true,
|
| 1617 |
+
"should_training_stop": true
|
| 1618 |
+
},
|
| 1619 |
+
"attributes": {}
|
| 1620 |
+
}
|
| 1621 |
+
},
|
| 1622 |
+
"total_flos": 0.0,
|
| 1623 |
+
"train_batch_size": 4,
|
| 1624 |
+
"trial_name": null,
|
| 1625 |
+
"trial_params": null
|
| 1626 |
+
}
|