Model save
Browse files- README.md +62 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May19_10-53-15_gpu4-119-5/events.out.tfevents.1716080053.gpu4-119-5.3224749.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +1822 -0
README.md
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: DUAL-GPO/phi-2-sft-lora-ultrachat-merged
|
| 9 |
+
model-index:
|
| 10 |
+
- name: phi-2-irepo-chatml-v2-i0
|
| 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 |
+
# phi-2-irepo-chatml-v2-i0
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [DUAL-GPO/phi-2-sft-lora-ultrachat-merged](https://huggingface.co/DUAL-GPO/phi-2-sft-lora-ultrachat-merged) on the None dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-06
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 4
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 3
|
| 44 |
+
- gradient_accumulation_steps: 4
|
| 45 |
+
- total_train_batch_size: 48
|
| 46 |
+
- total_eval_batch_size: 12
|
| 47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
+
- lr_scheduler_type: cosine
|
| 49 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 50 |
+
- num_epochs: 1
|
| 51 |
+
|
| 52 |
+
### Training results
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
### Framework versions
|
| 57 |
+
|
| 58 |
+
- PEFT 0.7.1
|
| 59 |
+
- Transformers 4.36.2
|
| 60 |
+
- Pytorch 2.1.2+cu121
|
| 61 |
+
- Datasets 2.14.6
|
| 62 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 335579632
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9bd270a0f5181d3292d21b3abb63c05fe850e935b9f4c2fa42044d7300cf50cf
|
| 3 |
size 335579632
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.04122506890033327,
|
| 4 |
+
"train_runtime": 12850.1249,
|
| 5 |
+
"train_samples": 61135,
|
| 6 |
+
"train_samples_per_second": 4.758,
|
| 7 |
+
"train_steps_per_second": 0.099
|
| 8 |
+
}
|
runs/May19_10-53-15_gpu4-119-5/events.out.tfevents.1716080053.gpu4-119-5.3224749.0
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 sha256:4c92a548e13b7be355be7f6914b555cd7cedcad491ab4b2821851f11ffb57212
|
| 3 |
+
size 86266
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.04122506890033327,
|
| 4 |
+
"train_runtime": 12850.1249,
|
| 5 |
+
"train_samples": 61135,
|
| 6 |
+
"train_samples_per_second": 4.758,
|
| 7 |
+
"train_steps_per_second": 0.099
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1822 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9994111874386653,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1273,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 3.90625e-08,
|
| 14 |
+
"logits/chosen": 0.571427047252655,
|
| 15 |
+
"logits/rejected": 0.7314685583114624,
|
| 16 |
+
"logps/chosen": -269.3661193847656,
|
| 17 |
+
"logps/rejected": -211.47052001953125,
|
| 18 |
+
"loss": 0.0442,
|
| 19 |
+
"rewards/accuracies": 0.0,
|
| 20 |
+
"rewards/chosen": 0.0,
|
| 21 |
+
"rewards/margins": 0.0,
|
| 22 |
+
"rewards/rejected": 0.0,
|
| 23 |
+
"step": 1
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"learning_rate": 3.90625e-07,
|
| 28 |
+
"logits/chosen": 0.693711519241333,
|
| 29 |
+
"logits/rejected": 0.7037237882614136,
|
| 30 |
+
"logps/chosen": -279.3418884277344,
|
| 31 |
+
"logps/rejected": -237.24127197265625,
|
| 32 |
+
"loss": 0.0615,
|
| 33 |
+
"rewards/accuracies": 0.3888888955116272,
|
| 34 |
+
"rewards/chosen": 0.00013600506645161659,
|
| 35 |
+
"rewards/margins": 9.131695696851239e-05,
|
| 36 |
+
"rewards/rejected": 4.4688102207146585e-05,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.02,
|
| 41 |
+
"learning_rate": 7.8125e-07,
|
| 42 |
+
"logits/chosen": 0.6187659502029419,
|
| 43 |
+
"logits/rejected": 0.6341584324836731,
|
| 44 |
+
"logps/chosen": -269.4906921386719,
|
| 45 |
+
"logps/rejected": -235.9794464111328,
|
| 46 |
+
"loss": 0.0505,
|
| 47 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 48 |
+
"rewards/chosen": 0.00019002554472535849,
|
| 49 |
+
"rewards/margins": 0.0005181421292945743,
|
| 50 |
+
"rewards/rejected": -0.0003281166427768767,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"learning_rate": 1.1718750000000001e-06,
|
| 56 |
+
"logits/chosen": 0.7198699712753296,
|
| 57 |
+
"logits/rejected": 0.7689396739006042,
|
| 58 |
+
"logps/chosen": -262.8460998535156,
|
| 59 |
+
"logps/rejected": -238.15896606445312,
|
| 60 |
+
"loss": 0.0612,
|
| 61 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 62 |
+
"rewards/chosen": 0.0001914335007313639,
|
| 63 |
+
"rewards/margins": -9.696001507109031e-05,
|
| 64 |
+
"rewards/rejected": 0.0002883935230784118,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.03,
|
| 69 |
+
"learning_rate": 1.5625e-06,
|
| 70 |
+
"logits/chosen": 0.6236015558242798,
|
| 71 |
+
"logits/rejected": 0.663158655166626,
|
| 72 |
+
"logps/chosen": -271.2866516113281,
|
| 73 |
+
"logps/rejected": -256.9141540527344,
|
| 74 |
+
"loss": 0.0451,
|
| 75 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 76 |
+
"rewards/chosen": -0.0013797181891277432,
|
| 77 |
+
"rewards/margins": 0.0008465956198051572,
|
| 78 |
+
"rewards/rejected": -0.0022263135761022568,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.04,
|
| 83 |
+
"learning_rate": 1.953125e-06,
|
| 84 |
+
"logits/chosen": 0.6586641073226929,
|
| 85 |
+
"logits/rejected": 0.7124195098876953,
|
| 86 |
+
"logps/chosen": -277.61431884765625,
|
| 87 |
+
"logps/rejected": -245.189697265625,
|
| 88 |
+
"loss": 0.0533,
|
| 89 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 90 |
+
"rewards/chosen": -0.0009271189337596297,
|
| 91 |
+
"rewards/margins": 0.0012959478190168738,
|
| 92 |
+
"rewards/rejected": -0.0022230667527765036,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.05,
|
| 97 |
+
"learning_rate": 2.3437500000000002e-06,
|
| 98 |
+
"logits/chosen": 0.6583735346794128,
|
| 99 |
+
"logits/rejected": 0.7763499021530151,
|
| 100 |
+
"logps/chosen": -295.41766357421875,
|
| 101 |
+
"logps/rejected": -274.49810791015625,
|
| 102 |
+
"loss": 0.0564,
|
| 103 |
+
"rewards/accuracies": 0.53125,
|
| 104 |
+
"rewards/chosen": -0.003529118839651346,
|
| 105 |
+
"rewards/margins": 0.0007951221195980906,
|
| 106 |
+
"rewards/rejected": -0.004324241541326046,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.05,
|
| 111 |
+
"learning_rate": 2.7343750000000004e-06,
|
| 112 |
+
"logits/chosen": 0.7303035855293274,
|
| 113 |
+
"logits/rejected": 0.7642933130264282,
|
| 114 |
+
"logps/chosen": -274.9570617675781,
|
| 115 |
+
"logps/rejected": -242.8776092529297,
|
| 116 |
+
"loss": 0.0458,
|
| 117 |
+
"rewards/accuracies": 0.46875,
|
| 118 |
+
"rewards/chosen": -0.005638590082526207,
|
| 119 |
+
"rewards/margins": 0.0016975027974694967,
|
| 120 |
+
"rewards/rejected": -0.00733609264716506,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.06,
|
| 125 |
+
"learning_rate": 3.125e-06,
|
| 126 |
+
"logits/chosen": 0.6080405712127686,
|
| 127 |
+
"logits/rejected": 0.7384923696517944,
|
| 128 |
+
"logps/chosen": -301.12823486328125,
|
| 129 |
+
"logps/rejected": -243.5036163330078,
|
| 130 |
+
"loss": 0.0526,
|
| 131 |
+
"rewards/accuracies": 0.612500011920929,
|
| 132 |
+
"rewards/chosen": -0.009938476607203484,
|
| 133 |
+
"rewards/margins": 0.004849700257182121,
|
| 134 |
+
"rewards/rejected": -0.01478817779570818,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.07,
|
| 139 |
+
"learning_rate": 3.5156250000000003e-06,
|
| 140 |
+
"logits/chosen": 0.733268141746521,
|
| 141 |
+
"logits/rejected": 0.7006837129592896,
|
| 142 |
+
"logps/chosen": -306.02166748046875,
|
| 143 |
+
"logps/rejected": -309.63238525390625,
|
| 144 |
+
"loss": 0.046,
|
| 145 |
+
"rewards/accuracies": 0.5625,
|
| 146 |
+
"rewards/chosen": -0.01198151707649231,
|
| 147 |
+
"rewards/margins": 0.005548244807869196,
|
| 148 |
+
"rewards/rejected": -0.017529763281345367,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.08,
|
| 153 |
+
"learning_rate": 3.90625e-06,
|
| 154 |
+
"logits/chosen": 0.6403722763061523,
|
| 155 |
+
"logits/rejected": 0.6439489126205444,
|
| 156 |
+
"logps/chosen": -265.2840881347656,
|
| 157 |
+
"logps/rejected": -230.4324493408203,
|
| 158 |
+
"loss": 0.0536,
|
| 159 |
+
"rewards/accuracies": 0.543749988079071,
|
| 160 |
+
"rewards/chosen": -0.018176386132836342,
|
| 161 |
+
"rewards/margins": 0.005613826680928469,
|
| 162 |
+
"rewards/rejected": -0.023790214210748672,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.09,
|
| 167 |
+
"learning_rate": 4.296875e-06,
|
| 168 |
+
"logits/chosen": 0.7128833532333374,
|
| 169 |
+
"logits/rejected": 0.5932031869888306,
|
| 170 |
+
"logps/chosen": -304.94171142578125,
|
| 171 |
+
"logps/rejected": -258.391845703125,
|
| 172 |
+
"loss": 0.0585,
|
| 173 |
+
"rewards/accuracies": 0.574999988079071,
|
| 174 |
+
"rewards/chosen": -0.028187787160277367,
|
| 175 |
+
"rewards/margins": 0.008357586339116096,
|
| 176 |
+
"rewards/rejected": -0.03654537722468376,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.09,
|
| 181 |
+
"learning_rate": 4.6875000000000004e-06,
|
| 182 |
+
"logits/chosen": 0.6879287958145142,
|
| 183 |
+
"logits/rejected": 0.7084887027740479,
|
| 184 |
+
"logps/chosen": -298.180419921875,
|
| 185 |
+
"logps/rejected": -226.73483276367188,
|
| 186 |
+
"loss": 0.0596,
|
| 187 |
+
"rewards/accuracies": 0.581250011920929,
|
| 188 |
+
"rewards/chosen": -0.04302723705768585,
|
| 189 |
+
"rewards/margins": 0.012504307553172112,
|
| 190 |
+
"rewards/rejected": -0.05553155019879341,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.1,
|
| 195 |
+
"learning_rate": 4.999962359300416e-06,
|
| 196 |
+
"logits/chosen": 0.7091008424758911,
|
| 197 |
+
"logits/rejected": 0.7590427994728088,
|
| 198 |
+
"logps/chosen": -296.51300048828125,
|
| 199 |
+
"logps/rejected": -253.19873046875,
|
| 200 |
+
"loss": 0.0479,
|
| 201 |
+
"rewards/accuracies": 0.5625,
|
| 202 |
+
"rewards/chosen": -0.0437840037047863,
|
| 203 |
+
"rewards/margins": 0.016129568219184875,
|
| 204 |
+
"rewards/rejected": -0.059913575649261475,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.11,
|
| 209 |
+
"learning_rate": 4.998645053824218e-06,
|
| 210 |
+
"logits/chosen": 0.6947115063667297,
|
| 211 |
+
"logits/rejected": 0.6665563583374023,
|
| 212 |
+
"logps/chosen": -335.786865234375,
|
| 213 |
+
"logps/rejected": -286.8999938964844,
|
| 214 |
+
"loss": 0.0472,
|
| 215 |
+
"rewards/accuracies": 0.606249988079071,
|
| 216 |
+
"rewards/chosen": -0.048582445830106735,
|
| 217 |
+
"rewards/margins": 0.025500774383544922,
|
| 218 |
+
"rewards/rejected": -0.07408322393894196,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.12,
|
| 223 |
+
"learning_rate": 4.9954468466732145e-06,
|
| 224 |
+
"logits/chosen": 0.6275844573974609,
|
| 225 |
+
"logits/rejected": 0.6720636487007141,
|
| 226 |
+
"logps/chosen": -296.78680419921875,
|
| 227 |
+
"logps/rejected": -272.51873779296875,
|
| 228 |
+
"loss": 0.0443,
|
| 229 |
+
"rewards/accuracies": 0.543749988079071,
|
| 230 |
+
"rewards/chosen": -0.07314912974834442,
|
| 231 |
+
"rewards/margins": 0.01452522724866867,
|
| 232 |
+
"rewards/rejected": -0.08767435699701309,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.13,
|
| 237 |
+
"learning_rate": 4.990370145357496e-06,
|
| 238 |
+
"logits/chosen": 0.5851293802261353,
|
| 239 |
+
"logits/rejected": 0.637703537940979,
|
| 240 |
+
"logps/chosen": -273.9035339355469,
|
| 241 |
+
"logps/rejected": -253.0749969482422,
|
| 242 |
+
"loss": 0.0501,
|
| 243 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 244 |
+
"rewards/chosen": -0.08011965453624725,
|
| 245 |
+
"rewards/margins": 0.028026139363646507,
|
| 246 |
+
"rewards/rejected": -0.10814578831195831,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.13,
|
| 251 |
+
"learning_rate": 4.983418771458684e-06,
|
| 252 |
+
"logits/chosen": 0.6668591499328613,
|
| 253 |
+
"logits/rejected": 0.6740854978561401,
|
| 254 |
+
"logps/chosen": -326.5847473144531,
|
| 255 |
+
"logps/rejected": -284.8072814941406,
|
| 256 |
+
"loss": 0.0466,
|
| 257 |
+
"rewards/accuracies": 0.581250011920929,
|
| 258 |
+
"rewards/chosen": -0.08650139719247818,
|
| 259 |
+
"rewards/margins": 0.02333686873316765,
|
| 260 |
+
"rewards/rejected": -0.10983824729919434,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.14,
|
| 265 |
+
"learning_rate": 4.97459795775315e-06,
|
| 266 |
+
"logits/chosen": 0.6102133393287659,
|
| 267 |
+
"logits/rejected": 0.6243614554405212,
|
| 268 |
+
"logps/chosen": -325.43829345703125,
|
| 269 |
+
"logps/rejected": -297.8895263671875,
|
| 270 |
+
"loss": 0.0469,
|
| 271 |
+
"rewards/accuracies": 0.59375,
|
| 272 |
+
"rewards/chosen": -0.10013189166784286,
|
| 273 |
+
"rewards/margins": 0.029614349827170372,
|
| 274 |
+
"rewards/rejected": -0.12974625825881958,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.15,
|
| 279 |
+
"learning_rate": 4.963914344272961e-06,
|
| 280 |
+
"logits/chosen": 0.5422903895378113,
|
| 281 |
+
"logits/rejected": 0.6029689908027649,
|
| 282 |
+
"logps/chosen": -304.6243591308594,
|
| 283 |
+
"logps/rejected": -265.31231689453125,
|
| 284 |
+
"loss": 0.0432,
|
| 285 |
+
"rewards/accuracies": 0.606249988079071,
|
| 286 |
+
"rewards/chosen": -0.10887263715267181,
|
| 287 |
+
"rewards/margins": 0.029181605204939842,
|
| 288 |
+
"rewards/rejected": -0.1380542516708374,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.16,
|
| 293 |
+
"learning_rate": 4.951375973307458e-06,
|
| 294 |
+
"logits/chosen": 0.5463661551475525,
|
| 295 |
+
"logits/rejected": 0.5838527679443359,
|
| 296 |
+
"logps/chosen": -301.7105712890625,
|
| 297 |
+
"logps/rejected": -276.5257568359375,
|
| 298 |
+
"loss": 0.0474,
|
| 299 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 300 |
+
"rewards/chosen": -0.11956067383289337,
|
| 301 |
+
"rewards/margins": 0.026623660698533058,
|
| 302 |
+
"rewards/rejected": -0.14618434011936188,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.16,
|
| 307 |
+
"learning_rate": 4.93699228334928e-06,
|
| 308 |
+
"logits/chosen": 0.6322023868560791,
|
| 309 |
+
"logits/rejected": 0.5507072806358337,
|
| 310 |
+
"logps/chosen": -316.19110107421875,
|
| 311 |
+
"logps/rejected": -265.67608642578125,
|
| 312 |
+
"loss": 0.0423,
|
| 313 |
+
"rewards/accuracies": 0.643750011920929,
|
| 314 |
+
"rewards/chosen": -0.11251308768987656,
|
| 315 |
+
"rewards/margins": 0.04010463505983353,
|
| 316 |
+
"rewards/rejected": -0.15261772274971008,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.17,
|
| 321 |
+
"learning_rate": 4.920774101989362e-06,
|
| 322 |
+
"logits/chosen": 0.5577625036239624,
|
| 323 |
+
"logits/rejected": 0.5779579877853394,
|
| 324 |
+
"logps/chosen": -311.96624755859375,
|
| 325 |
+
"logps/rejected": -279.3911437988281,
|
| 326 |
+
"loss": 0.046,
|
| 327 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 328 |
+
"rewards/chosen": -0.11701588332653046,
|
| 329 |
+
"rewards/margins": 0.032017000019550323,
|
| 330 |
+
"rewards/rejected": -0.14903289079666138,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.18,
|
| 335 |
+
"learning_rate": 4.902733637766261e-06,
|
| 336 |
+
"logits/chosen": 0.54422527551651,
|
| 337 |
+
"logits/rejected": 0.5549877285957336,
|
| 338 |
+
"logps/chosen": -326.54290771484375,
|
| 339 |
+
"logps/rejected": -289.0083923339844,
|
| 340 |
+
"loss": 0.0472,
|
| 341 |
+
"rewards/accuracies": 0.625,
|
| 342 |
+
"rewards/chosen": -0.12084343284368515,
|
| 343 |
+
"rewards/margins": 0.04505180940032005,
|
| 344 |
+
"rewards/rejected": -0.1658952534198761,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.19,
|
| 349 |
+
"learning_rate": 4.882884470975954e-06,
|
| 350 |
+
"logits/chosen": 0.4986969828605652,
|
| 351 |
+
"logits/rejected": 0.5934550166130066,
|
| 352 |
+
"logps/chosen": -280.1143798828125,
|
| 353 |
+
"logps/rejected": -251.39358520507812,
|
| 354 |
+
"loss": 0.0437,
|
| 355 |
+
"rewards/accuracies": 0.5625,
|
| 356 |
+
"rewards/chosen": -0.135459303855896,
|
| 357 |
+
"rewards/margins": 0.034046441316604614,
|
| 358 |
+
"rewards/rejected": -0.16950573027133942,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.2,
|
| 363 |
+
"learning_rate": 4.861241543449015e-06,
|
| 364 |
+
"logits/chosen": 0.5392612218856812,
|
| 365 |
+
"logits/rejected": 0.577301025390625,
|
| 366 |
+
"logps/chosen": -316.1842346191406,
|
| 367 |
+
"logps/rejected": -293.55126953125,
|
| 368 |
+
"loss": 0.0368,
|
| 369 |
+
"rewards/accuracies": 0.574999988079071,
|
| 370 |
+
"rewards/chosen": -0.14570352435112,
|
| 371 |
+
"rewards/margins": 0.02999885380268097,
|
| 372 |
+
"rewards/rejected": -0.17570236325263977,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.2,
|
| 377 |
+
"learning_rate": 4.8378211473028755e-06,
|
| 378 |
+
"logits/chosen": 0.5802425146102905,
|
| 379 |
+
"logits/rejected": 0.5668612122535706,
|
| 380 |
+
"logps/chosen": -284.0962829589844,
|
| 381 |
+
"logps/rejected": -251.66818237304688,
|
| 382 |
+
"loss": 0.0404,
|
| 383 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 384 |
+
"rewards/chosen": -0.12709495425224304,
|
| 385 |
+
"rewards/margins": 0.05544697493314743,
|
| 386 |
+
"rewards/rejected": -0.18254193663597107,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.21,
|
| 391 |
+
"learning_rate": 4.812640912677624e-06,
|
| 392 |
+
"logits/chosen": 0.485908180475235,
|
| 393 |
+
"logits/rejected": 0.5380377769470215,
|
| 394 |
+
"logps/chosen": -273.9252624511719,
|
| 395 |
+
"logps/rejected": -245.73794555664062,
|
| 396 |
+
"loss": 0.0488,
|
| 397 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 398 |
+
"rewards/chosen": -0.12756113708019257,
|
| 399 |
+
"rewards/margins": 0.020655876025557518,
|
| 400 |
+
"rewards/rejected": -0.14821702241897583,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.22,
|
| 405 |
+
"learning_rate": 4.785719794464596e-06,
|
| 406 |
+
"logits/chosen": 0.43520134687423706,
|
| 407 |
+
"logits/rejected": 0.5617900490760803,
|
| 408 |
+
"logps/chosen": -343.7205505371094,
|
| 409 |
+
"logps/rejected": -276.3643798828125,
|
| 410 |
+
"loss": 0.0404,
|
| 411 |
+
"rewards/accuracies": 0.59375,
|
| 412 |
+
"rewards/chosen": -0.13573798537254333,
|
| 413 |
+
"rewards/margins": 0.043094925582408905,
|
| 414 |
+
"rewards/rejected": -0.17883291840553284,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.23,
|
| 419 |
+
"learning_rate": 4.757078058037722e-06,
|
| 420 |
+
"logits/chosen": 0.4410250186920166,
|
| 421 |
+
"logits/rejected": 0.5906217694282532,
|
| 422 |
+
"logps/chosen": -331.30657958984375,
|
| 423 |
+
"logps/rejected": -277.7027282714844,
|
| 424 |
+
"loss": 0.0408,
|
| 425 |
+
"rewards/accuracies": 0.612500011920929,
|
| 426 |
+
"rewards/chosen": -0.11059264838695526,
|
| 427 |
+
"rewards/margins": 0.04809447377920151,
|
| 428 |
+
"rewards/rejected": -0.15868714451789856,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.24,
|
| 433 |
+
"learning_rate": 4.72673726399839e-06,
|
| 434 |
+
"logits/chosen": 0.5095167756080627,
|
| 435 |
+
"logits/rejected": 0.6414391398429871,
|
| 436 |
+
"logps/chosen": -262.4388122558594,
|
| 437 |
+
"logps/rejected": -231.455322265625,
|
| 438 |
+
"loss": 0.0392,
|
| 439 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 440 |
+
"rewards/chosen": -0.12092028558254242,
|
| 441 |
+
"rewards/margins": 0.02815370261669159,
|
| 442 |
+
"rewards/rejected": -0.1490740031003952,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.24,
|
| 447 |
+
"learning_rate": 4.694720251945298e-06,
|
| 448 |
+
"logits/chosen": 0.5478841066360474,
|
| 449 |
+
"logits/rejected": 0.4881111979484558,
|
| 450 |
+
"logps/chosen": -321.4375305175781,
|
| 451 |
+
"logps/rejected": -280.4942932128906,
|
| 452 |
+
"loss": 0.0413,
|
| 453 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 454 |
+
"rewards/chosen": -0.12603510916233063,
|
| 455 |
+
"rewards/margins": 0.04496049880981445,
|
| 456 |
+
"rewards/rejected": -0.17099562287330627,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.25,
|
| 461 |
+
"learning_rate": 4.661051123281528e-06,
|
| 462 |
+
"logits/chosen": 0.5421676635742188,
|
| 463 |
+
"logits/rejected": 0.5195788145065308,
|
| 464 |
+
"logps/chosen": -291.16778564453125,
|
| 465 |
+
"logps/rejected": -270.92779541015625,
|
| 466 |
+
"loss": 0.0416,
|
| 467 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 468 |
+
"rewards/chosen": -0.1264342963695526,
|
| 469 |
+
"rewards/margins": 0.042602840811014175,
|
| 470 |
+
"rewards/rejected": -0.1690371334552765,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.26,
|
| 475 |
+
"learning_rate": 4.6257552230717536e-06,
|
| 476 |
+
"logits/chosen": 0.567541778087616,
|
| 477 |
+
"logits/rejected": 0.5551782250404358,
|
| 478 |
+
"logps/chosen": -292.7509765625,
|
| 479 |
+
"logps/rejected": -310.8721618652344,
|
| 480 |
+
"loss": 0.0372,
|
| 481 |
+
"rewards/accuracies": 0.65625,
|
| 482 |
+
"rewards/chosen": -0.15939810872077942,
|
| 483 |
+
"rewards/margins": 0.0593896210193634,
|
| 484 |
+
"rewards/rejected": -0.21878774464130402,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.27,
|
| 489 |
+
"learning_rate": 4.588859120963282e-06,
|
| 490 |
+
"logits/chosen": 0.44213318824768066,
|
| 491 |
+
"logits/rejected": 0.4428648352622986,
|
| 492 |
+
"logps/chosen": -260.6324768066406,
|
| 493 |
+
"logps/rejected": -255.5272674560547,
|
| 494 |
+
"loss": 0.0452,
|
| 495 |
+
"rewards/accuracies": 0.581250011920929,
|
| 496 |
+
"rewards/chosen": -0.15107281506061554,
|
| 497 |
+
"rewards/margins": 0.03870910778641701,
|
| 498 |
+
"rewards/rejected": -0.18978191912174225,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.27,
|
| 503 |
+
"learning_rate": 4.5503905911852435e-06,
|
| 504 |
+
"logits/chosen": 0.5433288216590881,
|
| 505 |
+
"logits/rejected": 0.5616317987442017,
|
| 506 |
+
"logps/chosen": -283.3091735839844,
|
| 507 |
+
"logps/rejected": -249.4799041748047,
|
| 508 |
+
"loss": 0.0441,
|
| 509 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 510 |
+
"rewards/chosen": -0.163166344165802,
|
| 511 |
+
"rewards/margins": 0.03194192424416542,
|
| 512 |
+
"rewards/rejected": -0.19510827958583832,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.28,
|
| 517 |
+
"learning_rate": 4.510378591641036e-06,
|
| 518 |
+
"logits/chosen": 0.5152170658111572,
|
| 519 |
+
"logits/rejected": 0.6420885324478149,
|
| 520 |
+
"logps/chosen": -339.47576904296875,
|
| 521 |
+
"logps/rejected": -327.2655334472656,
|
| 522 |
+
"loss": 0.0351,
|
| 523 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 524 |
+
"rewards/chosen": -0.16583441197872162,
|
| 525 |
+
"rewards/margins": 0.04150914400815964,
|
| 526 |
+
"rewards/rejected": -0.20734354853630066,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.29,
|
| 531 |
+
"learning_rate": 4.468853242109712e-06,
|
| 532 |
+
"logits/chosen": 0.5886040925979614,
|
| 533 |
+
"logits/rejected": 0.5353233814239502,
|
| 534 |
+
"logps/chosen": -294.714599609375,
|
| 535 |
+
"logps/rejected": -272.34649658203125,
|
| 536 |
+
"loss": 0.0404,
|
| 537 |
+
"rewards/accuracies": 0.612500011920929,
|
| 538 |
+
"rewards/chosen": -0.16863426566123962,
|
| 539 |
+
"rewards/margins": 0.04221067577600479,
|
| 540 |
+
"rewards/rejected": -0.2108449637889862,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.3,
|
| 545 |
+
"learning_rate": 4.42584580157276e-06,
|
| 546 |
+
"logits/chosen": 0.4872845709323883,
|
| 547 |
+
"logits/rejected": 0.6496473550796509,
|
| 548 |
+
"logps/chosen": -322.4450988769531,
|
| 549 |
+
"logps/rejected": -284.5151062011719,
|
| 550 |
+
"loss": 0.0396,
|
| 551 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 552 |
+
"rewards/chosen": -0.1785622537136078,
|
| 553 |
+
"rewards/margins": 0.04358912259340286,
|
| 554 |
+
"rewards/rejected": -0.22215136885643005,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.31,
|
| 559 |
+
"learning_rate": 4.381388644683317e-06,
|
| 560 |
+
"logits/chosen": 0.5238572359085083,
|
| 561 |
+
"logits/rejected": 0.5852701663970947,
|
| 562 |
+
"logps/chosen": -330.46746826171875,
|
| 563 |
+
"logps/rejected": -284.7983093261719,
|
| 564 |
+
"loss": 0.0413,
|
| 565 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 566 |
+
"rewards/chosen": -0.1601935178041458,
|
| 567 |
+
"rewards/margins": 0.04235311970114708,
|
| 568 |
+
"rewards/rejected": -0.202546626329422,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.31,
|
| 573 |
+
"learning_rate": 4.33551523739555e-06,
|
| 574 |
+
"logits/chosen": 0.5760276913642883,
|
| 575 |
+
"logits/rejected": 0.6830281615257263,
|
| 576 |
+
"logps/chosen": -317.65252685546875,
|
| 577 |
+
"logps/rejected": -259.23590087890625,
|
| 578 |
+
"loss": 0.0411,
|
| 579 |
+
"rewards/accuracies": 0.643750011920929,
|
| 580 |
+
"rewards/chosen": -0.1589946150779724,
|
| 581 |
+
"rewards/margins": 0.04796143248677254,
|
| 582 |
+
"rewards/rejected": -0.20695605874061584,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.32,
|
| 587 |
+
"learning_rate": 4.288260111772535e-06,
|
| 588 |
+
"logits/chosen": 0.6023763418197632,
|
| 589 |
+
"logits/rejected": 0.6028311252593994,
|
| 590 |
+
"logps/chosen": -301.0093994140625,
|
| 591 |
+
"logps/rejected": -288.390625,
|
| 592 |
+
"loss": 0.0399,
|
| 593 |
+
"rewards/accuracies": 0.637499988079071,
|
| 594 |
+
"rewards/chosen": -0.15324203670024872,
|
| 595 |
+
"rewards/margins": 0.06706853210926056,
|
| 596 |
+
"rewards/rejected": -0.22031056880950928,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.33,
|
| 601 |
+
"learning_rate": 4.239658839991594e-06,
|
| 602 |
+
"logits/chosen": 0.4882682263851166,
|
| 603 |
+
"logits/rejected": 0.6300086975097656,
|
| 604 |
+
"logps/chosen": -326.09796142578125,
|
| 605 |
+
"logps/rejected": -292.12744140625,
|
| 606 |
+
"loss": 0.0367,
|
| 607 |
+
"rewards/accuracies": 0.675000011920929,
|
| 608 |
+
"rewards/chosen": -0.15980076789855957,
|
| 609 |
+
"rewards/margins": 0.05597474053502083,
|
| 610 |
+
"rewards/rejected": -0.2157754898071289,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.34,
|
| 615 |
+
"learning_rate": 4.189748007566686e-06,
|
| 616 |
+
"logits/chosen": 0.6084526181221008,
|
| 617 |
+
"logits/rejected": 0.6407297849655151,
|
| 618 |
+
"logps/chosen": -295.1040954589844,
|
| 619 |
+
"logps/rejected": -283.36199951171875,
|
| 620 |
+
"loss": 0.0415,
|
| 621 |
+
"rewards/accuracies": 0.612500011920929,
|
| 622 |
+
"rewards/chosen": -0.14586099982261658,
|
| 623 |
+
"rewards/margins": 0.044940222054719925,
|
| 624 |
+
"rewards/rejected": -0.1908012181520462,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.35,
|
| 629 |
+
"learning_rate": 4.138565185807972e-06,
|
| 630 |
+
"logits/chosen": 0.6028437614440918,
|
| 631 |
+
"logits/rejected": 0.5474241971969604,
|
| 632 |
+
"logps/chosen": -259.4440002441406,
|
| 633 |
+
"logps/rejected": -262.1929931640625,
|
| 634 |
+
"loss": 0.0457,
|
| 635 |
+
"rewards/accuracies": 0.543749988079071,
|
| 636 |
+
"rewards/chosen": -0.13820604979991913,
|
| 637 |
+
"rewards/margins": 0.04853081330657005,
|
| 638 |
+
"rewards/rejected": -0.18673686683177948,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.35,
|
| 643 |
+
"learning_rate": 4.086148903539311e-06,
|
| 644 |
+
"logits/chosen": 0.5923335552215576,
|
| 645 |
+
"logits/rejected": 0.6619054079055786,
|
| 646 |
+
"logps/chosen": -295.82440185546875,
|
| 647 |
+
"logps/rejected": -260.25653076171875,
|
| 648 |
+
"loss": 0.0387,
|
| 649 |
+
"rewards/accuracies": 0.606249988079071,
|
| 650 |
+
"rewards/chosen": -0.14740963280200958,
|
| 651 |
+
"rewards/margins": 0.04628303274512291,
|
| 652 |
+
"rewards/rejected": -0.19369268417358398,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.36,
|
| 657 |
+
"learning_rate": 4.032538618094972e-06,
|
| 658 |
+
"logits/chosen": 0.5214491486549377,
|
| 659 |
+
"logits/rejected": 0.6175715923309326,
|
| 660 |
+
"logps/chosen": -275.34234619140625,
|
| 661 |
+
"logps/rejected": -263.62628173828125,
|
| 662 |
+
"loss": 0.0375,
|
| 663 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 664 |
+
"rewards/chosen": -0.1500253975391388,
|
| 665 |
+
"rewards/margins": 0.058636367321014404,
|
| 666 |
+
"rewards/rejected": -0.2086617648601532,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.37,
|
| 671 |
+
"learning_rate": 3.977774685617386e-06,
|
| 672 |
+
"logits/chosen": 0.5247635841369629,
|
| 673 |
+
"logits/rejected": 0.5686976909637451,
|
| 674 |
+
"logps/chosen": -331.62969970703125,
|
| 675 |
+
"logps/rejected": -301.4044494628906,
|
| 676 |
+
"loss": 0.0342,
|
| 677 |
+
"rewards/accuracies": 0.53125,
|
| 678 |
+
"rewards/chosen": -0.15346726775169373,
|
| 679 |
+
"rewards/margins": 0.03650849312543869,
|
| 680 |
+
"rewards/rejected": -0.18997573852539062,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.38,
|
| 685 |
+
"learning_rate": 3.92189833067831e-06,
|
| 686 |
+
"logits/chosen": 0.48843416571617126,
|
| 687 |
+
"logits/rejected": 0.5862494707107544,
|
| 688 |
+
"logps/chosen": -315.828857421875,
|
| 689 |
+
"logps/rejected": -266.48931884765625,
|
| 690 |
+
"loss": 0.0396,
|
| 691 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 692 |
+
"rewards/chosen": -0.13729068636894226,
|
| 693 |
+
"rewards/margins": 0.05156175419688225,
|
| 694 |
+
"rewards/rejected": -0.18885241448879242,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.38,
|
| 699 |
+
"learning_rate": 3.864951615246261e-06,
|
| 700 |
+
"logits/chosen": 0.5482813119888306,
|
| 701 |
+
"logits/rejected": 0.5736523866653442,
|
| 702 |
+
"logps/chosen": -312.53289794921875,
|
| 703 |
+
"logps/rejected": -297.5661315917969,
|
| 704 |
+
"loss": 0.0393,
|
| 705 |
+
"rewards/accuracies": 0.668749988079071,
|
| 706 |
+
"rewards/chosen": -0.15364210307598114,
|
| 707 |
+
"rewards/margins": 0.06609909236431122,
|
| 708 |
+
"rewards/rejected": -0.21974118053913116,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.39,
|
| 713 |
+
"learning_rate": 3.806977407023581e-06,
|
| 714 |
+
"logits/chosen": 0.6043223142623901,
|
| 715 |
+
"logits/rejected": 0.5951470136642456,
|
| 716 |
+
"logps/chosen": -287.01416015625,
|
| 717 |
+
"logps/rejected": -266.211669921875,
|
| 718 |
+
"loss": 0.0317,
|
| 719 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 720 |
+
"rewards/chosen": -0.15719470381736755,
|
| 721 |
+
"rewards/margins": 0.05330941826105118,
|
| 722 |
+
"rewards/rejected": -0.21050409972667694,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.4,
|
| 727 |
+
"learning_rate": 3.7480193471769815e-06,
|
| 728 |
+
"logits/chosen": 0.508133053779602,
|
| 729 |
+
"logits/rejected": 0.5117858648300171,
|
| 730 |
+
"logps/chosen": -303.5716857910156,
|
| 731 |
+
"logps/rejected": -281.1479797363281,
|
| 732 |
+
"loss": 0.0433,
|
| 733 |
+
"rewards/accuracies": 0.550000011920929,
|
| 734 |
+
"rewards/chosen": -0.1687772274017334,
|
| 735 |
+
"rewards/margins": 0.04205339774489403,
|
| 736 |
+
"rewards/rejected": -0.21083059906959534,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.41,
|
| 741 |
+
"learning_rate": 3.6881218174858354e-06,
|
| 742 |
+
"logits/chosen": 0.5807247161865234,
|
| 743 |
+
"logits/rejected": 0.5294843912124634,
|
| 744 |
+
"logps/chosen": -325.0351867675781,
|
| 745 |
+
"logps/rejected": -303.7823791503906,
|
| 746 |
+
"loss": 0.0406,
|
| 747 |
+
"rewards/accuracies": 0.625,
|
| 748 |
+
"rewards/chosen": -0.17659473419189453,
|
| 749 |
+
"rewards/margins": 0.052628111094236374,
|
| 750 |
+
"rewards/rejected": -0.2292228490114212,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.42,
|
| 755 |
+
"learning_rate": 3.627329906932964e-06,
|
| 756 |
+
"logits/chosen": 0.5673174262046814,
|
| 757 |
+
"logits/rejected": 0.5497801303863525,
|
| 758 |
+
"logps/chosen": -339.43023681640625,
|
| 759 |
+
"logps/rejected": -292.1569519042969,
|
| 760 |
+
"loss": 0.039,
|
| 761 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 762 |
+
"rewards/chosen": -0.1955457627773285,
|
| 763 |
+
"rewards/margins": 0.05005335807800293,
|
| 764 |
+
"rewards/rejected": -0.24559912085533142,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.42,
|
| 769 |
+
"learning_rate": 3.5656893777630686e-06,
|
| 770 |
+
"logits/chosen": 0.5412625670433044,
|
| 771 |
+
"logits/rejected": 0.6069876551628113,
|
| 772 |
+
"logps/chosen": -285.06072998046875,
|
| 773 |
+
"logps/rejected": -270.3983459472656,
|
| 774 |
+
"loss": 0.038,
|
| 775 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 776 |
+
"rewards/chosen": -0.1897256076335907,
|
| 777 |
+
"rewards/margins": 0.0756111592054367,
|
| 778 |
+
"rewards/rejected": -0.2653367817401886,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.43,
|
| 783 |
+
"learning_rate": 3.503246631034345e-06,
|
| 784 |
+
"logits/chosen": 0.5480440855026245,
|
| 785 |
+
"logits/rejected": 0.5790668725967407,
|
| 786 |
+
"logps/chosen": -278.4317932128906,
|
| 787 |
+
"logps/rejected": -268.5929260253906,
|
| 788 |
+
"loss": 0.0412,
|
| 789 |
+
"rewards/accuracies": 0.543749988079071,
|
| 790 |
+
"rewards/chosen": -0.16996586322784424,
|
| 791 |
+
"rewards/margins": 0.05341961234807968,
|
| 792 |
+
"rewards/rejected": -0.22338548302650452,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.44,
|
| 797 |
+
"learning_rate": 3.440048671689219e-06,
|
| 798 |
+
"logits/chosen": 0.5717700123786926,
|
| 799 |
+
"logits/rejected": 0.6006878614425659,
|
| 800 |
+
"logps/chosen": -311.20782470703125,
|
| 801 |
+
"logps/rejected": -283.98321533203125,
|
| 802 |
+
"loss": 0.0373,
|
| 803 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 804 |
+
"rewards/chosen": -0.1866205632686615,
|
| 805 |
+
"rewards/margins": 0.05927601456642151,
|
| 806 |
+
"rewards/rejected": -0.24589654803276062,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.45,
|
| 811 |
+
"learning_rate": 3.3761430731705056e-06,
|
| 812 |
+
"logits/chosen": 0.5225729942321777,
|
| 813 |
+
"logits/rejected": 0.6431065797805786,
|
| 814 |
+
"logps/chosen": -302.95758056640625,
|
| 815 |
+
"logps/rejected": -277.3609313964844,
|
| 816 |
+
"loss": 0.0322,
|
| 817 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 818 |
+
"rewards/chosen": -0.18959183990955353,
|
| 819 |
+
"rewards/margins": 0.055399276316165924,
|
| 820 |
+
"rewards/rejected": -0.24499110877513885,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.46,
|
| 825 |
+
"learning_rate": 3.311577941609604e-06,
|
| 826 |
+
"logits/chosen": 0.6244104504585266,
|
| 827 |
+
"logits/rejected": 0.5891777873039246,
|
| 828 |
+
"logps/chosen": -266.7265319824219,
|
| 829 |
+
"logps/rejected": -242.4050750732422,
|
| 830 |
+
"loss": 0.043,
|
| 831 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 832 |
+
"rewards/chosen": -0.18597814440727234,
|
| 833 |
+
"rewards/margins": 0.04474898427724838,
|
| 834 |
+
"rewards/rejected": -0.2307271510362625,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.46,
|
| 839 |
+
"learning_rate": 3.2464018796137157e-06,
|
| 840 |
+
"logits/chosen": 0.5826702117919922,
|
| 841 |
+
"logits/rejected": 0.5641965270042419,
|
| 842 |
+
"logps/chosen": -303.88970947265625,
|
| 843 |
+
"logps/rejected": -268.25640869140625,
|
| 844 |
+
"loss": 0.0352,
|
| 845 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 846 |
+
"rewards/chosen": -0.17130930721759796,
|
| 847 |
+
"rewards/margins": 0.037886954843997955,
|
| 848 |
+
"rewards/rejected": -0.20919625461101532,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.47,
|
| 853 |
+
"learning_rate": 3.1806639496793245e-06,
|
| 854 |
+
"logits/chosen": 0.6168242692947388,
|
| 855 |
+
"logits/rejected": 0.6149759888648987,
|
| 856 |
+
"logps/chosen": -285.6898193359375,
|
| 857 |
+
"logps/rejected": -286.07379150390625,
|
| 858 |
+
"loss": 0.0402,
|
| 859 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 860 |
+
"rewards/chosen": -0.14284178614616394,
|
| 861 |
+
"rewards/margins": 0.050834327936172485,
|
| 862 |
+
"rewards/rejected": -0.19367611408233643,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.48,
|
| 867 |
+
"learning_rate": 3.114413637259484e-06,
|
| 868 |
+
"logits/chosen": 0.5781451463699341,
|
| 869 |
+
"logits/rejected": 0.6401593089103699,
|
| 870 |
+
"logps/chosen": -303.3341064453125,
|
| 871 |
+
"logps/rejected": -291.5891418457031,
|
| 872 |
+
"loss": 0.0394,
|
| 873 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 874 |
+
"rewards/chosen": -0.14258936047554016,
|
| 875 |
+
"rewards/margins": 0.04603290185332298,
|
| 876 |
+
"rewards/rejected": -0.18862226605415344,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.49,
|
| 881 |
+
"learning_rate": 3.0477008135127247e-06,
|
| 882 |
+
"logits/chosen": 0.5642613768577576,
|
| 883 |
+
"logits/rejected": 0.4775732159614563,
|
| 884 |
+
"logps/chosen": -308.5189208984375,
|
| 885 |
+
"logps/rejected": -277.7254333496094,
|
| 886 |
+
"loss": 0.0406,
|
| 887 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 888 |
+
"rewards/chosen": -0.15040606260299683,
|
| 889 |
+
"rewards/margins": 0.047120168805122375,
|
| 890 |
+
"rewards/rejected": -0.1975262314081192,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.49,
|
| 895 |
+
"learning_rate": 2.980575697761603e-06,
|
| 896 |
+
"logits/chosen": 0.5864249467849731,
|
| 897 |
+
"logits/rejected": 0.585323691368103,
|
| 898 |
+
"logps/chosen": -279.18060302734375,
|
| 899 |
+
"logps/rejected": -274.70880126953125,
|
| 900 |
+
"loss": 0.0431,
|
| 901 |
+
"rewards/accuracies": 0.612500011920929,
|
| 902 |
+
"rewards/chosen": -0.1575286090373993,
|
| 903 |
+
"rewards/margins": 0.05304408073425293,
|
| 904 |
+
"rewards/rejected": -0.21057268977165222,
|
| 905 |
+
"step": 630
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.5,
|
| 909 |
+
"learning_rate": 2.9130888196891755e-06,
|
| 910 |
+
"logits/chosen": 0.4903879165649414,
|
| 911 |
+
"logits/rejected": 0.5989735126495361,
|
| 912 |
+
"logps/chosen": -272.58428955078125,
|
| 913 |
+
"logps/rejected": -251.16531372070312,
|
| 914 |
+
"loss": 0.0431,
|
| 915 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 916 |
+
"rewards/chosen": -0.1445484459400177,
|
| 917 |
+
"rewards/margins": 0.022020401433110237,
|
| 918 |
+
"rewards/rejected": -0.1665688455104828,
|
| 919 |
+
"step": 640
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.51,
|
| 923 |
+
"learning_rate": 2.845290981301834e-06,
|
| 924 |
+
"logits/chosen": 0.5647401809692383,
|
| 925 |
+
"logits/rejected": 0.5633600950241089,
|
| 926 |
+
"logps/chosen": -294.28594970703125,
|
| 927 |
+
"logps/rejected": -258.2326965332031,
|
| 928 |
+
"loss": 0.0347,
|
| 929 |
+
"rewards/accuracies": 0.574999988079071,
|
| 930 |
+
"rewards/chosen": -0.15951719880104065,
|
| 931 |
+
"rewards/margins": 0.04437902569770813,
|
| 932 |
+
"rewards/rejected": -0.20389623939990997,
|
| 933 |
+
"step": 650
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.52,
|
| 937 |
+
"learning_rate": 2.7772332186871464e-06,
|
| 938 |
+
"logits/chosen": 0.5554920434951782,
|
| 939 |
+
"logits/rejected": 0.6650098562240601,
|
| 940 |
+
"logps/chosen": -307.80804443359375,
|
| 941 |
+
"logps/rejected": -271.55963134765625,
|
| 942 |
+
"loss": 0.0397,
|
| 943 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 944 |
+
"rewards/chosen": -0.1580692082643509,
|
| 945 |
+
"rewards/margins": 0.052956365048885345,
|
| 946 |
+
"rewards/rejected": -0.21102559566497803,
|
| 947 |
+
"step": 660
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.53,
|
| 951 |
+
"learning_rate": 2.708966763595493e-06,
|
| 952 |
+
"logits/chosen": 0.5610272884368896,
|
| 953 |
+
"logits/rejected": 0.6106783151626587,
|
| 954 |
+
"logps/chosen": -309.99945068359375,
|
| 955 |
+
"logps/rejected": -258.3206481933594,
|
| 956 |
+
"loss": 0.0395,
|
| 957 |
+
"rewards/accuracies": 0.612500011920929,
|
| 958 |
+
"rewards/chosen": -0.16326628625392914,
|
| 959 |
+
"rewards/margins": 0.06232718378305435,
|
| 960 |
+
"rewards/rejected": -0.22559349238872528,
|
| 961 |
+
"step": 670
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.53,
|
| 965 |
+
"learning_rate": 2.640543004874409e-06,
|
| 966 |
+
"logits/chosen": 0.4919437766075134,
|
| 967 |
+
"logits/rejected": 0.6173873543739319,
|
| 968 |
+
"logps/chosen": -316.8961486816406,
|
| 969 |
+
"logps/rejected": -278.5799255371094,
|
| 970 |
+
"loss": 0.0408,
|
| 971 |
+
"rewards/accuracies": 0.606249988079071,
|
| 972 |
+
"rewards/chosen": -0.15790703892707825,
|
| 973 |
+
"rewards/margins": 0.05107961967587471,
|
| 974 |
+
"rewards/rejected": -0.20898666977882385,
|
| 975 |
+
"step": 680
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.54,
|
| 979 |
+
"learning_rate": 2.572013449784671e-06,
|
| 980 |
+
"logits/chosen": 0.5977932810783386,
|
| 981 |
+
"logits/rejected": 0.5793666839599609,
|
| 982 |
+
"logps/chosen": -274.3351745605469,
|
| 983 |
+
"logps/rejected": -260.75408935546875,
|
| 984 |
+
"loss": 0.0364,
|
| 985 |
+
"rewards/accuracies": 0.574999988079071,
|
| 986 |
+
"rewards/chosen": -0.16240599751472473,
|
| 987 |
+
"rewards/margins": 0.04127690568566322,
|
| 988 |
+
"rewards/rejected": -0.20368289947509766,
|
| 989 |
+
"step": 690
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.55,
|
| 993 |
+
"learning_rate": 2.503429685227245e-06,
|
| 994 |
+
"logits/chosen": 0.5404385924339294,
|
| 995 |
+
"logits/rejected": 0.5224953293800354,
|
| 996 |
+
"logps/chosen": -333.06292724609375,
|
| 997 |
+
"logps/rejected": -277.642333984375,
|
| 998 |
+
"loss": 0.039,
|
| 999 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1000 |
+
"rewards/chosen": -0.1504869908094406,
|
| 1001 |
+
"rewards/margins": 0.05367060750722885,
|
| 1002 |
+
"rewards/rejected": -0.20415759086608887,
|
| 1003 |
+
"step": 700
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.56,
|
| 1007 |
+
"learning_rate": 2.434843338910286e-06,
|
| 1008 |
+
"logits/chosen": 0.5657359957695007,
|
| 1009 |
+
"logits/rejected": 0.6558262705802917,
|
| 1010 |
+
"logps/chosen": -338.18048095703125,
|
| 1011 |
+
"logps/rejected": -290.9669189453125,
|
| 1012 |
+
"loss": 0.0376,
|
| 1013 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1014 |
+
"rewards/chosen": -0.15950533747673035,
|
| 1015 |
+
"rewards/margins": 0.04541635885834694,
|
| 1016 |
+
"rewards/rejected": -0.2049216777086258,
|
| 1017 |
+
"step": 710
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.57,
|
| 1021 |
+
"learning_rate": 2.3663060404854155e-06,
|
| 1022 |
+
"logits/chosen": 0.5390611886978149,
|
| 1023 |
+
"logits/rejected": 0.5992618799209595,
|
| 1024 |
+
"logps/chosen": -287.187744140625,
|
| 1025 |
+
"logps/rejected": -266.4593505859375,
|
| 1026 |
+
"loss": 0.0403,
|
| 1027 |
+
"rewards/accuracies": 0.574999988079071,
|
| 1028 |
+
"rewards/chosen": -0.16842734813690186,
|
| 1029 |
+
"rewards/margins": 0.05528571456670761,
|
| 1030 |
+
"rewards/rejected": -0.22371308505535126,
|
| 1031 |
+
"step": 720
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.57,
|
| 1035 |
+
"learning_rate": 2.2978693826825406e-06,
|
| 1036 |
+
"logits/chosen": 0.5962172746658325,
|
| 1037 |
+
"logits/rejected": 0.618035614490509,
|
| 1038 |
+
"logps/chosen": -325.6087951660156,
|
| 1039 |
+
"logps/rejected": -282.9885559082031,
|
| 1040 |
+
"loss": 0.0349,
|
| 1041 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1042 |
+
"rewards/chosen": -0.1736389696598053,
|
| 1043 |
+
"rewards/margins": 0.05205898359417915,
|
| 1044 |
+
"rewards/rejected": -0.22569794952869415,
|
| 1045 |
+
"step": 730
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.58,
|
| 1049 |
+
"learning_rate": 2.2295848824724612e-06,
|
| 1050 |
+
"logits/chosen": 0.5612713098526001,
|
| 1051 |
+
"logits/rejected": 0.5915429592132568,
|
| 1052 |
+
"logps/chosen": -273.39300537109375,
|
| 1053 |
+
"logps/rejected": -248.25247192382812,
|
| 1054 |
+
"loss": 0.0474,
|
| 1055 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1056 |
+
"rewards/chosen": -0.1872544288635254,
|
| 1057 |
+
"rewards/margins": 0.06460748612880707,
|
| 1058 |
+
"rewards/rejected": -0.25186192989349365,
|
| 1059 |
+
"step": 740
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.59,
|
| 1063 |
+
"learning_rate": 2.1615039422865136e-06,
|
| 1064 |
+
"logits/chosen": 0.5421571731567383,
|
| 1065 |
+
"logits/rejected": 0.613379955291748,
|
| 1066 |
+
"logps/chosen": -293.1100769042969,
|
| 1067 |
+
"logps/rejected": -261.9356994628906,
|
| 1068 |
+
"loss": 0.0435,
|
| 1069 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1070 |
+
"rewards/chosen": -0.18293720483779907,
|
| 1071 |
+
"rewards/margins": 0.05960521101951599,
|
| 1072 |
+
"rewards/rejected": -0.24254241585731506,
|
| 1073 |
+
"step": 750
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.6,
|
| 1077 |
+
"learning_rate": 2.0936778113224253e-06,
|
| 1078 |
+
"logits/chosen": 0.6936706900596619,
|
| 1079 |
+
"logits/rejected": 0.6555561423301697,
|
| 1080 |
+
"logps/chosen": -301.51629638671875,
|
| 1081 |
+
"logps/rejected": -261.8467102050781,
|
| 1082 |
+
"loss": 0.0408,
|
| 1083 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1084 |
+
"rewards/chosen": -0.18969768285751343,
|
| 1085 |
+
"rewards/margins": 0.055843252688646317,
|
| 1086 |
+
"rewards/rejected": -0.24554093182086945,
|
| 1087 |
+
"step": 760
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.6,
|
| 1091 |
+
"learning_rate": 2.0261575469655304e-06,
|
| 1092 |
+
"logits/chosen": 0.5248943567276001,
|
| 1093 |
+
"logits/rejected": 0.548130214214325,
|
| 1094 |
+
"logps/chosen": -300.0675048828125,
|
| 1095 |
+
"logps/rejected": -273.765869140625,
|
| 1096 |
+
"loss": 0.0363,
|
| 1097 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1098 |
+
"rewards/chosen": -0.19320233166217804,
|
| 1099 |
+
"rewards/margins": 0.05187714099884033,
|
| 1100 |
+
"rewards/rejected": -0.24507947266101837,
|
| 1101 |
+
"step": 770
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.61,
|
| 1105 |
+
"learning_rate": 1.9589939763543693e-06,
|
| 1106 |
+
"logits/chosen": 0.6453245282173157,
|
| 1107 |
+
"logits/rejected": 0.48234623670578003,
|
| 1108 |
+
"logps/chosen": -298.9971008300781,
|
| 1109 |
+
"logps/rejected": -260.1150207519531,
|
| 1110 |
+
"loss": 0.0408,
|
| 1111 |
+
"rewards/accuracies": 0.625,
|
| 1112 |
+
"rewards/chosen": -0.19145645201206207,
|
| 1113 |
+
"rewards/margins": 0.05295104905962944,
|
| 1114 |
+
"rewards/rejected": -0.24440748989582062,
|
| 1115 |
+
"step": 780
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.62,
|
| 1119 |
+
"learning_rate": 1.8922376581196107e-06,
|
| 1120 |
+
"logits/chosen": 0.6130558252334595,
|
| 1121 |
+
"logits/rejected": 0.5399459004402161,
|
| 1122 |
+
"logps/chosen": -336.5523376464844,
|
| 1123 |
+
"logps/rejected": -287.4131774902344,
|
| 1124 |
+
"loss": 0.0434,
|
| 1125 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 1126 |
+
"rewards/chosen": -0.19996806979179382,
|
| 1127 |
+
"rewards/margins": 0.041921354830265045,
|
| 1128 |
+
"rewards/rejected": -0.24188943207263947,
|
| 1129 |
+
"step": 790
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.63,
|
| 1133 |
+
"learning_rate": 1.8259388443250993e-06,
|
| 1134 |
+
"logits/chosen": 0.48818960785865784,
|
| 1135 |
+
"logits/rejected": 0.5838159918785095,
|
| 1136 |
+
"logps/chosen": -295.0948791503906,
|
| 1137 |
+
"logps/rejected": -259.95452880859375,
|
| 1138 |
+
"loss": 0.0422,
|
| 1139 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1140 |
+
"rewards/chosen": -0.16656538844108582,
|
| 1141 |
+
"rewards/margins": 0.04429946839809418,
|
| 1142 |
+
"rewards/rejected": -0.21086485683918,
|
| 1143 |
+
"step": 800
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.64,
|
| 1147 |
+
"learning_rate": 1.760147442639679e-06,
|
| 1148 |
+
"logits/chosen": 0.4529576301574707,
|
| 1149 |
+
"logits/rejected": 0.6206762194633484,
|
| 1150 |
+
"logps/chosen": -310.306396484375,
|
| 1151 |
+
"logps/rejected": -250.24331665039062,
|
| 1152 |
+
"loss": 0.0406,
|
| 1153 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1154 |
+
"rewards/chosen": -0.18583405017852783,
|
| 1155 |
+
"rewards/margins": 0.061879683285951614,
|
| 1156 |
+
"rewards/rejected": -0.24771372973918915,
|
| 1157 |
+
"step": 810
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.64,
|
| 1161 |
+
"learning_rate": 1.6949129787682628e-06,
|
| 1162 |
+
"logits/chosen": 0.4996159076690674,
|
| 1163 |
+
"logits/rejected": 0.5564543008804321,
|
| 1164 |
+
"logps/chosen": -348.6169738769531,
|
| 1165 |
+
"logps/rejected": -285.4058532714844,
|
| 1166 |
+
"loss": 0.0494,
|
| 1167 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1168 |
+
"rewards/chosen": -0.18937376141548157,
|
| 1169 |
+
"rewards/margins": 0.05144171044230461,
|
| 1170 |
+
"rewards/rejected": -0.24081549048423767,
|
| 1171 |
+
"step": 820
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.65,
|
| 1175 |
+
"learning_rate": 1.6302845591704348e-06,
|
| 1176 |
+
"logits/chosen": 0.5382525324821472,
|
| 1177 |
+
"logits/rejected": 0.5930293202400208,
|
| 1178 |
+
"logps/chosen": -290.4473571777344,
|
| 1179 |
+
"logps/rejected": -286.49615478515625,
|
| 1180 |
+
"loss": 0.0448,
|
| 1181 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1182 |
+
"rewards/chosen": -0.17880043387413025,
|
| 1183 |
+
"rewards/margins": 0.05653693526983261,
|
| 1184 |
+
"rewards/rejected": -0.23533737659454346,
|
| 1185 |
+
"step": 830
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.66,
|
| 1189 |
+
"learning_rate": 1.5663108340946465e-06,
|
| 1190 |
+
"logits/chosen": 0.5701301097869873,
|
| 1191 |
+
"logits/rejected": 0.6285163760185242,
|
| 1192 |
+
"logps/chosen": -288.7821960449219,
|
| 1193 |
+
"logps/rejected": -275.9776611328125,
|
| 1194 |
+
"loss": 0.0354,
|
| 1195 |
+
"rewards/accuracies": 0.59375,
|
| 1196 |
+
"rewards/chosen": -0.18739423155784607,
|
| 1197 |
+
"rewards/margins": 0.05587046220898628,
|
| 1198 |
+
"rewards/rejected": -0.24326470494270325,
|
| 1199 |
+
"step": 840
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.67,
|
| 1203 |
+
"learning_rate": 1.5030399609558364e-06,
|
| 1204 |
+
"logits/chosen": 0.4821529984474182,
|
| 1205 |
+
"logits/rejected": 0.545883059501648,
|
| 1206 |
+
"logps/chosen": -295.86456298828125,
|
| 1207 |
+
"logps/rejected": -284.3241271972656,
|
| 1208 |
+
"loss": 0.0417,
|
| 1209 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1210 |
+
"rewards/chosen": -0.20055215060710907,
|
| 1211 |
+
"rewards/margins": 0.049835216253995895,
|
| 1212 |
+
"rewards/rejected": -0.25038737058639526,
|
| 1213 |
+
"step": 850
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.68,
|
| 1217 |
+
"learning_rate": 1.4405195680840357e-06,
|
| 1218 |
+
"logits/chosen": 0.5690888166427612,
|
| 1219 |
+
"logits/rejected": 0.6206023097038269,
|
| 1220 |
+
"logps/chosen": -282.1586608886719,
|
| 1221 |
+
"logps/rejected": -248.159423828125,
|
| 1222 |
+
"loss": 0.0371,
|
| 1223 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1224 |
+
"rewards/chosen": -0.17809875309467316,
|
| 1225 |
+
"rewards/margins": 0.05816931277513504,
|
| 1226 |
+
"rewards/rejected": -0.2362680435180664,
|
| 1227 |
+
"step": 860
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.68,
|
| 1231 |
+
"learning_rate": 1.378796718871252e-06,
|
| 1232 |
+
"logits/chosen": 0.5121451616287231,
|
| 1233 |
+
"logits/rejected": 0.6095593571662903,
|
| 1234 |
+
"logps/chosen": -339.46673583984375,
|
| 1235 |
+
"logps/rejected": -275.1758728027344,
|
| 1236 |
+
"loss": 0.0394,
|
| 1237 |
+
"rewards/accuracies": 0.625,
|
| 1238 |
+
"rewards/chosen": -0.18696454167366028,
|
| 1239 |
+
"rewards/margins": 0.05620839074254036,
|
| 1240 |
+
"rewards/rejected": -0.24317291378974915,
|
| 1241 |
+
"step": 870
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.69,
|
| 1245 |
+
"learning_rate": 1.3179178763436302e-06,
|
| 1246 |
+
"logits/chosen": 0.48259830474853516,
|
| 1247 |
+
"logits/rejected": 0.6230058073997498,
|
| 1248 |
+
"logps/chosen": -290.6861267089844,
|
| 1249 |
+
"logps/rejected": -267.58428955078125,
|
| 1250 |
+
"loss": 0.0312,
|
| 1251 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1252 |
+
"rewards/chosen": -0.1768282651901245,
|
| 1253 |
+
"rewards/margins": 0.061559922993183136,
|
| 1254 |
+
"rewards/rejected": -0.23838821053504944,
|
| 1255 |
+
"step": 880
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.7,
|
| 1259 |
+
"learning_rate": 1.2579288681855364e-06,
|
| 1260 |
+
"logits/chosen": 0.6349326968193054,
|
| 1261 |
+
"logits/rejected": 0.5690282583236694,
|
| 1262 |
+
"logps/chosen": -283.8800964355469,
|
| 1263 |
+
"logps/rejected": -267.09417724609375,
|
| 1264 |
+
"loss": 0.0435,
|
| 1265 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1266 |
+
"rewards/chosen": -0.19100067019462585,
|
| 1267 |
+
"rewards/margins": 0.05224493891000748,
|
| 1268 |
+
"rewards/rejected": -0.24324557185173035,
|
| 1269 |
+
"step": 890
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.71,
|
| 1273 |
+
"learning_rate": 1.1988748522419163e-06,
|
| 1274 |
+
"logits/chosen": 0.5791734457015991,
|
| 1275 |
+
"logits/rejected": 0.6113063097000122,
|
| 1276 |
+
"logps/chosen": -307.1993408203125,
|
| 1277 |
+
"logps/rejected": -278.4551696777344,
|
| 1278 |
+
"loss": 0.0402,
|
| 1279 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1280 |
+
"rewards/chosen": -0.18477237224578857,
|
| 1281 |
+
"rewards/margins": 0.057056255638599396,
|
| 1282 |
+
"rewards/rejected": -0.24182860553264618,
|
| 1283 |
+
"step": 900
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.71,
|
| 1287 |
+
"learning_rate": 1.1408002825248842e-06,
|
| 1288 |
+
"logits/chosen": 0.5439066886901855,
|
| 1289 |
+
"logits/rejected": 0.5941084027290344,
|
| 1290 |
+
"logps/chosen": -329.396240234375,
|
| 1291 |
+
"logps/rejected": -295.2839660644531,
|
| 1292 |
+
"loss": 0.0331,
|
| 1293 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1294 |
+
"rewards/chosen": -0.1919744461774826,
|
| 1295 |
+
"rewards/margins": 0.052353404462337494,
|
| 1296 |
+
"rewards/rejected": -0.2443278580904007,
|
| 1297 |
+
"step": 910
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.72,
|
| 1301 |
+
"learning_rate": 1.0837488757501369e-06,
|
| 1302 |
+
"logits/chosen": 0.5491483211517334,
|
| 1303 |
+
"logits/rejected": 0.6100614666938782,
|
| 1304 |
+
"logps/chosen": -292.9952392578125,
|
| 1305 |
+
"logps/rejected": -277.0441589355469,
|
| 1306 |
+
"loss": 0.0356,
|
| 1307 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1308 |
+
"rewards/chosen": -0.17984095215797424,
|
| 1309 |
+
"rewards/margins": 0.042489904910326004,
|
| 1310 |
+
"rewards/rejected": -0.22233083844184875,
|
| 1311 |
+
"step": 920
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.73,
|
| 1315 |
+
"learning_rate": 1.027763578428379e-06,
|
| 1316 |
+
"logits/chosen": 0.6163350939750671,
|
| 1317 |
+
"logits/rejected": 0.6365456581115723,
|
| 1318 |
+
"logps/chosen": -264.69342041015625,
|
| 1319 |
+
"logps/rejected": -243.10806274414062,
|
| 1320 |
+
"loss": 0.0417,
|
| 1321 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1322 |
+
"rewards/chosen": -0.19993123412132263,
|
| 1323 |
+
"rewards/margins": 0.046492286026477814,
|
| 1324 |
+
"rewards/rejected": -0.24642352759838104,
|
| 1325 |
+
"step": 930
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.74,
|
| 1329 |
+
"learning_rate": 9.728865345365379e-07,
|
| 1330 |
+
"logits/chosen": 0.5641825199127197,
|
| 1331 |
+
"logits/rejected": 0.606396496295929,
|
| 1332 |
+
"logps/chosen": -331.7998962402344,
|
| 1333 |
+
"logps/rejected": -272.21807861328125,
|
| 1334 |
+
"loss": 0.0449,
|
| 1335 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1336 |
+
"rewards/chosen": -0.19182895123958588,
|
| 1337 |
+
"rewards/margins": 0.047703493386507034,
|
| 1338 |
+
"rewards/rejected": -0.239532470703125,
|
| 1339 |
+
"step": 940
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.75,
|
| 1343 |
+
"learning_rate": 9.191590537930975e-07,
|
| 1344 |
+
"logits/chosen": 0.5647382736206055,
|
| 1345 |
+
"logits/rejected": 0.5931411981582642,
|
| 1346 |
+
"logps/chosen": -301.1795959472656,
|
| 1347 |
+
"logps/rejected": -249.8610076904297,
|
| 1348 |
+
"loss": 0.0382,
|
| 1349 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1350 |
+
"rewards/chosen": -0.1782718151807785,
|
| 1351 |
+
"rewards/margins": 0.065487340092659,
|
| 1352 |
+
"rewards/rejected": -0.2437591254711151,
|
| 1353 |
+
"step": 950
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.75,
|
| 1357 |
+
"learning_rate": 8.666215805614373e-07,
|
| 1358 |
+
"logits/chosen": 0.5855926871299744,
|
| 1359 |
+
"logits/rejected": 0.6292095184326172,
|
| 1360 |
+
"logps/chosen": -314.7355041503906,
|
| 1361 |
+
"logps/rejected": -275.4202575683594,
|
| 1362 |
+
"loss": 0.0394,
|
| 1363 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1364 |
+
"rewards/chosen": -0.20517146587371826,
|
| 1365 |
+
"rewards/margins": 0.041300006210803986,
|
| 1366 |
+
"rewards/rejected": -0.24647144973278046,
|
| 1367 |
+
"step": 960
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.76,
|
| 1371 |
+
"learning_rate": 8.153136634045844e-07,
|
| 1372 |
+
"logits/chosen": 0.5548075437545776,
|
| 1373 |
+
"logits/rejected": 0.6082533001899719,
|
| 1374 |
+
"logps/chosen": -298.9161071777344,
|
| 1375 |
+
"logps/rejected": -263.01263427734375,
|
| 1376 |
+
"loss": 0.0347,
|
| 1377 |
+
"rewards/accuracies": 0.625,
|
| 1378 |
+
"rewards/chosen": -0.20182470977306366,
|
| 1379 |
+
"rewards/margins": 0.05161082744598389,
|
| 1380 |
+
"rewards/rejected": -0.25343555212020874,
|
| 1381 |
+
"step": 970
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.77,
|
| 1385 |
+
"learning_rate": 7.652739253142915e-07,
|
| 1386 |
+
"logits/chosen": 0.6577678322792053,
|
| 1387 |
+
"logits/rejected": 0.7113202810287476,
|
| 1388 |
+
"logps/chosen": -305.44976806640625,
|
| 1389 |
+
"logps/rejected": -270.2418212890625,
|
| 1390 |
+
"loss": 0.0368,
|
| 1391 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1392 |
+
"rewards/chosen": -0.18101605772972107,
|
| 1393 |
+
"rewards/margins": 0.06050257757306099,
|
| 1394 |
+
"rewards/rejected": -0.24151864647865295,
|
| 1395 |
+
"step": 980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.78,
|
| 1399 |
+
"learning_rate": 7.165400346368648e-07,
|
| 1400 |
+
"logits/chosen": 0.6242294311523438,
|
| 1401 |
+
"logits/rejected": 0.6780456304550171,
|
| 1402 |
+
"logps/chosen": -332.4653625488281,
|
| 1403 |
+
"logps/rejected": -270.494384765625,
|
| 1404 |
+
"loss": 0.0428,
|
| 1405 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1406 |
+
"rewards/chosen": -0.18956415355205536,
|
| 1407 |
+
"rewards/margins": 0.056781280785799026,
|
| 1408 |
+
"rewards/rejected": -0.24634544551372528,
|
| 1409 |
+
"step": 990
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.79,
|
| 1413 |
+
"learning_rate": 6.691486767176092e-07,
|
| 1414 |
+
"logits/chosen": 0.5437681078910828,
|
| 1415 |
+
"logits/rejected": 0.5767791867256165,
|
| 1416 |
+
"logps/chosen": -271.98846435546875,
|
| 1417 |
+
"logps/rejected": -275.7591247558594,
|
| 1418 |
+
"loss": 0.042,
|
| 1419 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1420 |
+
"rewards/chosen": -0.15761049091815948,
|
| 1421 |
+
"rewards/margins": 0.060720719397068024,
|
| 1422 |
+
"rewards/rejected": -0.2183312177658081,
|
| 1423 |
+
"step": 1000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.79,
|
| 1427 |
+
"learning_rate": 6.231355262852529e-07,
|
| 1428 |
+
"logits/chosen": 0.5681263208389282,
|
| 1429 |
+
"logits/rejected": 0.5934298038482666,
|
| 1430 |
+
"logps/chosen": -362.6823425292969,
|
| 1431 |
+
"logps/rejected": -315.0286560058594,
|
| 1432 |
+
"loss": 0.0427,
|
| 1433 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 1434 |
+
"rewards/chosen": -0.19303514063358307,
|
| 1435 |
+
"rewards/margins": 0.03555145114660263,
|
| 1436 |
+
"rewards/rejected": -0.2285865992307663,
|
| 1437 |
+
"step": 1010
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.8,
|
| 1441 |
+
"learning_rate": 5.785352205971275e-07,
|
| 1442 |
+
"logits/chosen": 0.5157848596572876,
|
| 1443 |
+
"logits/rejected": 0.6102535724639893,
|
| 1444 |
+
"logps/chosen": -334.3918151855469,
|
| 1445 |
+
"logps/rejected": -284.1390075683594,
|
| 1446 |
+
"loss": 0.0442,
|
| 1447 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1448 |
+
"rewards/chosen": -0.1846066266298294,
|
| 1449 |
+
"rewards/margins": 0.06148224323987961,
|
| 1450 |
+
"rewards/rejected": -0.24608883261680603,
|
| 1451 |
+
"step": 1020
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.81,
|
| 1455 |
+
"learning_rate": 5.353813333653287e-07,
|
| 1456 |
+
"logits/chosen": 0.539936363697052,
|
| 1457 |
+
"logits/rejected": 0.6743307709693909,
|
| 1458 |
+
"logps/chosen": -305.96270751953125,
|
| 1459 |
+
"logps/rejected": -280.2798767089844,
|
| 1460 |
+
"loss": 0.0339,
|
| 1461 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1462 |
+
"rewards/chosen": -0.1911364048719406,
|
| 1463 |
+
"rewards/margins": 0.05630878359079361,
|
| 1464 |
+
"rewards/rejected": -0.24744519591331482,
|
| 1465 |
+
"step": 1030
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.82,
|
| 1469 |
+
"learning_rate": 4.937063494834774e-07,
|
| 1470 |
+
"logits/chosen": 0.5440152883529663,
|
| 1471 |
+
"logits/rejected": 0.5379213094711304,
|
| 1472 |
+
"logps/chosen": -341.1717224121094,
|
| 1473 |
+
"logps/rejected": -311.44549560546875,
|
| 1474 |
+
"loss": 0.0393,
|
| 1475 |
+
"rewards/accuracies": 0.574999988079071,
|
| 1476 |
+
"rewards/chosen": -0.18748323619365692,
|
| 1477 |
+
"rewards/margins": 0.0409691259264946,
|
| 1478 |
+
"rewards/rejected": -0.22845235466957092,
|
| 1479 |
+
"step": 1040
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.82,
|
| 1483 |
+
"learning_rate": 4.5354164057310857e-07,
|
| 1484 |
+
"logits/chosen": 0.5548704862594604,
|
| 1485 |
+
"logits/rejected": 0.689899742603302,
|
| 1486 |
+
"logps/chosen": -312.20428466796875,
|
| 1487 |
+
"logps/rejected": -294.8197021484375,
|
| 1488 |
+
"loss": 0.0314,
|
| 1489 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1490 |
+
"rewards/chosen": -0.18130271136760712,
|
| 1491 |
+
"rewards/margins": 0.04764419421553612,
|
| 1492 |
+
"rewards/rejected": -0.22894692420959473,
|
| 1493 |
+
"step": 1050
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.83,
|
| 1497 |
+
"learning_rate": 4.1491744136810066e-07,
|
| 1498 |
+
"logits/chosen": 0.4567073881626129,
|
| 1499 |
+
"logits/rejected": 0.574688732624054,
|
| 1500 |
+
"logps/chosen": -294.9923400878906,
|
| 1501 |
+
"logps/rejected": -275.3056945800781,
|
| 1502 |
+
"loss": 0.0326,
|
| 1503 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1504 |
+
"rewards/chosen": -0.1727767288684845,
|
| 1505 |
+
"rewards/margins": 0.04382780194282532,
|
| 1506 |
+
"rewards/rejected": -0.216604545712471,
|
| 1507 |
+
"step": 1060
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.84,
|
| 1511 |
+
"learning_rate": 3.7786282695491313e-07,
|
| 1512 |
+
"logits/chosen": 0.6358944773674011,
|
| 1513 |
+
"logits/rejected": 0.6744691133499146,
|
| 1514 |
+
"logps/chosen": -316.3371276855469,
|
| 1515 |
+
"logps/rejected": -270.2391662597656,
|
| 1516 |
+
"loss": 0.037,
|
| 1517 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1518 |
+
"rewards/chosen": -0.1606210619211197,
|
| 1519 |
+
"rewards/margins": 0.05917732045054436,
|
| 1520 |
+
"rewards/rejected": -0.21979837119579315,
|
| 1521 |
+
"step": 1070
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.85,
|
| 1525 |
+
"learning_rate": 3.4240569088577564e-07,
|
| 1526 |
+
"logits/chosen": 0.5884599089622498,
|
| 1527 |
+
"logits/rejected": 0.6169819831848145,
|
| 1528 |
+
"logps/chosen": -307.93701171875,
|
| 1529 |
+
"logps/rejected": -279.9734191894531,
|
| 1530 |
+
"loss": 0.0364,
|
| 1531 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1532 |
+
"rewards/chosen": -0.17860962450504303,
|
| 1533 |
+
"rewards/margins": 0.05303224176168442,
|
| 1534 |
+
"rewards/rejected": -0.23164185881614685,
|
| 1535 |
+
"step": 1080
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.86,
|
| 1539 |
+
"learning_rate": 3.0857272418129136e-07,
|
| 1540 |
+
"logits/chosen": 0.47011032700538635,
|
| 1541 |
+
"logits/rejected": 0.595490574836731,
|
| 1542 |
+
"logps/chosen": -283.87786865234375,
|
| 1543 |
+
"logps/rejected": -263.67724609375,
|
| 1544 |
+
"loss": 0.0334,
|
| 1545 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 1546 |
+
"rewards/chosen": -0.17268426716327667,
|
| 1547 |
+
"rewards/margins": 0.049144428223371506,
|
| 1548 |
+
"rewards/rejected": -0.22182869911193848,
|
| 1549 |
+
"step": 1090
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.86,
|
| 1553 |
+
"learning_rate": 2.7638939523827956e-07,
|
| 1554 |
+
"logits/chosen": 0.6203007698059082,
|
| 1555 |
+
"logits/rejected": 0.6173182129859924,
|
| 1556 |
+
"logps/chosen": -283.423828125,
|
| 1557 |
+
"logps/rejected": -276.0241394042969,
|
| 1558 |
+
"loss": 0.0371,
|
| 1559 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1560 |
+
"rewards/chosen": -0.17366403341293335,
|
| 1561 |
+
"rewards/margins": 0.06443198025226593,
|
| 1562 |
+
"rewards/rejected": -0.23809599876403809,
|
| 1563 |
+
"step": 1100
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.87,
|
| 1567 |
+
"learning_rate": 2.4587993065795983e-07,
|
| 1568 |
+
"logits/chosen": 0.6205240488052368,
|
| 1569 |
+
"logits/rejected": 0.6329373121261597,
|
| 1570 |
+
"logps/chosen": -326.6821594238281,
|
| 1571 |
+
"logps/rejected": -297.9998474121094,
|
| 1572 |
+
"loss": 0.0429,
|
| 1573 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1574 |
+
"rewards/chosen": -0.18457256257534027,
|
| 1575 |
+
"rewards/margins": 0.05959055945277214,
|
| 1576 |
+
"rewards/rejected": -0.2441631555557251,
|
| 1577 |
+
"step": 1110
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.88,
|
| 1581 |
+
"learning_rate": 2.170672970089291e-07,
|
| 1582 |
+
"logits/chosen": 0.5973344445228577,
|
| 1583 |
+
"logits/rejected": 0.6125074028968811,
|
| 1584 |
+
"logps/chosen": -318.9927978515625,
|
| 1585 |
+
"logps/rejected": -277.43475341796875,
|
| 1586 |
+
"loss": 0.0384,
|
| 1587 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 1588 |
+
"rewards/chosen": -0.17308750748634338,
|
| 1589 |
+
"rewards/margins": 0.03716738894581795,
|
| 1590 |
+
"rewards/rejected": -0.21025490760803223,
|
| 1591 |
+
"step": 1120
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.89,
|
| 1595 |
+
"learning_rate": 1.8997318353864673e-07,
|
| 1596 |
+
"logits/chosen": 0.5647025108337402,
|
| 1597 |
+
"logits/rejected": 0.6136811971664429,
|
| 1598 |
+
"logps/chosen": -306.97698974609375,
|
| 1599 |
+
"logps/rejected": -276.1070251464844,
|
| 1600 |
+
"loss": 0.0394,
|
| 1601 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1602 |
+
"rewards/chosen": -0.1707993894815445,
|
| 1603 |
+
"rewards/margins": 0.05791207030415535,
|
| 1604 |
+
"rewards/rejected": -0.22871145606040955,
|
| 1605 |
+
"step": 1130
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.89,
|
| 1609 |
+
"learning_rate": 1.6461798584644944e-07,
|
| 1610 |
+
"logits/chosen": 0.6306375861167908,
|
| 1611 |
+
"logits/rejected": 0.677390456199646,
|
| 1612 |
+
"logps/chosen": -310.4776306152344,
|
| 1613 |
+
"logps/rejected": -283.60137939453125,
|
| 1614 |
+
"loss": 0.0358,
|
| 1615 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1616 |
+
"rewards/chosen": -0.17375603318214417,
|
| 1617 |
+
"rewards/margins": 0.06438947468996048,
|
| 1618 |
+
"rewards/rejected": -0.23814550042152405,
|
| 1619 |
+
"step": 1140
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.9,
|
| 1623 |
+
"learning_rate": 1.4102079053038454e-07,
|
| 1624 |
+
"logits/chosen": 0.5880599021911621,
|
| 1625 |
+
"logits/rejected": 0.5578813552856445,
|
| 1626 |
+
"logps/chosen": -300.5694580078125,
|
| 1627 |
+
"logps/rejected": -258.61004638671875,
|
| 1628 |
+
"loss": 0.0377,
|
| 1629 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 1630 |
+
"rewards/chosen": -0.1807607263326645,
|
| 1631 |
+
"rewards/margins": 0.037232767790555954,
|
| 1632 |
+
"rewards/rejected": -0.21799346804618835,
|
| 1633 |
+
"step": 1150
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.91,
|
| 1637 |
+
"learning_rate": 1.1919936081941585e-07,
|
| 1638 |
+
"logits/chosen": 0.5770727396011353,
|
| 1639 |
+
"logits/rejected": 0.6810525059700012,
|
| 1640 |
+
"logps/chosen": -277.33453369140625,
|
| 1641 |
+
"logps/rejected": -288.65362548828125,
|
| 1642 |
+
"loss": 0.036,
|
| 1643 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1644 |
+
"rewards/chosen": -0.1911756694316864,
|
| 1645 |
+
"rewards/margins": 0.05452916771173477,
|
| 1646 |
+
"rewards/rejected": -0.24570484459400177,
|
| 1647 |
+
"step": 1160
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.92,
|
| 1651 |
+
"learning_rate": 9.917012320182245e-08,
|
| 1652 |
+
"logits/chosen": 0.5721954703330994,
|
| 1653 |
+
"logits/rejected": 0.6260513067245483,
|
| 1654 |
+
"logps/chosen": -284.9855651855469,
|
| 1655 |
+
"logps/rejected": -293.6813659667969,
|
| 1656 |
+
"loss": 0.0456,
|
| 1657 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1658 |
+
"rewards/chosen": -0.1743406057357788,
|
| 1659 |
+
"rewards/margins": 0.05407222360372543,
|
| 1660 |
+
"rewards/rejected": -0.22841282188892365,
|
| 1661 |
+
"step": 1170
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.93,
|
| 1665 |
+
"learning_rate": 8.094815505985315e-08,
|
| 1666 |
+
"logits/chosen": 0.6424626111984253,
|
| 1667 |
+
"logits/rejected": 0.5862604975700378,
|
| 1668 |
+
"logps/chosen": -305.09735107421875,
|
| 1669 |
+
"logps/rejected": -268.40179443359375,
|
| 1670 |
+
"loss": 0.0364,
|
| 1671 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1672 |
+
"rewards/chosen": -0.17467457056045532,
|
| 1673 |
+
"rewards/margins": 0.03977496176958084,
|
| 1674 |
+
"rewards/rejected": -0.21444952487945557,
|
| 1675 |
+
"step": 1180
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.93,
|
| 1679 |
+
"learning_rate": 6.454717331994542e-08,
|
| 1680 |
+
"logits/chosen": 0.6076738238334656,
|
| 1681 |
+
"logits/rejected": 0.5825602412223816,
|
| 1682 |
+
"logps/chosen": -316.4405212402344,
|
| 1683 |
+
"logps/rejected": -298.938232421875,
|
| 1684 |
+
"loss": 0.0406,
|
| 1685 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1686 |
+
"rewards/chosen": -0.18210789561271667,
|
| 1687 |
+
"rewards/margins": 0.03575897961854935,
|
| 1688 |
+
"rewards/rejected": -0.2178668975830078,
|
| 1689 |
+
"step": 1190
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.94,
|
| 1693 |
+
"learning_rate": 4.9979524127052595e-08,
|
| 1694 |
+
"logits/chosen": 0.573727011680603,
|
| 1695 |
+
"logits/rejected": 0.6119558811187744,
|
| 1696 |
+
"logps/chosen": -308.22174072265625,
|
| 1697 |
+
"logps/rejected": -276.9308776855469,
|
| 1698 |
+
"loss": 0.0392,
|
| 1699 |
+
"rewards/accuracies": 0.53125,
|
| 1700 |
+
"rewards/chosen": -0.16624239087104797,
|
| 1701 |
+
"rewards/margins": 0.043992962688207626,
|
| 1702 |
+
"rewards/rejected": -0.2102353572845459,
|
| 1703 |
+
"step": 1200
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.95,
|
| 1707 |
+
"learning_rate": 3.725617355085476e-08,
|
| 1708 |
+
"logits/chosen": 0.5791876316070557,
|
| 1709 |
+
"logits/rejected": 0.6267887353897095,
|
| 1710 |
+
"logps/chosen": -314.5113830566406,
|
| 1711 |
+
"logps/rejected": -280.70892333984375,
|
| 1712 |
+
"loss": 0.039,
|
| 1713 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1714 |
+
"rewards/chosen": -0.20197847485542297,
|
| 1715 |
+
"rewards/margins": 0.04940859228372574,
|
| 1716 |
+
"rewards/rejected": -0.2513870596885681,
|
| 1717 |
+
"step": 1210
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.96,
|
| 1721 |
+
"learning_rate": 2.63866993308437e-08,
|
| 1722 |
+
"logits/chosen": 0.5729714632034302,
|
| 1723 |
+
"logits/rejected": 0.6838079690933228,
|
| 1724 |
+
"logps/chosen": -325.7432861328125,
|
| 1725 |
+
"logps/rejected": -262.1849060058594,
|
| 1726 |
+
"loss": 0.0383,
|
| 1727 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1728 |
+
"rewards/chosen": -0.16999685764312744,
|
| 1729 |
+
"rewards/margins": 0.05414390563964844,
|
| 1730 |
+
"rewards/rejected": -0.22414080798625946,
|
| 1731 |
+
"step": 1220
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.97,
|
| 1735 |
+
"learning_rate": 1.737928366650099e-08,
|
| 1736 |
+
"logits/chosen": 0.537979006767273,
|
| 1737 |
+
"logits/rejected": 0.6233702898025513,
|
| 1738 |
+
"logps/chosen": -347.04754638671875,
|
| 1739 |
+
"logps/rejected": -299.84564208984375,
|
| 1740 |
+
"loss": 0.0378,
|
| 1741 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1742 |
+
"rewards/chosen": -0.1912025660276413,
|
| 1743 |
+
"rewards/margins": 0.057091712951660156,
|
| 1744 |
+
"rewards/rejected": -0.24829427897930145,
|
| 1745 |
+
"step": 1230
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.97,
|
| 1749 |
+
"learning_rate": 1.0240707057995735e-08,
|
| 1750 |
+
"logits/chosen": 0.5721354484558105,
|
| 1751 |
+
"logits/rejected": 0.6452512741088867,
|
| 1752 |
+
"logps/chosen": -291.0641784667969,
|
| 1753 |
+
"logps/rejected": -280.39947509765625,
|
| 1754 |
+
"loss": 0.0384,
|
| 1755 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1756 |
+
"rewards/chosen": -0.18037080764770508,
|
| 1757 |
+
"rewards/margins": 0.05573723465204239,
|
| 1758 |
+
"rewards/rejected": -0.23610803484916687,
|
| 1759 |
+
"step": 1240
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.98,
|
| 1763 |
+
"learning_rate": 4.976343202034717e-09,
|
| 1764 |
+
"logits/chosen": 0.5639461278915405,
|
| 1765 |
+
"logits/rejected": 0.6287204027175903,
|
| 1766 |
+
"logps/chosen": -303.62811279296875,
|
| 1767 |
+
"logps/rejected": -280.3854675292969,
|
| 1768 |
+
"loss": 0.0316,
|
| 1769 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1770 |
+
"rewards/chosen": -0.18784946203231812,
|
| 1771 |
+
"rewards/margins": 0.043084483593702316,
|
| 1772 |
+
"rewards/rejected": -0.23093394935131073,
|
| 1773 |
+
"step": 1250
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.99,
|
| 1777 |
+
"learning_rate": 1.5901549467139953e-09,
|
| 1778 |
+
"logits/chosen": 0.4898532032966614,
|
| 1779 |
+
"logits/rejected": 0.5764004588127136,
|
| 1780 |
+
"logps/chosen": -281.58013916015625,
|
| 1781 |
+
"logps/rejected": -252.1945343017578,
|
| 1782 |
+
"loss": 0.0345,
|
| 1783 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1784 |
+
"rewards/chosen": -0.1634509116411209,
|
| 1785 |
+
"rewards/margins": 0.06314502656459808,
|
| 1786 |
+
"rewards/rejected": -0.2265959233045578,
|
| 1787 |
+
"step": 1260
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 1.0,
|
| 1791 |
+
"learning_rate": 8.469130840960127e-11,
|
| 1792 |
+
"logits/chosen": 0.6247270107269287,
|
| 1793 |
+
"logits/rejected": 0.6422919034957886,
|
| 1794 |
+
"logps/chosen": -290.09881591796875,
|
| 1795 |
+
"logps/rejected": -274.88592529296875,
|
| 1796 |
+
"loss": 0.0406,
|
| 1797 |
+
"rewards/accuracies": 0.59375,
|
| 1798 |
+
"rewards/chosen": -0.16841378808021545,
|
| 1799 |
+
"rewards/margins": 0.0537291057407856,
|
| 1800 |
+
"rewards/rejected": -0.22214286029338837,
|
| 1801 |
+
"step": 1270
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 1.0,
|
| 1805 |
+
"step": 1273,
|
| 1806 |
+
"total_flos": 0.0,
|
| 1807 |
+
"train_loss": 0.04122506890033327,
|
| 1808 |
+
"train_runtime": 12850.1249,
|
| 1809 |
+
"train_samples_per_second": 4.758,
|
| 1810 |
+
"train_steps_per_second": 0.099
|
| 1811 |
+
}
|
| 1812 |
+
],
|
| 1813 |
+
"logging_steps": 10,
|
| 1814 |
+
"max_steps": 1273,
|
| 1815 |
+
"num_input_tokens_seen": 0,
|
| 1816 |
+
"num_train_epochs": 1,
|
| 1817 |
+
"save_steps": 100,
|
| 1818 |
+
"total_flos": 0.0,
|
| 1819 |
+
"train_batch_size": 4,
|
| 1820 |
+
"trial_name": null,
|
| 1821 |
+
"trial_params": null
|
| 1822 |
+
}
|