Commit
•
3829c7f
1
Parent(s):
c6f9ff2
Model save
Browse files- README.md +74 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +624 -0
README.md
ADDED
|
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
tags:
|
| 4 |
+
- trl
|
| 5 |
+
- dpo
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
base_model: NbAiLab/nb-gpt-j-6B-v2
|
| 8 |
+
model-index:
|
| 9 |
+
- name: aftonposten-6b-align-scan
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# aftonposten-6b-align-scan
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [NbAiLab/nb-gpt-j-6B-v2](https://huggingface.co/NbAiLab/nb-gpt-j-6B-v2) on an unknown dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.6931
|
| 21 |
+
- Rewards/chosen: 0.0001
|
| 22 |
+
- Rewards/rejected: 0.0001
|
| 23 |
+
- Rewards/accuracies: 0.4958
|
| 24 |
+
- Rewards/margins: 0.0000
|
| 25 |
+
- Logps/rejected: -37.5088
|
| 26 |
+
- Logps/chosen: -34.0255
|
| 27 |
+
- Logits/rejected: -2.2394
|
| 28 |
+
- Logits/chosen: -2.2443
|
| 29 |
+
|
| 30 |
+
## Model description
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Intended uses & limitations
|
| 35 |
+
|
| 36 |
+
More information needed
|
| 37 |
+
|
| 38 |
+
## Training and evaluation data
|
| 39 |
+
|
| 40 |
+
More information needed
|
| 41 |
+
|
| 42 |
+
## Training procedure
|
| 43 |
+
|
| 44 |
+
### Training hyperparameters
|
| 45 |
+
|
| 46 |
+
The following hyperparameters were used during training:
|
| 47 |
+
- learning_rate: 5e-07
|
| 48 |
+
- train_batch_size: 4
|
| 49 |
+
- eval_batch_size: 8
|
| 50 |
+
- seed: 42
|
| 51 |
+
- distributed_type: multi-GPU
|
| 52 |
+
- gradient_accumulation_steps: 2
|
| 53 |
+
- total_train_batch_size: 8
|
| 54 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 55 |
+
- lr_scheduler_type: cosine
|
| 56 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 57 |
+
- num_epochs: 1
|
| 58 |
+
|
| 59 |
+
### Training results
|
| 60 |
+
|
| 61 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 62 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 63 |
+
| 0.6932 | 0.26 | 100 | 0.6932 | 0.0001 | 0.0001 | 0.4605 | -0.0001 | -37.5038 | -34.0293 | -2.2393 | -2.2442 |
|
| 64 |
+
| 0.693 | 0.52 | 200 | 0.6932 | 0.0001 | 0.0002 | 0.4992 | -0.0001 | -37.5001 | -34.0271 | -2.2390 | -2.2439 |
|
| 65 |
+
| 0.693 | 0.78 | 300 | 0.6931 | 0.0001 | 0.0001 | 0.4958 | 0.0000 | -37.5088 | -34.0255 | -2.2394 | -2.2443 |
|
| 66 |
+
|
| 67 |
+
|
| 68 |
+
### Framework versions
|
| 69 |
+
|
| 70 |
+
- PEFT 0.8.2
|
| 71 |
+
- Transformers 4.37.2
|
| 72 |
+
- Pytorch 2.1.2+cu121
|
| 73 |
+
- Datasets 2.17.0
|
| 74 |
+
- Tokenizers 0.15.1
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 176183216
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81a12530c8d521b3699b0a5fe263f2b51bc1b52ef24c389c68c92ba8e08e4da3
|
| 3 |
size 176183216
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.6930816006350827,
|
| 4 |
+
"train_runtime": 3252.6347,
|
| 5 |
+
"train_samples": 3079,
|
| 6 |
+
"train_samples_per_second": 0.947,
|
| 7 |
+
"train_steps_per_second": 0.118
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.6930816006350827,
|
| 4 |
+
"train_runtime": 3252.6347,
|
| 5 |
+
"train_samples": 3079,
|
| 6 |
+
"train_samples_per_second": 0.947,
|
| 7 |
+
"train_steps_per_second": 0.118
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,624 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 385,
|
| 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": 1.282051282051282e-08,
|
| 14 |
+
"logits/chosen": -1.7278180122375488,
|
| 15 |
+
"logits/rejected": -1.7377450466156006,
|
| 16 |
+
"logps/chosen": -29.553977966308594,
|
| 17 |
+
"logps/rejected": -42.813133239746094,
|
| 18 |
+
"loss": 0.6931,
|
| 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": 0.03,
|
| 27 |
+
"learning_rate": 1.2820512820512818e-07,
|
| 28 |
+
"logits/chosen": -1.86637282371521,
|
| 29 |
+
"logits/rejected": -1.8706889152526855,
|
| 30 |
+
"logps/chosen": -36.99605941772461,
|
| 31 |
+
"logps/rejected": -33.66212844848633,
|
| 32 |
+
"loss": 0.6929,
|
| 33 |
+
"rewards/accuracies": 0.4861111044883728,
|
| 34 |
+
"rewards/chosen": 0.00010540737275732681,
|
| 35 |
+
"rewards/margins": 0.00042594841215759516,
|
| 36 |
+
"rewards/rejected": -0.0003205410612281412,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.05,
|
| 41 |
+
"learning_rate": 2.5641025641025636e-07,
|
| 42 |
+
"logits/chosen": -1.9977174997329712,
|
| 43 |
+
"logits/rejected": -2.0003600120544434,
|
| 44 |
+
"logps/chosen": -29.638538360595703,
|
| 45 |
+
"logps/rejected": -29.072032928466797,
|
| 46 |
+
"loss": 0.6931,
|
| 47 |
+
"rewards/accuracies": 0.4375,
|
| 48 |
+
"rewards/chosen": 3.6563815228873864e-05,
|
| 49 |
+
"rewards/margins": 0.00012243764649610966,
|
| 50 |
+
"rewards/rejected": -8.587385673308745e-05,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.08,
|
| 55 |
+
"learning_rate": 3.8461538461538463e-07,
|
| 56 |
+
"logits/chosen": -1.91976797580719,
|
| 57 |
+
"logits/rejected": -1.917059302330017,
|
| 58 |
+
"logps/chosen": -31.427032470703125,
|
| 59 |
+
"logps/rejected": -33.21014404296875,
|
| 60 |
+
"loss": 0.6932,
|
| 61 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 62 |
+
"rewards/chosen": -0.00010937436309177428,
|
| 63 |
+
"rewards/margins": -0.000192504987353459,
|
| 64 |
+
"rewards/rejected": 8.313067519338802e-05,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.1,
|
| 69 |
+
"learning_rate": 4.999896948438433e-07,
|
| 70 |
+
"logits/chosen": -2.0164756774902344,
|
| 71 |
+
"logits/rejected": -2.0077216625213623,
|
| 72 |
+
"logps/chosen": -32.58618927001953,
|
| 73 |
+
"logps/rejected": -32.47720718383789,
|
| 74 |
+
"loss": 0.6934,
|
| 75 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 76 |
+
"rewards/chosen": -9.37797303777188e-05,
|
| 77 |
+
"rewards/margins": -0.00044843609794043005,
|
| 78 |
+
"rewards/rejected": 0.0003546563093550503,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.13,
|
| 83 |
+
"learning_rate": 4.987541037542186e-07,
|
| 84 |
+
"logits/chosen": -1.864516019821167,
|
| 85 |
+
"logits/rejected": -1.8537566661834717,
|
| 86 |
+
"logps/chosen": -33.572547912597656,
|
| 87 |
+
"logps/rejected": -35.45746994018555,
|
| 88 |
+
"loss": 0.6932,
|
| 89 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 90 |
+
"rewards/chosen": -0.00015066245396155864,
|
| 91 |
+
"rewards/margins": -9.940571908373386e-05,
|
| 92 |
+
"rewards/rejected": -5.125668758410029e-05,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.16,
|
| 97 |
+
"learning_rate": 4.954691471941118e-07,
|
| 98 |
+
"logits/chosen": -1.9447959661483765,
|
| 99 |
+
"logits/rejected": -1.946738839149475,
|
| 100 |
+
"logps/chosen": -32.576927185058594,
|
| 101 |
+
"logps/rejected": -33.18446350097656,
|
| 102 |
+
"loss": 0.693,
|
| 103 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 104 |
+
"rewards/chosen": 0.0002326248650206253,
|
| 105 |
+
"rewards/margins": 0.0003634841996245086,
|
| 106 |
+
"rewards/rejected": -0.00013085929094813764,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.18,
|
| 111 |
+
"learning_rate": 4.901618883413548e-07,
|
| 112 |
+
"logits/chosen": -2.079601764678955,
|
| 113 |
+
"logits/rejected": -2.0845787525177,
|
| 114 |
+
"logps/chosen": -33.98583221435547,
|
| 115 |
+
"logps/rejected": -36.567405700683594,
|
| 116 |
+
"loss": 0.6931,
|
| 117 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 118 |
+
"rewards/chosen": 2.6532879928709008e-05,
|
| 119 |
+
"rewards/margins": 1.1868818546645343e-05,
|
| 120 |
+
"rewards/rejected": 1.4664057744084857e-05,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.21,
|
| 125 |
+
"learning_rate": 4.828760511501322e-07,
|
| 126 |
+
"logits/chosen": -1.9421491622924805,
|
| 127 |
+
"logits/rejected": -1.9453084468841553,
|
| 128 |
+
"logps/chosen": -34.42374038696289,
|
| 129 |
+
"logps/rejected": -34.57770538330078,
|
| 130 |
+
"loss": 0.6932,
|
| 131 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 132 |
+
"rewards/chosen": 4.0700066165300086e-05,
|
| 133 |
+
"rewards/margins": -0.0001062863739207387,
|
| 134 |
+
"rewards/rejected": 0.0001469864509999752,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.23,
|
| 139 |
+
"learning_rate": 4.736716601303429e-07,
|
| 140 |
+
"logits/chosen": -1.9507558345794678,
|
| 141 |
+
"logits/rejected": -1.9552711248397827,
|
| 142 |
+
"logps/chosen": -32.4448356628418,
|
| 143 |
+
"logps/rejected": -32.35200881958008,
|
| 144 |
+
"loss": 0.6931,
|
| 145 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 146 |
+
"rewards/chosen": 0.0002258731983602047,
|
| 147 |
+
"rewards/margins": 0.00015688061830587685,
|
| 148 |
+
"rewards/rejected": 6.899260188220069e-05,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.26,
|
| 153 |
+
"learning_rate": 4.62624545834521e-07,
|
| 154 |
+
"logits/chosen": -2.0490055084228516,
|
| 155 |
+
"logits/rejected": -2.0470049381256104,
|
| 156 |
+
"logps/chosen": -32.24321365356445,
|
| 157 |
+
"logps/rejected": -31.265995025634766,
|
| 158 |
+
"loss": 0.6932,
|
| 159 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 160 |
+
"rewards/chosen": -2.38026914303191e-05,
|
| 161 |
+
"rewards/margins": -2.5972036382881925e-05,
|
| 162 |
+
"rewards/rejected": 2.169329718526569e-06,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.26,
|
| 167 |
+
"eval_logits/chosen": -2.244168996810913,
|
| 168 |
+
"eval_logits/rejected": -2.23928165435791,
|
| 169 |
+
"eval_logps/chosen": -34.029258728027344,
|
| 170 |
+
"eval_logps/rejected": -37.50376510620117,
|
| 171 |
+
"eval_loss": 0.6931847929954529,
|
| 172 |
+
"eval_rewards/accuracies": 0.46054819226264954,
|
| 173 |
+
"eval_rewards/chosen": 5.296626841300167e-05,
|
| 174 |
+
"eval_rewards/margins": -7.559578079963103e-05,
|
| 175 |
+
"eval_rewards/rejected": 0.00012856203829869628,
|
| 176 |
+
"eval_runtime": 146.0358,
|
| 177 |
+
"eval_samples_per_second": 2.349,
|
| 178 |
+
"eval_steps_per_second": 0.294,
|
| 179 |
+
"step": 100
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.29,
|
| 183 |
+
"learning_rate": 4.4982572012636904e-07,
|
| 184 |
+
"logits/chosen": -2.005403518676758,
|
| 185 |
+
"logits/rejected": -2.0029940605163574,
|
| 186 |
+
"logps/chosen": -33.23992156982422,
|
| 187 |
+
"logps/rejected": -34.01720428466797,
|
| 188 |
+
"loss": 0.6932,
|
| 189 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 190 |
+
"rewards/chosen": 3.894489418598823e-05,
|
| 191 |
+
"rewards/margins": -8.27050898806192e-05,
|
| 192 |
+
"rewards/rejected": 0.00012164998042862862,
|
| 193 |
+
"step": 110
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.31,
|
| 197 |
+
"learning_rate": 4.353806263777677e-07,
|
| 198 |
+
"logits/chosen": -2.0166852474212646,
|
| 199 |
+
"logits/rejected": -2.008302688598633,
|
| 200 |
+
"logps/chosen": -32.434364318847656,
|
| 201 |
+
"logps/rejected": -32.170448303222656,
|
| 202 |
+
"loss": 0.6931,
|
| 203 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 204 |
+
"rewards/chosen": 0.00010542555537540466,
|
| 205 |
+
"rewards/margins": 0.00019486629753373563,
|
| 206 |
+
"rewards/rejected": -8.94407567102462e-05,
|
| 207 |
+
"step": 120
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.34,
|
| 211 |
+
"learning_rate": 4.194082707715275e-07,
|
| 212 |
+
"logits/chosen": -2.0464212894439697,
|
| 213 |
+
"logits/rejected": -2.0383687019348145,
|
| 214 |
+
"logps/chosen": -30.498855590820312,
|
| 215 |
+
"logps/rejected": -32.038177490234375,
|
| 216 |
+
"loss": 0.6933,
|
| 217 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 218 |
+
"rewards/chosen": -0.0002313825534656644,
|
| 219 |
+
"rewards/margins": -0.0003796928212977946,
|
| 220 |
+
"rewards/rejected": 0.00014831028238404542,
|
| 221 |
+
"step": 130
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.36,
|
| 225 |
+
"learning_rate": 4.020402418666621e-07,
|
| 226 |
+
"logits/chosen": -1.9772886037826538,
|
| 227 |
+
"logits/rejected": -1.987566590309143,
|
| 228 |
+
"logps/chosen": -31.40262794494629,
|
| 229 |
+
"logps/rejected": -32.533485412597656,
|
| 230 |
+
"loss": 0.6931,
|
| 231 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 232 |
+
"rewards/chosen": 0.00017036919598467648,
|
| 233 |
+
"rewards/margins": 0.00015645996609237045,
|
| 234 |
+
"rewards/rejected": 1.3909203516959678e-05,
|
| 235 |
+
"step": 140
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.39,
|
| 239 |
+
"learning_rate": 3.8341962650351185e-07,
|
| 240 |
+
"logits/chosen": -1.8902900218963623,
|
| 241 |
+
"logits/rejected": -1.8913724422454834,
|
| 242 |
+
"logps/chosen": -34.174896240234375,
|
| 243 |
+
"logps/rejected": -34.764488220214844,
|
| 244 |
+
"loss": 0.693,
|
| 245 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 246 |
+
"rewards/chosen": 0.00017796576139517128,
|
| 247 |
+
"rewards/margins": 0.00025028776144608855,
|
| 248 |
+
"rewards/rejected": -7.232203643070534e-05,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.42,
|
| 253 |
+
"learning_rate": 3.636998309800572e-07,
|
| 254 |
+
"logits/chosen": -1.9422016143798828,
|
| 255 |
+
"logits/rejected": -1.938711404800415,
|
| 256 |
+
"logps/chosen": -36.14549255371094,
|
| 257 |
+
"logps/rejected": -32.743194580078125,
|
| 258 |
+
"loss": 0.6929,
|
| 259 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 260 |
+
"rewards/chosen": 0.00026183511363342404,
|
| 261 |
+
"rewards/margins": 0.0004662507271859795,
|
| 262 |
+
"rewards/rejected": -0.00020441561355255544,
|
| 263 |
+
"step": 160
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.44,
|
| 267 |
+
"learning_rate": 3.430433172111807e-07,
|
| 268 |
+
"logits/chosen": -2.041975498199463,
|
| 269 |
+
"logits/rejected": -2.034566879272461,
|
| 270 |
+
"logps/chosen": -33.78548049926758,
|
| 271 |
+
"logps/rejected": -31.344890594482422,
|
| 272 |
+
"loss": 0.6931,
|
| 273 |
+
"rewards/accuracies": 0.512499988079071,
|
| 274 |
+
"rewards/chosen": 0.0001237928809132427,
|
| 275 |
+
"rewards/margins": 5.422577305580489e-05,
|
| 276 |
+
"rewards/rejected": 6.956708239158615e-05,
|
| 277 |
+
"step": 170
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 0.47,
|
| 281 |
+
"learning_rate": 3.216202642830543e-07,
|
| 282 |
+
"logits/chosen": -2.047132968902588,
|
| 283 |
+
"logits/rejected": -2.0523972511291504,
|
| 284 |
+
"logps/chosen": -32.526214599609375,
|
| 285 |
+
"logps/rejected": -32.523216247558594,
|
| 286 |
+
"loss": 0.6929,
|
| 287 |
+
"rewards/accuracies": 0.5625,
|
| 288 |
+
"rewards/chosen": 0.00018572784028947353,
|
| 289 |
+
"rewards/margins": 0.000566480855923146,
|
| 290 |
+
"rewards/rejected": -0.0003807530156336725,
|
| 291 |
+
"step": 180
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.49,
|
| 295 |
+
"learning_rate": 2.9960716642946403e-07,
|
| 296 |
+
"logits/chosen": -2.048360824584961,
|
| 297 |
+
"logits/rejected": -2.0455949306488037,
|
| 298 |
+
"logps/chosen": -31.48598289489746,
|
| 299 |
+
"logps/rejected": -31.33578109741211,
|
| 300 |
+
"loss": 0.693,
|
| 301 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 302 |
+
"rewards/chosen": 2.9682874810532667e-05,
|
| 303 |
+
"rewards/margins": 0.00020535592921078205,
|
| 304 |
+
"rewards/rejected": -0.00017567307804711163,
|
| 305 |
+
"step": 190
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.52,
|
| 309 |
+
"learning_rate": 2.771853789806683e-07,
|
| 310 |
+
"logits/chosen": -1.918714165687561,
|
| 311 |
+
"logits/rejected": -1.9233949184417725,
|
| 312 |
+
"logps/chosen": -31.575008392333984,
|
| 313 |
+
"logps/rejected": -32.78480911254883,
|
| 314 |
+
"loss": 0.693,
|
| 315 |
+
"rewards/accuracies": 0.5625,
|
| 316 |
+
"rewards/chosen": 0.0003127219679299742,
|
| 317 |
+
"rewards/margins": 0.00025148707209154963,
|
| 318 |
+
"rewards/rejected": 6.123496132204309e-05,
|
| 319 |
+
"step": 200
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 0.52,
|
| 323 |
+
"eval_logits/chosen": -2.2438974380493164,
|
| 324 |
+
"eval_logits/rejected": -2.2390170097351074,
|
| 325 |
+
"eval_logps/chosen": -34.02705764770508,
|
| 326 |
+
"eval_logps/rejected": -37.500091552734375,
|
| 327 |
+
"eval_loss": 0.6931942701339722,
|
| 328 |
+
"eval_rewards/accuracies": 0.4991694688796997,
|
| 329 |
+
"eval_rewards/chosen": 7.496162288589403e-05,
|
| 330 |
+
"eval_rewards/margins": -9.032993693836033e-05,
|
| 331 |
+
"eval_rewards/rejected": 0.0001652915816521272,
|
| 332 |
+
"eval_runtime": 145.6901,
|
| 333 |
+
"eval_samples_per_second": 2.354,
|
| 334 |
+
"eval_steps_per_second": 0.295,
|
| 335 |
+
"step": 200
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.55,
|
| 339 |
+
"learning_rate": 2.5453962426402e-07,
|
| 340 |
+
"logits/chosen": -2.0324437618255615,
|
| 341 |
+
"logits/rejected": -2.0431275367736816,
|
| 342 |
+
"logps/chosen": -31.958581924438477,
|
| 343 |
+
"logps/rejected": -33.874977111816406,
|
| 344 |
+
"loss": 0.693,
|
| 345 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 346 |
+
"rewards/chosen": 0.00012724331463687122,
|
| 347 |
+
"rewards/margins": 0.00025004256167449057,
|
| 348 |
+
"rewards/rejected": -0.00012279924703761935,
|
| 349 |
+
"step": 210
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 0.57,
|
| 353 |
+
"learning_rate": 2.318564697655179e-07,
|
| 354 |
+
"logits/chosen": -1.9262778759002686,
|
| 355 |
+
"logits/rejected": -1.9411571025848389,
|
| 356 |
+
"logps/chosen": -30.077739715576172,
|
| 357 |
+
"logps/rejected": -31.5812931060791,
|
| 358 |
+
"loss": 0.6929,
|
| 359 |
+
"rewards/accuracies": 0.612500011920929,
|
| 360 |
+
"rewards/chosen": 0.0003380063863005489,
|
| 361 |
+
"rewards/margins": 0.0005691683036275208,
|
| 362 |
+
"rewards/rejected": -0.00023116180091165006,
|
| 363 |
+
"step": 220
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.6,
|
| 367 |
+
"learning_rate": 2.093227910899832e-07,
|
| 368 |
+
"logits/chosen": -1.984165906906128,
|
| 369 |
+
"logits/rejected": -1.9881302118301392,
|
| 370 |
+
"logps/chosen": -33.39875030517578,
|
| 371 |
+
"logps/rejected": -31.574642181396484,
|
| 372 |
+
"loss": 0.6929,
|
| 373 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 374 |
+
"rewards/chosen": 0.00018722013919614255,
|
| 375 |
+
"rewards/margins": 0.0004406363586895168,
|
| 376 |
+
"rewards/rejected": -0.00025341621949337423,
|
| 377 |
+
"step": 230
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.62,
|
| 381 |
+
"learning_rate": 1.8712423238279356e-07,
|
| 382 |
+
"logits/chosen": -1.9830011129379272,
|
| 383 |
+
"logits/rejected": -1.9610159397125244,
|
| 384 |
+
"logps/chosen": -34.161964416503906,
|
| 385 |
+
"logps/rejected": -34.9656867980957,
|
| 386 |
+
"loss": 0.6931,
|
| 387 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 388 |
+
"rewards/chosen": -8.213297405745834e-05,
|
| 389 |
+
"rewards/margins": 6.679827492916957e-05,
|
| 390 |
+
"rewards/rejected": -0.0001489312417106703,
|
| 391 |
+
"step": 240
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.65,
|
| 395 |
+
"learning_rate": 1.654436768970182e-07,
|
| 396 |
+
"logits/chosen": -2.024148464202881,
|
| 397 |
+
"logits/rejected": -2.020843982696533,
|
| 398 |
+
"logps/chosen": -32.924686431884766,
|
| 399 |
+
"logps/rejected": -36.229591369628906,
|
| 400 |
+
"loss": 0.6931,
|
| 401 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 402 |
+
"rewards/chosen": 3.8046332520025317e-06,
|
| 403 |
+
"rewards/margins": 5.0471979193389416e-05,
|
| 404 |
+
"rewards/rejected": -4.666733366320841e-05,
|
| 405 |
+
"step": 250
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.68,
|
| 409 |
+
"learning_rate": 1.444597403062196e-07,
|
| 410 |
+
"logits/chosen": -1.8910725116729736,
|
| 411 |
+
"logits/rejected": -1.8886131048202515,
|
| 412 |
+
"logps/chosen": -34.17741012573242,
|
| 413 |
+
"logps/rejected": -35.50103759765625,
|
| 414 |
+
"loss": 0.6931,
|
| 415 |
+
"rewards/accuracies": 0.512499988079071,
|
| 416 |
+
"rewards/chosen": 0.00010990710143232718,
|
| 417 |
+
"rewards/margins": 7.404615462291986e-05,
|
| 418 |
+
"rewards/rejected": 3.586096136132255e-05,
|
| 419 |
+
"step": 260
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.7,
|
| 423 |
+
"learning_rate": 1.2434529917578887e-07,
|
| 424 |
+
"logits/chosen": -1.8763153553009033,
|
| 425 |
+
"logits/rejected": -1.873784065246582,
|
| 426 |
+
"logps/chosen": -34.39301300048828,
|
| 427 |
+
"logps/rejected": -31.753448486328125,
|
| 428 |
+
"loss": 0.6932,
|
| 429 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 430 |
+
"rewards/chosen": 4.5871554902987555e-05,
|
| 431 |
+
"rewards/margins": -0.00013292724906932563,
|
| 432 |
+
"rewards/rejected": 0.00017879880033433437,
|
| 433 |
+
"step": 270
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.73,
|
| 437 |
+
"learning_rate": 1.0526606671603521e-07,
|
| 438 |
+
"logits/chosen": -1.9792474508285522,
|
| 439 |
+
"logits/rejected": -1.9685996770858765,
|
| 440 |
+
"logps/chosen": -35.32111740112305,
|
| 441 |
+
"logps/rejected": -31.84047508239746,
|
| 442 |
+
"loss": 0.693,
|
| 443 |
+
"rewards/accuracies": 0.5,
|
| 444 |
+
"rewards/chosen": 0.00035524697159416974,
|
| 445 |
+
"rewards/margins": 0.0003500745224300772,
|
| 446 |
+
"rewards/rejected": 5.172446435608435e-06,
|
| 447 |
+
"step": 280
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.75,
|
| 451 |
+
"learning_rate": 8.737922755071453e-08,
|
| 452 |
+
"logits/chosen": -2.0755350589752197,
|
| 453 |
+
"logits/rejected": -2.0605063438415527,
|
| 454 |
+
"logps/chosen": -30.92108154296875,
|
| 455 |
+
"logps/rejected": -32.65219497680664,
|
| 456 |
+
"loss": 0.6931,
|
| 457 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 458 |
+
"rewards/chosen": 3.043042670469731e-05,
|
| 459 |
+
"rewards/margins": 3.4023898479063064e-05,
|
| 460 |
+
"rewards/rejected": -3.593444034777349e-06,
|
| 461 |
+
"step": 290
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.78,
|
| 465 |
+
"learning_rate": 7.08321427484816e-08,
|
| 466 |
+
"logits/chosen": -1.9464104175567627,
|
| 467 |
+
"logits/rejected": -1.9438741207122803,
|
| 468 |
+
"logps/chosen": -32.899200439453125,
|
| 469 |
+
"logps/rejected": -30.818578720092773,
|
| 470 |
+
"loss": 0.693,
|
| 471 |
+
"rewards/accuracies": 0.574999988079071,
|
| 472 |
+
"rewards/chosen": 0.00027910774224437773,
|
| 473 |
+
"rewards/margins": 0.00029683284810744226,
|
| 474 |
+
"rewards/rejected": -1.772511677700095e-05,
|
| 475 |
+
"step": 300
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 0.78,
|
| 479 |
+
"eval_logits/chosen": -2.244286060333252,
|
| 480 |
+
"eval_logits/rejected": -2.2394051551818848,
|
| 481 |
+
"eval_logps/chosen": -34.025508880615234,
|
| 482 |
+
"eval_logps/rejected": -37.50877380371094,
|
| 483 |
+
"eval_loss": 0.6931424140930176,
|
| 484 |
+
"eval_rewards/accuracies": 0.4958471655845642,
|
| 485 |
+
"eval_rewards/chosen": 9.041713929036632e-05,
|
| 486 |
+
"eval_rewards/margins": 1.1897234799107537e-05,
|
| 487 |
+
"eval_rewards/rejected": 7.851990085327998e-05,
|
| 488 |
+
"eval_runtime": 145.6886,
|
| 489 |
+
"eval_samples_per_second": 2.354,
|
| 490 |
+
"eval_steps_per_second": 0.295,
|
| 491 |
+
"step": 300
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.81,
|
| 495 |
+
"learning_rate": 5.576113578589034e-08,
|
| 496 |
+
"logits/chosen": -1.9285697937011719,
|
| 497 |
+
"logits/rejected": -1.9253038167953491,
|
| 498 |
+
"logps/chosen": -31.567317962646484,
|
| 499 |
+
"logps/rejected": -33.728389739990234,
|
| 500 |
+
"loss": 0.693,
|
| 501 |
+
"rewards/accuracies": 0.550000011920929,
|
| 502 |
+
"rewards/chosen": 0.00036028382601216435,
|
| 503 |
+
"rewards/margins": 0.00022268423344939947,
|
| 504 |
+
"rewards/rejected": 0.00013759959256276488,
|
| 505 |
+
"step": 310
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 0.83,
|
| 509 |
+
"learning_rate": 4.229036944380912e-08,
|
| 510 |
+
"logits/chosen": -1.9806677103042603,
|
| 511 |
+
"logits/rejected": -1.9683544635772705,
|
| 512 |
+
"logps/chosen": -34.57938766479492,
|
| 513 |
+
"logps/rejected": -33.576698303222656,
|
| 514 |
+
"loss": 0.6929,
|
| 515 |
+
"rewards/accuracies": 0.625,
|
| 516 |
+
"rewards/chosen": 0.0001669025223236531,
|
| 517 |
+
"rewards/margins": 0.0005454975180327892,
|
| 518 |
+
"rewards/rejected": -0.0003785950248129666,
|
| 519 |
+
"step": 320
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.86,
|
| 523 |
+
"learning_rate": 3.053082288996112e-08,
|
| 524 |
+
"logits/chosen": -2.0156922340393066,
|
| 525 |
+
"logits/rejected": -2.014228105545044,
|
| 526 |
+
"logps/chosen": -33.48464584350586,
|
| 527 |
+
"logps/rejected": -32.47287368774414,
|
| 528 |
+
"loss": 0.6932,
|
| 529 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 530 |
+
"rewards/chosen": 1.1468652701296378e-05,
|
| 531 |
+
"rewards/margins": -5.562617297982797e-05,
|
| 532 |
+
"rewards/rejected": 6.709482113365084e-05,
|
| 533 |
+
"step": 330
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.88,
|
| 537 |
+
"learning_rate": 2.05793773749158e-08,
|
| 538 |
+
"logits/chosen": -2.1031441688537598,
|
| 539 |
+
"logits/rejected": -2.0873398780822754,
|
| 540 |
+
"logps/chosen": -34.18089294433594,
|
| 541 |
+
"logps/rejected": -33.10489273071289,
|
| 542 |
+
"loss": 0.6932,
|
| 543 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 544 |
+
"rewards/chosen": -1.5041228834888898e-05,
|
| 545 |
+
"rewards/margins": -0.000146475198562257,
|
| 546 |
+
"rewards/rejected": 0.0001314339751843363,
|
| 547 |
+
"step": 340
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 0.91,
|
| 551 |
+
"learning_rate": 1.251801807404168e-08,
|
| 552 |
+
"logits/chosen": -1.9747949838638306,
|
| 553 |
+
"logits/rejected": -1.9738445281982422,
|
| 554 |
+
"logps/chosen": -33.250999450683594,
|
| 555 |
+
"logps/rejected": -32.463138580322266,
|
| 556 |
+
"loss": 0.693,
|
| 557 |
+
"rewards/accuracies": 0.574999988079071,
|
| 558 |
+
"rewards/chosen": 0.0003109553945250809,
|
| 559 |
+
"rewards/margins": 0.00029085634741932154,
|
| 560 |
+
"rewards/rejected": 2.009904164879117e-05,
|
| 561 |
+
"step": 350
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.94,
|
| 565 |
+
"learning_rate": 6.41315865106129e-09,
|
| 566 |
+
"logits/chosen": -1.9307000637054443,
|
| 567 |
+
"logits/rejected": -1.9410593509674072,
|
| 568 |
+
"logps/chosen": -32.20253372192383,
|
| 569 |
+
"logps/rejected": -35.28581619262695,
|
| 570 |
+
"loss": 0.6932,
|
| 571 |
+
"rewards/accuracies": 0.5,
|
| 572 |
+
"rewards/chosen": 2.157145536330063e-05,
|
| 573 |
+
"rewards/margins": -7.86231794336345e-06,
|
| 574 |
+
"rewards/rejected": 2.9433791496558115e-05,
|
| 575 |
+
"step": 360
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 0.96,
|
| 579 |
+
"learning_rate": 2.3150941078050324e-09,
|
| 580 |
+
"logits/chosen": -2.069913148880005,
|
| 581 |
+
"logits/rejected": -2.0633533000946045,
|
| 582 |
+
"logps/chosen": -33.65105438232422,
|
| 583 |
+
"logps/rejected": -29.224279403686523,
|
| 584 |
+
"loss": 0.6932,
|
| 585 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 586 |
+
"rewards/chosen": 9.127380508289207e-06,
|
| 587 |
+
"rewards/margins": -1.2333152881183196e-05,
|
| 588 |
+
"rewards/rejected": 2.1460538846440613e-05,
|
| 589 |
+
"step": 370
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 0.99,
|
| 593 |
+
"learning_rate": 2.575864278703266e-10,
|
| 594 |
+
"logits/chosen": -1.9291718006134033,
|
| 595 |
+
"logits/rejected": -1.9313350915908813,
|
| 596 |
+
"logps/chosen": -34.24486541748047,
|
| 597 |
+
"logps/rejected": -30.897415161132812,
|
| 598 |
+
"loss": 0.6931,
|
| 599 |
+
"rewards/accuracies": 0.5625,
|
| 600 |
+
"rewards/chosen": -5.8152021665591747e-05,
|
| 601 |
+
"rewards/margins": 0.00018974675913341343,
|
| 602 |
+
"rewards/rejected": -0.000247898802626878,
|
| 603 |
+
"step": 380
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.0,
|
| 607 |
+
"step": 385,
|
| 608 |
+
"total_flos": 0.0,
|
| 609 |
+
"train_loss": 0.6930816006350827,
|
| 610 |
+
"train_runtime": 3252.6347,
|
| 611 |
+
"train_samples_per_second": 0.947,
|
| 612 |
+
"train_steps_per_second": 0.118
|
| 613 |
+
}
|
| 614 |
+
],
|
| 615 |
+
"logging_steps": 10,
|
| 616 |
+
"max_steps": 385,
|
| 617 |
+
"num_input_tokens_seen": 0,
|
| 618 |
+
"num_train_epochs": 1,
|
| 619 |
+
"save_steps": 100,
|
| 620 |
+
"total_flos": 0.0,
|
| 621 |
+
"train_batch_size": 4,
|
| 622 |
+
"trial_name": null,
|
| 623 |
+
"trial_params": null
|
| 624 |
+
}
|