Model save
Browse files- README.md +62 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +897 -0
README.md
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: llama3
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: meta-llama/Meta-Llama-3-8B
|
| 9 |
+
model-index:
|
| 10 |
+
- name: llama3-sudo-dpo-3epochs-medical-1e-5
|
| 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 |
+
# llama3-sudo-dpo-3epochs-medical-1e-5
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on an unknown 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: 1e-05
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 8
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 4
|
| 44 |
+
- gradient_accumulation_steps: 4
|
| 45 |
+
- total_train_batch_size: 64
|
| 46 |
+
- total_eval_batch_size: 32
|
| 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: 3
|
| 51 |
+
|
| 52 |
+
### Training results
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
### Framework versions
|
| 57 |
+
|
| 58 |
+
- PEFT 0.12.0
|
| 59 |
+
- Transformers 4.44.0
|
| 60 |
+
- Pytorch 2.1.2
|
| 61 |
+
- Datasets 3.0.0
|
| 62 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.992,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.049965105635520604,
|
| 5 |
+
"train_runtime": 8901.9362,
|
| 6 |
+
"train_samples": 12000,
|
| 7 |
+
"train_samples_per_second": 4.044,
|
| 8 |
+
"train_steps_per_second": 0.063
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.992,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.049965105635520604,
|
| 5 |
+
"train_runtime": 8901.9362,
|
| 6 |
+
"train_samples": 12000,
|
| 7 |
+
"train_samples_per_second": 4.044,
|
| 8 |
+
"train_steps_per_second": 0.063
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,897 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.992,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 561,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.005333333333333333,
|
| 13 |
+
"grad_norm": 0.7276772136468688,
|
| 14 |
+
"learning_rate": 1.7543859649122808e-07,
|
| 15 |
+
"logits/chosen": -1.480194330215454,
|
| 16 |
+
"logits/rejected": -1.445469856262207,
|
| 17 |
+
"logps/chosen": -200.37672424316406,
|
| 18 |
+
"logps/rejected": -252.16812133789062,
|
| 19 |
+
"loss": 0.6931,
|
| 20 |
+
"rewards/accuracies": 0.0,
|
| 21 |
+
"rewards/chosen": 0.0,
|
| 22 |
+
"rewards/margins": 0.0,
|
| 23 |
+
"rewards/rejected": 0.0,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.05333333333333334,
|
| 28 |
+
"grad_norm": 0.7041154730069451,
|
| 29 |
+
"learning_rate": 1.7543859649122807e-06,
|
| 30 |
+
"logits/chosen": -1.4611287117004395,
|
| 31 |
+
"logits/rejected": -1.460557222366333,
|
| 32 |
+
"logps/chosen": -251.83726501464844,
|
| 33 |
+
"logps/rejected": -254.89569091796875,
|
| 34 |
+
"loss": 0.6926,
|
| 35 |
+
"rewards/accuracies": 0.5208333134651184,
|
| 36 |
+
"rewards/chosen": 0.0008822197560220957,
|
| 37 |
+
"rewards/margins": 0.0008583312737755477,
|
| 38 |
+
"rewards/rejected": 2.3888502255431376e-05,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.10666666666666667,
|
| 43 |
+
"grad_norm": 0.6867796553002457,
|
| 44 |
+
"learning_rate": 3.5087719298245615e-06,
|
| 45 |
+
"logits/chosen": -1.4618045091629028,
|
| 46 |
+
"logits/rejected": -1.460260272026062,
|
| 47 |
+
"logps/chosen": -249.17172241210938,
|
| 48 |
+
"logps/rejected": -243.20840454101562,
|
| 49 |
+
"loss": 0.6825,
|
| 50 |
+
"rewards/accuracies": 0.875,
|
| 51 |
+
"rewards/chosen": 0.011592308059334755,
|
| 52 |
+
"rewards/margins": 0.01997370831668377,
|
| 53 |
+
"rewards/rejected": -0.00838139932602644,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.16,
|
| 58 |
+
"grad_norm": 0.9201550701932262,
|
| 59 |
+
"learning_rate": 5.263157894736842e-06,
|
| 60 |
+
"logits/chosen": -1.449467420578003,
|
| 61 |
+
"logits/rejected": -1.4497499465942383,
|
| 62 |
+
"logps/chosen": -244.9186248779297,
|
| 63 |
+
"logps/rejected": -258.53594970703125,
|
| 64 |
+
"loss": 0.6393,
|
| 65 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 66 |
+
"rewards/chosen": 0.0312686562538147,
|
| 67 |
+
"rewards/margins": 0.10987410694360733,
|
| 68 |
+
"rewards/rejected": -0.07860545814037323,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.21333333333333335,
|
| 73 |
+
"grad_norm": 2.166122053069534,
|
| 74 |
+
"learning_rate": 7.017543859649123e-06,
|
| 75 |
+
"logits/chosen": -1.3264390230178833,
|
| 76 |
+
"logits/rejected": -1.2739282846450806,
|
| 77 |
+
"logps/chosen": -248.67361450195312,
|
| 78 |
+
"logps/rejected": -315.01190185546875,
|
| 79 |
+
"loss": 0.493,
|
| 80 |
+
"rewards/accuracies": 1.0,
|
| 81 |
+
"rewards/chosen": 0.05745680257678032,
|
| 82 |
+
"rewards/margins": 0.5744289755821228,
|
| 83 |
+
"rewards/rejected": -0.5169721841812134,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.26666666666666666,
|
| 88 |
+
"grad_norm": 0.5760301626216394,
|
| 89 |
+
"learning_rate": 8.771929824561405e-06,
|
| 90 |
+
"logits/chosen": -1.0679562091827393,
|
| 91 |
+
"logits/rejected": -0.7663384675979614,
|
| 92 |
+
"logps/chosen": -266.1653747558594,
|
| 93 |
+
"logps/rejected": -710.7122802734375,
|
| 94 |
+
"loss": 0.227,
|
| 95 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 96 |
+
"rewards/chosen": -0.03745045140385628,
|
| 97 |
+
"rewards/margins": 4.618867874145508,
|
| 98 |
+
"rewards/rejected": -4.656317710876465,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.32,
|
| 103 |
+
"grad_norm": 0.1657312187471681,
|
| 104 |
+
"learning_rate": 9.999125804563732e-06,
|
| 105 |
+
"logits/chosen": -0.7736077308654785,
|
| 106 |
+
"logits/rejected": -0.20556649565696716,
|
| 107 |
+
"logps/chosen": -297.69696044921875,
|
| 108 |
+
"logps/rejected": -1320.8775634765625,
|
| 109 |
+
"loss": 0.0582,
|
| 110 |
+
"rewards/accuracies": 1.0,
|
| 111 |
+
"rewards/chosen": -0.49763980507850647,
|
| 112 |
+
"rewards/margins": 10.216059684753418,
|
| 113 |
+
"rewards/rejected": -10.713699340820312,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.37333333333333335,
|
| 118 |
+
"grad_norm": 0.0045546225580099286,
|
| 119 |
+
"learning_rate": 9.983593054785776e-06,
|
| 120 |
+
"logits/chosen": -0.42266106605529785,
|
| 121 |
+
"logits/rejected": 0.3064606785774231,
|
| 122 |
+
"logps/chosen": -336.424072265625,
|
| 123 |
+
"logps/rejected": -2042.226318359375,
|
| 124 |
+
"loss": 0.0033,
|
| 125 |
+
"rewards/accuracies": 1.0,
|
| 126 |
+
"rewards/chosen": -0.825412392616272,
|
| 127 |
+
"rewards/margins": 16.916912078857422,
|
| 128 |
+
"rewards/rejected": -17.74232292175293,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.4266666666666667,
|
| 133 |
+
"grad_norm": 0.009511386498326472,
|
| 134 |
+
"learning_rate": 9.948703190391131e-06,
|
| 135 |
+
"logits/chosen": -0.2592306137084961,
|
| 136 |
+
"logits/rejected": 0.5354924201965332,
|
| 137 |
+
"logps/chosen": -330.1050109863281,
|
| 138 |
+
"logps/rejected": -2244.28857421875,
|
| 139 |
+
"loss": 0.0007,
|
| 140 |
+
"rewards/accuracies": 1.0,
|
| 141 |
+
"rewards/chosen": -0.9192193150520325,
|
| 142 |
+
"rewards/margins": 18.9534912109375,
|
| 143 |
+
"rewards/rejected": -19.872713088989258,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.48,
|
| 148 |
+
"grad_norm": 0.04227567141303192,
|
| 149 |
+
"learning_rate": 9.894591729488243e-06,
|
| 150 |
+
"logits/chosen": -0.03826441615819931,
|
| 151 |
+
"logits/rejected": 0.9021320343017578,
|
| 152 |
+
"logps/chosen": -337.2659912109375,
|
| 153 |
+
"logps/rejected": -2535.1923828125,
|
| 154 |
+
"loss": 0.0004,
|
| 155 |
+
"rewards/accuracies": 1.0,
|
| 156 |
+
"rewards/chosen": -1.057660460472107,
|
| 157 |
+
"rewards/margins": 21.727779388427734,
|
| 158 |
+
"rewards/rejected": -22.785442352294922,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.5333333333333333,
|
| 163 |
+
"grad_norm": 0.018772815508711668,
|
| 164 |
+
"learning_rate": 9.82146885010314e-06,
|
| 165 |
+
"logits/chosen": -0.08644755929708481,
|
| 166 |
+
"logits/rejected": 0.9283016324043274,
|
| 167 |
+
"logps/chosen": -351.6574401855469,
|
| 168 |
+
"logps/rejected": -2719.936767578125,
|
| 169 |
+
"loss": 0.0003,
|
| 170 |
+
"rewards/accuracies": 1.0,
|
| 171 |
+
"rewards/chosen": -0.9892961382865906,
|
| 172 |
+
"rewards/margins": 23.653213500976562,
|
| 173 |
+
"rewards/rejected": -24.642505645751953,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.5866666666666667,
|
| 178 |
+
"grad_norm": 0.00020416355795206898,
|
| 179 |
+
"learning_rate": 9.72961857381258e-06,
|
| 180 |
+
"logits/chosen": 0.040968358516693115,
|
| 181 |
+
"logits/rejected": 1.012926697731018,
|
| 182 |
+
"logps/chosen": -324.0478210449219,
|
| 183 |
+
"logps/rejected": -2775.6318359375,
|
| 184 |
+
"loss": 0.0001,
|
| 185 |
+
"rewards/accuracies": 1.0,
|
| 186 |
+
"rewards/chosen": -0.9146736264228821,
|
| 187 |
+
"rewards/margins": 24.339845657348633,
|
| 188 |
+
"rewards/rejected": -25.254520416259766,
|
| 189 |
+
"step": 110
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.64,
|
| 193 |
+
"grad_norm": 0.00020565435473269487,
|
| 194 |
+
"learning_rate": 9.619397662556434e-06,
|
| 195 |
+
"logits/chosen": -0.06887398660182953,
|
| 196 |
+
"logits/rejected": 0.8867729902267456,
|
| 197 |
+
"logps/chosen": -389.4439697265625,
|
| 198 |
+
"logps/rejected": -2713.948974609375,
|
| 199 |
+
"loss": 0.0011,
|
| 200 |
+
"rewards/accuracies": 1.0,
|
| 201 |
+
"rewards/chosen": -1.1910871267318726,
|
| 202 |
+
"rewards/margins": 23.515239715576172,
|
| 203 |
+
"rewards/rejected": -24.706327438354492,
|
| 204 |
+
"step": 120
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.6933333333333334,
|
| 208 |
+
"grad_norm": 0.07831943931959111,
|
| 209 |
+
"learning_rate": 9.49123423291422e-06,
|
| 210 |
+
"logits/chosen": -0.011257803067564964,
|
| 211 |
+
"logits/rejected": 1.0701903104782104,
|
| 212 |
+
"logps/chosen": -373.259033203125,
|
| 213 |
+
"logps/rejected": -2927.7978515625,
|
| 214 |
+
"loss": 0.0005,
|
| 215 |
+
"rewards/accuracies": 1.0,
|
| 216 |
+
"rewards/chosen": -1.190225601196289,
|
| 217 |
+
"rewards/margins": 25.528148651123047,
|
| 218 |
+
"rewards/rejected": -26.718372344970703,
|
| 219 |
+
"step": 130
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.7466666666666667,
|
| 223 |
+
"grad_norm": 0.0004181495770625809,
|
| 224 |
+
"learning_rate": 9.345626093228233e-06,
|
| 225 |
+
"logits/chosen": 0.1232583150267601,
|
| 226 |
+
"logits/rejected": 1.1058192253112793,
|
| 227 |
+
"logps/chosen": -368.08062744140625,
|
| 228 |
+
"logps/rejected": -2857.86767578125,
|
| 229 |
+
"loss": 0.0,
|
| 230 |
+
"rewards/accuracies": 1.0,
|
| 231 |
+
"rewards/chosen": -1.0861897468566895,
|
| 232 |
+
"rewards/margins": 25.07080841064453,
|
| 233 |
+
"rewards/rejected": -26.156997680664062,
|
| 234 |
+
"step": 140
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.8,
|
| 238 |
+
"grad_norm": 0.0003206578966702994,
|
| 239 |
+
"learning_rate": 9.1831388100321e-06,
|
| 240 |
+
"logits/chosen": 0.22049780189990997,
|
| 241 |
+
"logits/rejected": 1.3007761240005493,
|
| 242 |
+
"logps/chosen": -348.95294189453125,
|
| 243 |
+
"logps/rejected": -2912.65673828125,
|
| 244 |
+
"loss": 0.0001,
|
| 245 |
+
"rewards/accuracies": 1.0,
|
| 246 |
+
"rewards/chosen": -0.9624420404434204,
|
| 247 |
+
"rewards/margins": 25.764638900756836,
|
| 248 |
+
"rewards/rejected": -26.72708511352539,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.8533333333333334,
|
| 253 |
+
"grad_norm": 0.0002525473318545097,
|
| 254 |
+
"learning_rate": 9.004403511295141e-06,
|
| 255 |
+
"logits/chosen": 0.25433215498924255,
|
| 256 |
+
"logits/rejected": 1.5667898654937744,
|
| 257 |
+
"logps/chosen": -354.92938232421875,
|
| 258 |
+
"logps/rejected": -3299.885986328125,
|
| 259 |
+
"loss": 0.0002,
|
| 260 |
+
"rewards/accuracies": 1.0,
|
| 261 |
+
"rewards/chosen": -0.9957138895988464,
|
| 262 |
+
"rewards/margins": 29.404260635375977,
|
| 263 |
+
"rewards/rejected": -30.39996910095215,
|
| 264 |
+
"step": 160
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.9066666666666666,
|
| 268 |
+
"grad_norm": 0.00013079558655504706,
|
| 269 |
+
"learning_rate": 8.810114435015055e-06,
|
| 270 |
+
"logits/chosen": 0.4517189562320709,
|
| 271 |
+
"logits/rejected": 1.8216791152954102,
|
| 272 |
+
"logps/chosen": -349.7914733886719,
|
| 273 |
+
"logps/rejected": -3455.74755859375,
|
| 274 |
+
"loss": 0.0005,
|
| 275 |
+
"rewards/accuracies": 1.0,
|
| 276 |
+
"rewards/chosen": -1.1400539875030518,
|
| 277 |
+
"rewards/margins": 30.764062881469727,
|
| 278 |
+
"rewards/rejected": -31.90411376953125,
|
| 279 |
+
"step": 170
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.96,
|
| 283 |
+
"grad_norm": 0.0003519489046110966,
|
| 284 |
+
"learning_rate": 8.601026232680634e-06,
|
| 285 |
+
"logits/chosen": 0.08518455922603607,
|
| 286 |
+
"logits/rejected": 1.142094373703003,
|
| 287 |
+
"logps/chosen": -370.423828125,
|
| 288 |
+
"logps/rejected": -3112.953125,
|
| 289 |
+
"loss": 0.0,
|
| 290 |
+
"rewards/accuracies": 1.0,
|
| 291 |
+
"rewards/chosen": -1.124570608139038,
|
| 292 |
+
"rewards/margins": 27.508525848388672,
|
| 293 |
+
"rewards/rejected": -28.633098602294922,
|
| 294 |
+
"step": 180
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 1.0133333333333334,
|
| 298 |
+
"grad_norm": 9.418447063320835e-05,
|
| 299 |
+
"learning_rate": 8.377951038078303e-06,
|
| 300 |
+
"logits/chosen": 0.09172124415636063,
|
| 301 |
+
"logits/rejected": 1.2133045196533203,
|
| 302 |
+
"logps/chosen": -377.74609375,
|
| 303 |
+
"logps/rejected": -3092.125,
|
| 304 |
+
"loss": 0.0001,
|
| 305 |
+
"rewards/accuracies": 1.0,
|
| 306 |
+
"rewards/chosen": -1.2323256731033325,
|
| 307 |
+
"rewards/margins": 27.186447143554688,
|
| 308 |
+
"rewards/rejected": -28.418771743774414,
|
| 309 |
+
"step": 190
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 1.0666666666666667,
|
| 313 |
+
"grad_norm": 5.617178391629701e-05,
|
| 314 |
+
"learning_rate": 8.141755312827737e-06,
|
| 315 |
+
"logits/chosen": 0.26229405403137207,
|
| 316 |
+
"logits/rejected": 1.4855743646621704,
|
| 317 |
+
"logps/chosen": -363.44500732421875,
|
| 318 |
+
"logps/rejected": -3202.00830078125,
|
| 319 |
+
"loss": 0.0003,
|
| 320 |
+
"rewards/accuracies": 1.0,
|
| 321 |
+
"rewards/chosen": -1.1322368383407593,
|
| 322 |
+
"rewards/margins": 28.364980697631836,
|
| 323 |
+
"rewards/rejected": -29.49721908569336,
|
| 324 |
+
"step": 200
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 1.12,
|
| 328 |
+
"grad_norm": 0.0006911681869226501,
|
| 329 |
+
"learning_rate": 7.89335648089903e-06,
|
| 330 |
+
"logits/chosen": 0.27199310064315796,
|
| 331 |
+
"logits/rejected": 1.5671254396438599,
|
| 332 |
+
"logps/chosen": -383.5852966308594,
|
| 333 |
+
"logps/rejected": -3335.39453125,
|
| 334 |
+
"loss": 0.0002,
|
| 335 |
+
"rewards/accuracies": 1.0,
|
| 336 |
+
"rewards/chosen": -1.2137231826782227,
|
| 337 |
+
"rewards/margins": 29.575260162353516,
|
| 338 |
+
"rewards/rejected": -30.78898048400879,
|
| 339 |
+
"step": 210
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.1733333333333333,
|
| 343 |
+
"grad_norm": 0.0001949526485184765,
|
| 344 |
+
"learning_rate": 7.633719365183505e-06,
|
| 345 |
+
"logits/chosen": 0.48210054636001587,
|
| 346 |
+
"logits/rejected": 2.022475481033325,
|
| 347 |
+
"logps/chosen": -336.2737121582031,
|
| 348 |
+
"logps/rejected": -3361.983642578125,
|
| 349 |
+
"loss": 0.0001,
|
| 350 |
+
"rewards/accuracies": 1.0,
|
| 351 |
+
"rewards/chosen": -1.048081636428833,
|
| 352 |
+
"rewards/margins": 30.004507064819336,
|
| 353 |
+
"rewards/rejected": -31.052587509155273,
|
| 354 |
+
"step": 220
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.2266666666666666,
|
| 358 |
+
"grad_norm": 0.0003148592569737297,
|
| 359 |
+
"learning_rate": 7.363852439959136e-06,
|
| 360 |
+
"logits/chosen": 0.2750876247882843,
|
| 361 |
+
"logits/rejected": 1.6405880451202393,
|
| 362 |
+
"logps/chosen": -348.73944091796875,
|
| 363 |
+
"logps/rejected": -3349.078857421875,
|
| 364 |
+
"loss": 0.0,
|
| 365 |
+
"rewards/accuracies": 1.0,
|
| 366 |
+
"rewards/chosen": -0.9442297220230103,
|
| 367 |
+
"rewards/margins": 30.076862335205078,
|
| 368 |
+
"rewards/rejected": -31.02109146118164,
|
| 369 |
+
"step": 230
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.28,
|
| 373 |
+
"grad_norm": 0.0001452254973725037,
|
| 374 |
+
"learning_rate": 7.084803913806642e-06,
|
| 375 |
+
"logits/chosen": 0.2403218001127243,
|
| 376 |
+
"logits/rejected": 1.4615638256072998,
|
| 377 |
+
"logps/chosen": -350.90313720703125,
|
| 378 |
+
"logps/rejected": -3641.379638671875,
|
| 379 |
+
"loss": 0.0,
|
| 380 |
+
"rewards/accuracies": 1.0,
|
| 381 |
+
"rewards/chosen": -0.9335042238235474,
|
| 382 |
+
"rewards/margins": 32.821163177490234,
|
| 383 |
+
"rewards/rejected": -33.75466537475586,
|
| 384 |
+
"step": 240
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.3333333333333333,
|
| 388 |
+
"grad_norm": 0.0017729334665338538,
|
| 389 |
+
"learning_rate": 6.797657658190838e-06,
|
| 390 |
+
"logits/chosen": 0.20432284474372864,
|
| 391 |
+
"logits/rejected": 1.5277740955352783,
|
| 392 |
+
"logps/chosen": -366.70648193359375,
|
| 393 |
+
"logps/rejected": -3363.55615234375,
|
| 394 |
+
"loss": 0.0,
|
| 395 |
+
"rewards/accuracies": 1.0,
|
| 396 |
+
"rewards/chosen": -0.9871833920478821,
|
| 397 |
+
"rewards/margins": 30.192123413085938,
|
| 398 |
+
"rewards/rejected": -31.179306030273438,
|
| 399 |
+
"step": 250
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 1.3866666666666667,
|
| 403 |
+
"grad_norm": 4.839956002070519e-05,
|
| 404 |
+
"learning_rate": 6.503528997521365e-06,
|
| 405 |
+
"logits/chosen": 0.35634833574295044,
|
| 406 |
+
"logits/rejected": 1.6659228801727295,
|
| 407 |
+
"logps/chosen": -353.92230224609375,
|
| 408 |
+
"logps/rejected": -3349.120361328125,
|
| 409 |
+
"loss": 0.0,
|
| 410 |
+
"rewards/accuracies": 1.0,
|
| 411 |
+
"rewards/chosen": -0.9974803924560547,
|
| 412 |
+
"rewards/margins": 30.04400062561035,
|
| 413 |
+
"rewards/rejected": -31.04148292541504,
|
| 414 |
+
"step": 260
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 1.44,
|
| 418 |
+
"grad_norm": 0.00010640905181507166,
|
| 419 |
+
"learning_rate": 6.2035603770448664e-06,
|
| 420 |
+
"logits/chosen": 0.20499210059642792,
|
| 421 |
+
"logits/rejected": 1.4853886365890503,
|
| 422 |
+
"logps/chosen": -349.24420166015625,
|
| 423 |
+
"logps/rejected": -3330.802001953125,
|
| 424 |
+
"loss": 0.0001,
|
| 425 |
+
"rewards/accuracies": 1.0,
|
| 426 |
+
"rewards/chosen": -0.9278079271316528,
|
| 427 |
+
"rewards/margins": 29.96002197265625,
|
| 428 |
+
"rewards/rejected": -30.887828826904297,
|
| 429 |
+
"step": 270
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.4933333333333334,
|
| 433 |
+
"grad_norm": 0.0008385515993980591,
|
| 434 |
+
"learning_rate": 5.8989169253952635e-06,
|
| 435 |
+
"logits/chosen": 0.48648738861083984,
|
| 436 |
+
"logits/rejected": 1.7824747562408447,
|
| 437 |
+
"logps/chosen": -365.4317321777344,
|
| 438 |
+
"logps/rejected": -3538.17529296875,
|
| 439 |
+
"loss": 0.0,
|
| 440 |
+
"rewards/accuracies": 1.0,
|
| 441 |
+
"rewards/chosen": -1.0642529726028442,
|
| 442 |
+
"rewards/margins": 31.774890899658203,
|
| 443 |
+
"rewards/rejected": -32.83914566040039,
|
| 444 |
+
"step": 280
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.5466666666666666,
|
| 448 |
+
"grad_norm": 0.006853385916627181,
|
| 449 |
+
"learning_rate": 5.590781929037965e-06,
|
| 450 |
+
"logits/chosen": 0.32097750902175903,
|
| 451 |
+
"logits/rejected": 1.6314767599105835,
|
| 452 |
+
"logps/chosen": -339.53387451171875,
|
| 453 |
+
"logps/rejected": -3474.714111328125,
|
| 454 |
+
"loss": 0.0,
|
| 455 |
+
"rewards/accuracies": 1.0,
|
| 456 |
+
"rewards/chosen": -0.9878216981887817,
|
| 457 |
+
"rewards/margins": 31.292749404907227,
|
| 458 |
+
"rewards/rejected": -32.28057098388672,
|
| 459 |
+
"step": 290
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.6,
|
| 463 |
+
"grad_norm": 5.799629098534532e-05,
|
| 464 |
+
"learning_rate": 5.2803522361859596e-06,
|
| 465 |
+
"logits/chosen": 0.33614903688430786,
|
| 466 |
+
"logits/rejected": 1.679677963256836,
|
| 467 |
+
"logps/chosen": -359.19586181640625,
|
| 468 |
+
"logps/rejected": -3670.31787109375,
|
| 469 |
+
"loss": 0.0001,
|
| 470 |
+
"rewards/accuracies": 1.0,
|
| 471 |
+
"rewards/chosen": -1.0481712818145752,
|
| 472 |
+
"rewards/margins": 33.0955924987793,
|
| 473 |
+
"rewards/rejected": -34.143768310546875,
|
| 474 |
+
"step": 300
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 1.6533333333333333,
|
| 478 |
+
"grad_norm": 4.24276661954427e-05,
|
| 479 |
+
"learning_rate": 4.968833608039832e-06,
|
| 480 |
+
"logits/chosen": 0.534166693687439,
|
| 481 |
+
"logits/rejected": 1.9130985736846924,
|
| 482 |
+
"logps/chosen": -358.578125,
|
| 483 |
+
"logps/rejected": -3584.983154296875,
|
| 484 |
+
"loss": 0.0001,
|
| 485 |
+
"rewards/accuracies": 1.0,
|
| 486 |
+
"rewards/chosen": -1.0471018552780151,
|
| 487 |
+
"rewards/margins": 32.28336715698242,
|
| 488 |
+
"rewards/rejected": -33.33047103881836,
|
| 489 |
+
"step": 310
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.7066666666666666,
|
| 493 |
+
"grad_norm": 0.00023914971694466748,
|
| 494 |
+
"learning_rate": 4.657436035408217e-06,
|
| 495 |
+
"logits/chosen": 0.5404218435287476,
|
| 496 |
+
"logits/rejected": 2.118741273880005,
|
| 497 |
+
"logps/chosen": -337.90423583984375,
|
| 498 |
+
"logps/rejected": -3778.611328125,
|
| 499 |
+
"loss": 0.0001,
|
| 500 |
+
"rewards/accuracies": 1.0,
|
| 501 |
+
"rewards/chosen": -0.9940769076347351,
|
| 502 |
+
"rewards/margins": 34.16416549682617,
|
| 503 |
+
"rewards/rejected": -35.15824508666992,
|
| 504 |
+
"step": 320
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 1.76,
|
| 508 |
+
"grad_norm": 5.3234427116302415e-05,
|
| 509 |
+
"learning_rate": 4.347369038899744e-06,
|
| 510 |
+
"logits/chosen": 0.2553441822528839,
|
| 511 |
+
"logits/rejected": 1.6488205194473267,
|
| 512 |
+
"logps/chosen": -381.4933166503906,
|
| 513 |
+
"logps/rejected": -3569.164794921875,
|
| 514 |
+
"loss": 0.0007,
|
| 515 |
+
"rewards/accuracies": 1.0,
|
| 516 |
+
"rewards/chosen": -1.153830885887146,
|
| 517 |
+
"rewards/margins": 32.051170349121094,
|
| 518 |
+
"rewards/rejected": -33.20500183105469,
|
| 519 |
+
"step": 330
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.8133333333333335,
|
| 523 |
+
"grad_norm": 3.284575847964367e-05,
|
| 524 |
+
"learning_rate": 4.0398369709413195e-06,
|
| 525 |
+
"logits/chosen": 0.4040505886077881,
|
| 526 |
+
"logits/rejected": 1.7738959789276123,
|
| 527 |
+
"logps/chosen": -354.72760009765625,
|
| 528 |
+
"logps/rejected": -3770.258544921875,
|
| 529 |
+
"loss": 0.0,
|
| 530 |
+
"rewards/accuracies": 1.0,
|
| 531 |
+
"rewards/chosen": -1.0133658647537231,
|
| 532 |
+
"rewards/margins": 34.131263732910156,
|
| 533 |
+
"rewards/rejected": -35.14462661743164,
|
| 534 |
+
"step": 340
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 1.8666666666666667,
|
| 538 |
+
"grad_norm": 7.061849396581742e-05,
|
| 539 |
+
"learning_rate": 3.736034337870512e-06,
|
| 540 |
+
"logits/chosen": 0.33974844217300415,
|
| 541 |
+
"logits/rejected": 1.8519665002822876,
|
| 542 |
+
"logps/chosen": -352.6244201660156,
|
| 543 |
+
"logps/rejected": -3790.14453125,
|
| 544 |
+
"loss": 0.0001,
|
| 545 |
+
"rewards/accuracies": 1.0,
|
| 546 |
+
"rewards/chosen": -1.0042585134506226,
|
| 547 |
+
"rewards/margins": 34.339210510253906,
|
| 548 |
+
"rewards/rejected": -35.34347152709961,
|
| 549 |
+
"step": 350
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.92,
|
| 553 |
+
"grad_norm": 0.0037822509705480955,
|
| 554 |
+
"learning_rate": 3.4371411602717785e-06,
|
| 555 |
+
"logits/chosen": 0.2804495394229889,
|
| 556 |
+
"logits/rejected": 1.6284122467041016,
|
| 557 |
+
"logps/chosen": -372.8134460449219,
|
| 558 |
+
"logps/rejected": -3452.75537109375,
|
| 559 |
+
"loss": 0.0,
|
| 560 |
+
"rewards/accuracies": 1.0,
|
| 561 |
+
"rewards/chosen": -1.0843387842178345,
|
| 562 |
+
"rewards/margins": 31.060293197631836,
|
| 563 |
+
"rewards/rejected": -32.144630432128906,
|
| 564 |
+
"step": 360
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 1.9733333333333334,
|
| 568 |
+
"grad_norm": 0.0032698987512048447,
|
| 569 |
+
"learning_rate": 3.1443183895778105e-06,
|
| 570 |
+
"logits/chosen": 0.39163050055503845,
|
| 571 |
+
"logits/rejected": 1.7898658514022827,
|
| 572 |
+
"logps/chosen": -356.69293212890625,
|
| 573 |
+
"logps/rejected": -3789.28662109375,
|
| 574 |
+
"loss": 0.0002,
|
| 575 |
+
"rewards/accuracies": 1.0,
|
| 576 |
+
"rewards/chosen": -1.0389353036880493,
|
| 577 |
+
"rewards/margins": 34.344696044921875,
|
| 578 |
+
"rewards/rejected": -35.383628845214844,
|
| 579 |
+
"step": 370
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 2.026666666666667,
|
| 583 |
+
"grad_norm": 7.942034446736206e-05,
|
| 584 |
+
"learning_rate": 2.8587033987386857e-06,
|
| 585 |
+
"logits/chosen": 0.646866500377655,
|
| 586 |
+
"logits/rejected": 2.1818718910217285,
|
| 587 |
+
"logps/chosen": -347.6737365722656,
|
| 588 |
+
"logps/rejected": -3894.88525390625,
|
| 589 |
+
"loss": 0.0,
|
| 590 |
+
"rewards/accuracies": 1.0,
|
| 591 |
+
"rewards/chosen": -0.9586876630783081,
|
| 592 |
+
"rewards/margins": 35.39875030517578,
|
| 593 |
+
"rewards/rejected": -36.35744094848633,
|
| 594 |
+
"step": 380
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 2.08,
|
| 598 |
+
"grad_norm": 0.007394158789638307,
|
| 599 |
+
"learning_rate": 2.5814055644738013e-06,
|
| 600 |
+
"logits/chosen": 0.6307215094566345,
|
| 601 |
+
"logits/rejected": 2.338592529296875,
|
| 602 |
+
"logps/chosen": -339.7705993652344,
|
| 603 |
+
"logps/rejected": -4155.8330078125,
|
| 604 |
+
"loss": 0.0,
|
| 605 |
+
"rewards/accuracies": 1.0,
|
| 606 |
+
"rewards/chosen": -1.053788661956787,
|
| 607 |
+
"rewards/margins": 37.81401824951172,
|
| 608 |
+
"rewards/rejected": -38.8678092956543,
|
| 609 |
+
"step": 390
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 2.1333333333333333,
|
| 613 |
+
"grad_norm": 0.00019979686681868912,
|
| 614 |
+
"learning_rate": 2.3135019582658803e-06,
|
| 615 |
+
"logits/chosen": 0.3243342936038971,
|
| 616 |
+
"logits/rejected": 1.6741666793823242,
|
| 617 |
+
"logps/chosen": -371.49658203125,
|
| 618 |
+
"logps/rejected": -3645.52880859375,
|
| 619 |
+
"loss": 0.0,
|
| 620 |
+
"rewards/accuracies": 1.0,
|
| 621 |
+
"rewards/chosen": -1.1034276485443115,
|
| 622 |
+
"rewards/margins": 32.93607711791992,
|
| 623 |
+
"rewards/rejected": -34.03950881958008,
|
| 624 |
+
"step": 400
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 2.1866666666666665,
|
| 628 |
+
"grad_norm": 0.00033866067681241304,
|
| 629 |
+
"learning_rate": 2.056033162833977e-06,
|
| 630 |
+
"logits/chosen": 0.493365615606308,
|
| 631 |
+
"logits/rejected": 2.103060007095337,
|
| 632 |
+
"logps/chosen": -371.1989440917969,
|
| 633 |
+
"logps/rejected": -3984.021484375,
|
| 634 |
+
"loss": 0.0,
|
| 635 |
+
"rewards/accuracies": 1.0,
|
| 636 |
+
"rewards/chosen": -1.1410273313522339,
|
| 637 |
+
"rewards/margins": 36.14289093017578,
|
| 638 |
+
"rewards/rejected": -37.283912658691406,
|
| 639 |
+
"step": 410
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 2.24,
|
| 643 |
+
"grad_norm": 0.00014521714619967945,
|
| 644 |
+
"learning_rate": 1.809999230334958e-06,
|
| 645 |
+
"logits/chosen": 0.731221079826355,
|
| 646 |
+
"logits/rejected": 2.3306963443756104,
|
| 647 |
+
"logps/chosen": -349.55670166015625,
|
| 648 |
+
"logps/rejected": -3872.18896484375,
|
| 649 |
+
"loss": 0.0002,
|
| 650 |
+
"rewards/accuracies": 1.0,
|
| 651 |
+
"rewards/chosen": -1.0938142538070679,
|
| 652 |
+
"rewards/margins": 35.13497543334961,
|
| 653 |
+
"rewards/rejected": -36.22879409790039,
|
| 654 |
+
"step": 420
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 2.2933333333333334,
|
| 658 |
+
"grad_norm": 0.00020768738588162756,
|
| 659 |
+
"learning_rate": 1.5763557979925326e-06,
|
| 660 |
+
"logits/chosen": 0.5012251138687134,
|
| 661 |
+
"logits/rejected": 1.9018030166625977,
|
| 662 |
+
"logps/chosen": -372.3318176269531,
|
| 663 |
+
"logps/rejected": -3688.88134765625,
|
| 664 |
+
"loss": 0.0,
|
| 665 |
+
"rewards/accuracies": 1.0,
|
| 666 |
+
"rewards/chosen": -1.1943247318267822,
|
| 667 |
+
"rewards/margins": 33.225135803222656,
|
| 668 |
+
"rewards/rejected": -34.419456481933594,
|
| 669 |
+
"step": 430
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.3466666666666667,
|
| 673 |
+
"grad_norm": 7.310800602605835e-05,
|
| 674 |
+
"learning_rate": 1.3560103762413584e-06,
|
| 675 |
+
"logits/chosen": 0.5643904805183411,
|
| 676 |
+
"logits/rejected": 2.1081748008728027,
|
| 677 |
+
"logps/chosen": -370.6814880371094,
|
| 678 |
+
"logps/rejected": -4141.08740234375,
|
| 679 |
+
"loss": 0.0004,
|
| 680 |
+
"rewards/accuracies": 1.0,
|
| 681 |
+
"rewards/chosen": -1.1442639827728271,
|
| 682 |
+
"rewards/margins": 37.65648651123047,
|
| 683 |
+
"rewards/rejected": -38.800743103027344,
|
| 684 |
+
"step": 440
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 2.4,
|
| 688 |
+
"grad_norm": 0.00024154584051485787,
|
| 689 |
+
"learning_rate": 1.1498188238036862e-06,
|
| 690 |
+
"logits/chosen": 0.644447922706604,
|
| 691 |
+
"logits/rejected": 2.188000440597534,
|
| 692 |
+
"logps/chosen": -363.6133728027344,
|
| 693 |
+
"logps/rejected": -3879.87939453125,
|
| 694 |
+
"loss": 0.0001,
|
| 695 |
+
"rewards/accuracies": 1.0,
|
| 696 |
+
"rewards/chosen": -1.153423547744751,
|
| 697 |
+
"rewards/margins": 35.144954681396484,
|
| 698 |
+
"rewards/rejected": -36.298377990722656,
|
| 699 |
+
"step": 450
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.453333333333333,
|
| 703 |
+
"grad_norm": 0.00016005089303980743,
|
| 704 |
+
"learning_rate": 9.585820233899739e-07,
|
| 705 |
+
"logits/chosen": 0.5130532383918762,
|
| 706 |
+
"logits/rejected": 2.036015033721924,
|
| 707 |
+
"logps/chosen": -363.82366943359375,
|
| 708 |
+
"logps/rejected": -3935.55712890625,
|
| 709 |
+
"loss": 0.0,
|
| 710 |
+
"rewards/accuracies": 1.0,
|
| 711 |
+
"rewards/chosen": -1.1580547094345093,
|
| 712 |
+
"rewards/margins": 35.62702560424805,
|
| 713 |
+
"rewards/rejected": -36.78507995605469,
|
| 714 |
+
"step": 460
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 2.506666666666667,
|
| 718 |
+
"grad_norm": 0.00036818670972699614,
|
| 719 |
+
"learning_rate": 7.830427709355726e-07,
|
| 720 |
+
"logits/chosen": 0.5470496416091919,
|
| 721 |
+
"logits/rejected": 2.2380988597869873,
|
| 722 |
+
"logps/chosen": -364.50726318359375,
|
| 723 |
+
"logps/rejected": -4161.2060546875,
|
| 724 |
+
"loss": 0.0,
|
| 725 |
+
"rewards/accuracies": 1.0,
|
| 726 |
+
"rewards/chosen": -1.1360366344451904,
|
| 727 |
+
"rewards/margins": 37.85028839111328,
|
| 728 |
+
"rewards/rejected": -38.986324310302734,
|
| 729 |
+
"step": 470
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.56,
|
| 733 |
+
"grad_norm": 0.00012922481748816765,
|
| 734 |
+
"learning_rate": 6.238828904562316e-07,
|
| 735 |
+
"logits/chosen": 0.6629059314727783,
|
| 736 |
+
"logits/rejected": 2.3191587924957275,
|
| 737 |
+
"logps/chosen": -363.80462646484375,
|
| 738 |
+
"logps/rejected": -4275.02099609375,
|
| 739 |
+
"loss": 0.0,
|
| 740 |
+
"rewards/accuracies": 1.0,
|
| 741 |
+
"rewards/chosen": -1.0997378826141357,
|
| 742 |
+
"rewards/margins": 38.973384857177734,
|
| 743 |
+
"rewards/rejected": -40.073123931884766,
|
| 744 |
+
"step": 480
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 2.6133333333333333,
|
| 748 |
+
"grad_norm": 5.441705463376079e-05,
|
| 749 |
+
"learning_rate": 4.817205857288176e-07,
|
| 750 |
+
"logits/chosen": 0.45334744453430176,
|
| 751 |
+
"logits/rejected": 1.8800970315933228,
|
| 752 |
+
"logps/chosen": -363.90594482421875,
|
| 753 |
+
"logps/rejected": -3663.01806640625,
|
| 754 |
+
"loss": 0.0,
|
| 755 |
+
"rewards/accuracies": 1.0,
|
| 756 |
+
"rewards/chosen": -1.1525468826293945,
|
| 757 |
+
"rewards/margins": 33.03852081298828,
|
| 758 |
+
"rewards/rejected": -34.19106674194336,
|
| 759 |
+
"step": 490
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.6666666666666665,
|
| 763 |
+
"grad_norm": 5.433719318270229e-05,
|
| 764 |
+
"learning_rate": 3.5710803908373226e-07,
|
| 765 |
+
"logits/chosen": 0.6295723915100098,
|
| 766 |
+
"logits/rejected": 2.0429253578186035,
|
| 767 |
+
"logps/chosen": -356.60894775390625,
|
| 768 |
+
"logps/rejected": -3701.161376953125,
|
| 769 |
+
"loss": 0.0,
|
| 770 |
+
"rewards/accuracies": 1.0,
|
| 771 |
+
"rewards/chosen": -1.0938208103179932,
|
| 772 |
+
"rewards/margins": 33.477996826171875,
|
| 773 |
+
"rewards/rejected": -34.571815490722656,
|
| 774 |
+
"step": 500
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 2.7199999999999998,
|
| 778 |
+
"grad_norm": 0.0008541567247667036,
|
| 779 |
+
"learning_rate": 2.5052926663577006e-07,
|
| 780 |
+
"logits/chosen": 0.5647148489952087,
|
| 781 |
+
"logits/rejected": 2.2748687267303467,
|
| 782 |
+
"logps/chosen": -367.0702209472656,
|
| 783 |
+
"logps/rejected": -4109.466796875,
|
| 784 |
+
"loss": 0.0,
|
| 785 |
+
"rewards/accuracies": 1.0,
|
| 786 |
+
"rewards/chosen": -1.1226731538772583,
|
| 787 |
+
"rewards/margins": 37.33594512939453,
|
| 788 |
+
"rewards/rejected": -38.458614349365234,
|
| 789 |
+
"step": 510
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 2.7733333333333334,
|
| 793 |
+
"grad_norm": 9.149130407162097e-05,
|
| 794 |
+
"learning_rate": 1.6239823828401945e-07,
|
| 795 |
+
"logits/chosen": 0.8614528775215149,
|
| 796 |
+
"logits/rejected": 2.60591721534729,
|
| 797 |
+
"logps/chosen": -346.8439636230469,
|
| 798 |
+
"logps/rejected": -4330.7294921875,
|
| 799 |
+
"loss": 0.0,
|
| 800 |
+
"rewards/accuracies": 1.0,
|
| 801 |
+
"rewards/chosen": -1.1519685983657837,
|
| 802 |
+
"rewards/margins": 39.40824890136719,
|
| 803 |
+
"rewards/rejected": -40.56021499633789,
|
| 804 |
+
"step": 520
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 2.8266666666666667,
|
| 808 |
+
"grad_norm": 6.199788721668637e-05,
|
| 809 |
+
"learning_rate": 9.305726978306173e-08,
|
| 810 |
+
"logits/chosen": 0.4548477530479431,
|
| 811 |
+
"logits/rejected": 1.8193861246109009,
|
| 812 |
+
"logps/chosen": -368.3951721191406,
|
| 813 |
+
"logps/rejected": -3689.09765625,
|
| 814 |
+
"loss": 0.0,
|
| 815 |
+
"rewards/accuracies": 1.0,
|
| 816 |
+
"rewards/chosen": -1.1117929220199585,
|
| 817 |
+
"rewards/margins": 33.348350524902344,
|
| 818 |
+
"rewards/rejected": -34.460140228271484,
|
| 819 |
+
"step": 530
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.88,
|
| 823 |
+
"grad_norm": 0.00048754024723981533,
|
| 824 |
+
"learning_rate": 4.2775693130948094e-08,
|
| 825 |
+
"logits/chosen": 0.601469874382019,
|
| 826 |
+
"logits/rejected": 1.8848406076431274,
|
| 827 |
+
"logps/chosen": -371.9225158691406,
|
| 828 |
+
"logps/rejected": -3655.969482421875,
|
| 829 |
+
"loss": 0.0,
|
| 830 |
+
"rewards/accuracies": 1.0,
|
| 831 |
+
"rewards/chosen": -1.2053592205047607,
|
| 832 |
+
"rewards/margins": 32.930843353271484,
|
| 833 |
+
"rewards/rejected": -34.13620376586914,
|
| 834 |
+
"step": 540
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 2.9333333333333336,
|
| 838 |
+
"grad_norm": 3.647961634896807e-05,
|
| 839 |
+
"learning_rate": 1.174881043835563e-08,
|
| 840 |
+
"logits/chosen": 0.16144999861717224,
|
| 841 |
+
"logits/rejected": 1.4864904880523682,
|
| 842 |
+
"logps/chosen": -379.0067138671875,
|
| 843 |
+
"logps/rejected": -3737.619873046875,
|
| 844 |
+
"loss": 0.0,
|
| 845 |
+
"rewards/accuracies": 1.0,
|
| 846 |
+
"rewards/chosen": -1.1336318254470825,
|
| 847 |
+
"rewards/margins": 33.75095748901367,
|
| 848 |
+
"rewards/rejected": -34.88459014892578,
|
| 849 |
+
"step": 550
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.986666666666667,
|
| 853 |
+
"grad_norm": 0.008316077390120028,
|
| 854 |
+
"learning_rate": 9.713534230904043e-11,
|
| 855 |
+
"logits/chosen": 0.6818611025810242,
|
| 856 |
+
"logits/rejected": 2.075132131576538,
|
| 857 |
+
"logps/chosen": -358.7864685058594,
|
| 858 |
+
"logps/rejected": -3695.71484375,
|
| 859 |
+
"loss": 0.0001,
|
| 860 |
+
"rewards/accuracies": 1.0,
|
| 861 |
+
"rewards/chosen": -1.12270188331604,
|
| 862 |
+
"rewards/margins": 33.42768859863281,
|
| 863 |
+
"rewards/rejected": -34.550392150878906,
|
| 864 |
+
"step": 560
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 2.992,
|
| 868 |
+
"step": 561,
|
| 869 |
+
"total_flos": 0.0,
|
| 870 |
+
"train_loss": 0.049965105635520604,
|
| 871 |
+
"train_runtime": 8901.9362,
|
| 872 |
+
"train_samples_per_second": 4.044,
|
| 873 |
+
"train_steps_per_second": 0.063
|
| 874 |
+
}
|
| 875 |
+
],
|
| 876 |
+
"logging_steps": 10,
|
| 877 |
+
"max_steps": 561,
|
| 878 |
+
"num_input_tokens_seen": 0,
|
| 879 |
+
"num_train_epochs": 3,
|
| 880 |
+
"save_steps": 1000,
|
| 881 |
+
"stateful_callbacks": {
|
| 882 |
+
"TrainerControl": {
|
| 883 |
+
"args": {
|
| 884 |
+
"should_epoch_stop": false,
|
| 885 |
+
"should_evaluate": false,
|
| 886 |
+
"should_log": false,
|
| 887 |
+
"should_save": true,
|
| 888 |
+
"should_training_stop": true
|
| 889 |
+
},
|
| 890 |
+
"attributes": {}
|
| 891 |
+
}
|
| 892 |
+
},
|
| 893 |
+
"total_flos": 0.0,
|
| 894 |
+
"train_batch_size": 4,
|
| 895 |
+
"trial_name": null,
|
| 896 |
+
"trial_params": null
|
| 897 |
+
}
|