metadata
library_name: transformers
license: llama3
base_model: meta-llama/Meta-Llama-3-8B-Instruct
tags:
- alignment-handbook
- trl
- simpo
- generated_from_trainer
- trl
- simpo
- generated_from_trainer
datasets:
- yakazimir/llama3-ultrafeedback-armorm
model-index:
- name: llama3_l5_best_entropy
results: []
llama3_l5_best_entropy
This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the yakazimir/llama3-ultrafeedback-armorm dataset. It achieves the following results on the evaluation set:
- Loss: 1.4398
- Rewards/chosen: -8.4238
- Rewards/rejected: -21.7089
- Rewards/accuracies: 0.8795
- Rewards/margins: 13.2850
- Logps/rejected: -2.1709
- Logps/chosen: -0.8424
- Logits/rejected: -1.4192
- Logits/chosen: -1.5108
- Semantic Entropy: 0.8327
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-06
- train_batch_size: 2
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 16
- total_train_batch_size: 128
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 2.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen | Semantic Entropy |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1.5778 | 0.8743 | 400 | 1.6373 | -8.6865 | -18.8603 | 0.8735 | 10.1738 | -1.8860 | -0.8687 | -1.4323 | -1.5078 | 0.8519 |
| 0.9552 | 1.7486 | 800 | 1.4402 | -8.2804 | -21.2503 | 0.8795 | 12.9699 | -2.1250 | -0.8280 | -1.4434 | -1.5360 | 0.8377 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.2.2+cu121
- Datasets 2.18.0
- Tokenizers 0.19.1