metadata

license: apache-2.0
base_model: alignment-handbook/zephyr-7b-sft-full
tags:
  - alignment-handbook
  - trl
  - dpo
  - generated_from_trainer
  - trl
  - dpo
  - alignment-handbook
  - generated_from_trainer
datasets:
  - HuggingFaceH4/ultrafeedback_binarized
model-index:
  - name: zephyr-7b-dpo-full-ultrabin-high-margin
    results: []

zephyr-7b-dpo-full-ultrabin-high-margin

This model is a fine-tuned version of alignment-handbook/zephyr-7b-sft-full on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:

Loss: 0.5565
Rewards/chosen: -0.6538
Rewards/rejected: -1.5442
Rewards/accuracies: 0.7578
Rewards/margins: 0.8904
Logps/rejected: -417.0811
Logps/chosen: -328.0004
Logits/rejected: 0.3127
Logits/chosen: -0.1625

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: 5e-07
train_batch_size: 8
eval_batch_size: 8
seed: 55
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 2
total_train_batch_size: 128
total_eval_batch_size: 64
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.3228	0.6969	100	0.5641	-0.9601	-1.8247	0.7539	0.8645	-445.1261	-358.6317	0.5872	0.2254

Framework versions

Transformers 4.41.2
Pytorch 2.1.2
Datasets 2.20.0
Tokenizers 0.19.1