ZHLiu627/zephyr-gemma-rpo

gemma_rdpo_eta0.005_no_decay

This model is a fine-tuned version of HuggingFaceH4/zephyr-7b-gemma-sft-v0.1 on the argilla/dpo-mix-7k dataset. It achieves the following results on the evaluation set:

• Loss: 0.5899
• Rewards/chosen: -0.7561
• Rewards/rejected: -2.1676
• Rewards/accuracies: 0.7234
• Rewards/margins: 1.4115
• Logps/rejected: -464.4193
• Logps/chosen: -408.9966
• Logits/rejected: 161.6372
• Logits/chosen: 163.3166
• Eta: 0.0050

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: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 16
• total_train_batch_size: 128
• total_eval_batch_size: 8
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen	Eta
0.5915	0.9479	50	0.5745	-0.7212	-1.8546	0.7021	1.1334	-458.1599	-408.3000	171.6370	174.0368	0.0050
0.2599	1.8957	100	0.5899	-0.7561	-2.1676	0.7234	1.4115	-464.4193	-408.9966	161.6372	163.3166	0.0050

Framework versions

• Transformers 4.40.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1