OpenELM-1_1B-DPO

This model is a fine-tuned version of data/OpenELM-1_1B-SFT on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:

• Loss: 1.0009
• Rewards/chosen: -14.5625
• Rewards/rejected: -17.125
• Rewards/accuracies: 0.7188
• Rewards/margins: 2.6094
• Logps/rejected: -632.0
• Logps/chosen: -612.0
• Logits/rejected: -13.0
• Logits/chosen: -13.0

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-05
• train_batch_size: 8
• eval_batch_size: 16
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• 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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.7494	1.0	1911	0.7879	-11.6875	-12.8125	0.6797	1.1797	-548.0	-552.0	-12.8125	-12.5625
0.0999	2.0	3822	0.8019	-14.125	-15.9375	0.6992	1.8125	-608.0	-604.0	-13.0625	-13.0
0.011	3.0	5733	1.0009	-14.5625	-17.125	0.7188	2.6094	-632.0	-612.0	-13.0	-13.0

Framework versions

• Transformers 4.44.2
• Pytorch 2.3.0
• Datasets 2.21.0
• Tokenizers 0.19.1