Na_M2_100steps_1e7rate_03beta_cSFTDPO

This model is a fine-tuned version of tsavage68/Na_M2_1000steps_1e7_SFT on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0000
• Rewards/chosen: 2.8383
• Rewards/rejected: -9.2541
• Rewards/accuracies: 1.0
• Rewards/margins: 12.0924
• Logps/rejected: -110.7703
• Logps/chosen: -38.6713
• Logits/rejected: -2.5103
• Logits/chosen: -2.5250

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-07
• train_batch_size: 2
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• training_steps: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.0004	0.2667	50	0.0000	2.3360	-8.0685	1.0	10.4045	-106.8185	-40.3458	-2.5169	-2.5309
0.0	0.5333	100	0.0000	2.8383	-9.2541	1.0	12.0924	-110.7703	-38.6713	-2.5103	-2.5250

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1