UTI2_L3_1000steps_1e5rate_05beta_CSFTDPO

This model is a fine-tuned version of tsavage68/UTI_L3_1000steps_1e5rate_SFT on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6931
• Rewards/chosen: 0.0
• Rewards/rejected: 0.0
• Rewards/accuracies: 0.0
• Rewards/margins: 0.0
• Logps/rejected: 0.0
• Logps/chosen: 0.0
• Logits/rejected: -1.1794
• Logits/chosen: -1.1794

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-05
• 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: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.6931	0.3333	25	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	0.6667	50	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	1.0	75	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	1.3333	100	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	1.6667	125	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	2.0	150	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	2.3333	175	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	2.6667	200	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	3.0	225	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	3.3333	250	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	3.6667	275	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	4.0	300	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	4.3333	325	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	4.6667	350	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	5.0	375	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	5.3333	400	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	5.6667	425	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	6.0	450	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	6.3333	475	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	6.6667	500	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	7.0	525	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	7.3333	550	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	7.6667	575	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	8.0	600	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	8.3333	625	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	8.6667	650	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	9.0	675	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	9.3333	700	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	9.6667	725	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	10.0	750	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	10.3333	775	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	10.6667	800	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	11.0	825	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	11.3333	850	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	11.6667	875	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	12.0	900	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	12.3333	925	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	12.6667	950	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	13.0	975	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794
0.6931	13.3333	1000	0.6931	0.0	0.0	0.0	0.0	0.0	0.0	-1.1794	-1.1794

Framework versions

• Transformers 4.41.2
• Pytorch 2.0.0+cu117
• Datasets 2.19.2
• Tokenizers 0.19.1