pairwise-reward-sft-zephyr-7b-sft-qlora-ultrafeedback-ultrafeedback-binarized-20241013-124646

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4739
• Accuracy: 0.7592

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: 1.5e-05
• train_batch_size: 16
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6209	0.0526	100	0.6427	0.6784
0.6346	0.1052	200	0.5829	0.7165
0.5945	0.1578	300	0.5333	0.7351
0.5258	0.2104	400	0.5169	0.7461
0.4914	0.2630	500	0.5209	0.7346
0.4995	0.3155	600	0.5056	0.7536
0.5272	0.3681	700	0.5041	0.7541
0.4993	0.4207	800	0.4943	0.7471
0.5317	0.4733	900	0.4970	0.7602
0.5193	0.5259	1000	0.4850	0.7597
0.4534	0.5785	1100	0.4931	0.7582
0.4828	0.6311	1200	0.4808	0.7582
0.5432	0.6837	1300	0.4836	0.7491
0.4343	0.7363	1400	0.4797	0.7582
0.4287	0.7889	1500	0.4794	0.7612
0.5117	0.8414	1600	0.4799	0.7587
0.4369	0.8940	1700	0.4770	0.7582
0.4537	0.9466	1800	0.4750	0.7566
0.451	0.9992	1900	0.4739	0.7592

Framework versions

• PEFT 0.12.0
• Transformers 4.45.2
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.20.0