Deepseek-R1-Distill-Qwen-14B-peft-p-tuning

This model is a fine-tuned version of deepseek-ai/DeepSeek-R1-Distill-Qwen-14B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1008
• Accuracy: 0.9711
• Precision: 0.8926
• Recall: 0.9234
• F1: 0.9077

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: 0.001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.2967	1.0	1477	0.1583	0.9399	0.8841	0.7011	0.7821
0.1192	2.0	2954	0.2277	0.9216	0.9267	0.5326	0.6764
0.1061	3.0	4431	0.0979	0.9705	0.8893	0.9234	0.9060
0.0733	4.0	5908	0.0953	0.9664	0.8953	0.8851	0.8902
0.0547	5.0	7385	0.1008	0.9711	0.8926	0.9234	0.9077

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.1