SFT-Qwen2.5-Coder-3B_long_v1

This model is a fine-tuned version of Qwen/Qwen2.5-Coder-3B-Instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7569

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: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT 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: 0.03
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
0.9906	0.2807	20	0.9487
0.8528	0.5614	40	0.8620
0.8721	0.8421	60	0.8238
0.8059	1.1123	80	0.8018
0.8141	1.3930	100	0.7868
0.7353	1.6737	120	0.7767
0.6779	1.9544	140	0.7647
0.6273	2.2246	160	0.7629
0.6983	2.5053	180	0.7597
0.6958	2.7860	200	0.7569

Framework versions

• PEFT 0.18.0
• Transformers 4.57.1
• Pytorch 2.9.0+cu126
• Datasets 4.4.1
• Tokenizers 0.22.1