Qwen3-4B_Paper_Impact_code_SFT_1ep

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

Qwen3-4B_Paper_Impact_code_SFT_1ep

This model is a fine-tuned version of Qwen/Qwen3-4B-Instruct-2507 on the paper_impact_code_train dataset. It achieves the following results on the evaluation set:

• Loss: 0.0773

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: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1.0

Training results

Framework versions

• Transformers 4.57.1
• Pytorch 2.6.0+cu124
• Datasets 4.5.0
• Tokenizers 0.22.1