| library_name: peft | |
| license: apache-2.0 | |
| base_model: Qwen/Qwen3-4B | |
| tags: | |
| - base_model:adapter:Qwen/Qwen3-4B | |
| - lora | |
| - sft | |
| - transformers | |
| - trl | |
| pipeline_tag: text-generation | |
| model-index: | |
| - name: qwen3_4b_hacker | |
| results: [] | |
| <!-- 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_hacker | |
| This model is a fine-tuned version of [Qwen/Qwen3-4B](https://huggingface.co/Qwen/Qwen3-4B) on the None dataset. | |
| ## 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: 0.0002 | |
| - train_batch_size: 4 | |
| - eval_batch_size: 8 | |
| - seed: 42 | |
| - gradient_accumulation_steps: 4 | |
| - total_train_batch_size: 16 | |
| - 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_steps: 100 | |
| - num_epochs: 1 | |
| - mixed_precision_training: Native AMP | |
| ### Training results | |
| ### Framework versions | |
| - PEFT 0.16.0 | |
| - Transformers 4.53.1 | |
| - Pytorch 2.7.0+cu126 | |
| - Datasets 3.6.0 | |
| - Tokenizers 0.21.2 |