voidful/auv-codec-librispeech

1---2library_name: peft3license: apache-2.04base_model: Qwen/Qwen3-4B-Instruct-25075tags:6- generated_from_trainer7model-index:8- name: auv-codec-librispeech9  results: []10---11 12<!-- This model card has been generated automatically according to the information the Trainer had access to. You13should probably proofread and complete it, then remove this comment. -->14 15# auv-codec-librispeech16 17This model is a fine-tuned version of [Qwen/Qwen3-4B-Instruct-2507](https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507) on an unknown dataset.18It achieves the following results on the evaluation set:19- Loss: 11.969920 21## Model description22 23More information needed24 25## Intended uses & limitations26 27More information needed28 29## Training and evaluation data30 31More information needed32 33## Training procedure34 35### Training hyperparameters36 37The following hyperparameters were used during training:38- learning_rate: 0.000139- train_batch_size: 140- eval_batch_size: 141- seed: 4242- gradient_accumulation_steps: 843- total_train_batch_size: 844- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments45- lr_scheduler_type: cosine46- lr_scheduler_warmup_ratio: 0.147- num_epochs: 348 49### Training results50 51| Training Loss | Epoch | Step   | Validation Loss |52|:-------------:|:-----:|:------:|:---------------:|53| 11.9929       | 1.0   | 35156  | 11.9689         |54| 11.992        | 2.0   | 70312  | 11.9734         |55| 11.9116       | 3.0   | 105468 | 11.9699         |56 57 58### Framework versions59 60- PEFT 0.15.061- Transformers 4.57.362- Pytorch 2.8.0+cu12663- Datasets 4.0.064- Tokenizers 0.22.0