Speecht5 Finetuned Gcr 10

--- library_name: transformers language: - gcr license: mit base_model: ckandemir/speecht5_finetuned_voxpopuli_fr tags: - generated_from_trainer model-index: - name: TTS SpeechT5 GCR results: [] --- # TTS SpeechT5 GCR This model is a fine-tuned version of [ckandemir/speecht5_finetuned_voxpopuli_fr](https://huggingface.co/ckandemir/speecht5_finetuned_voxpopuli_fr) on the GuianeseCreole dataset. It achieves the following results on the evaluation set: - Loss: 0.4737 ## 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: 1e-05 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 32 - 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_steps: 100 - training_steps: 900 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-------:|:----:|:---------------:| | 0.6881 | 3.4622 | 100 | 0.5387 | | 0.5268 | 6.9244 | 200 | 0.4970 | | 0.4991 | 10.3556 | 300 | 0.4897 | | 0.4916 | 13.8178 | 400 | 0.4779 | | 0.4782 | 17.2489 | 500 | 0.4770 | | 0.4686 | 20.7111 | 600 | 0.4779 | | 0.4668 | 24.1422 | 700 | 0.4771 | | 0.456 | 27.6044 | 800 | 0.4695 | | 0.4556 | 31.0356 | 900 | 0.4737 | ### Framework versions - Transformers 4.52.0 - Pytorch 2.11.0+cu128 - Datasets 3.6.0 - Tokenizers 0.21.4