metadata
base_model: facebook/musicgen-large
library_name: peft
license: cc-by-nc-4.0
tags:
- text-to-audio
- ylacombe/tiny-punk
- generated_from_trainer
model-index:
- name: punk-music-test
results: []
punk-music-test
This model is a fine-tuned version of facebook/musicgen-large on the YLACOMBE/TINY-PUNK - DEFAULT dataset. It achieves the following results on the evaluation set:
- Loss: 4.6589
- Clap: 0.0487
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: 4
- eval_batch_size: 1
- seed: 123
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.99) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Clap |
|---|---|---|---|---|
| 8.875 | 0.8571 | 3 | 4.4074 | 0.0678 |
| 8.7074 | 2.0 | 7 | 4.4241 | 0.0586 |
| 8.6372 | 2.8571 | 10 | 4.4617 | 0.0704 |
| 8.5145 | 4.0 | 14 | 4.5377 | 0.0566 |
| 8.4215 | 4.8571 | 17 | 4.5900 | 0.0567 |
| 8.4121 | 6.0 | 21 | 4.6385 | 0.0320 |
| 8.377 | 6.8571 | 24 | 4.6589 | 0.0539 |
Framework versions
- PEFT 0.11.1
- Transformers 4.42.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.20.0
- Tokenizers 0.19.1