Hugging Face's logo

j-11045
/
musicgen-medium

Text-to-Audio
PEFT
Safetensors
musicgen
j-11045/indian-music-with-metadata-2
Generated from Trainer
Model card Files Community
musicgen-medium / README.md
j-11045's picture
j-11045
End of training
7c718bc verified
preview code
|
raw
history blame contribute delete
1.83 kB
metadata
base_model: facebook/musicgen-large
library_name: peft
license: cc-by-nc-4.0
tags:
  - text-to-audio
  - j-11045/indian-music-with-metadata-2
  - generated_from_trainer
model-index:
  - name: musicgen-medium
    results: []

musicgen-medium

This model is a fine-tuned version of facebook/musicgen-large on the J-11045/INDIAN-MUSIC-WITH-METADATA-2 - DEFAULT dataset. It achieves the following results on the evaluation set:

  • Loss: 5.8703
  • Clap: -0.0918

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: 2
  • eval_batch_size: 1
  • seed: 456
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 16
  • optimizer: Use adamw_torch with betas=(0.9,0.99) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • num_epochs: 2.0
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Clap
42.3441 0.4739 25 11.2322 0.1068
37.4697 0.9479 50 5.8964 -0.0773
37.8209 1.4265 75 5.9483 -0.1071
36.2328 1.9005 100 5.8852 -0.1152

Framework versions

  • PEFT 0.13.2
  • Transformers 4.46.0.dev0
  • Pytorch 2.1.2+cu121
  • Datasets 3.0.2
  • Tokenizers 0.20.1