musicgen-synths-mono

This model is a fine-tuned version of facebook/musicgen-small on the TRULYSENYA/SYNTH-SOUNDS-SLICE-MONO - DEFAULT dataset. It achieves the following results on the evaluation set:

• Loss: 2.7421
• Clap: 0.0827

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: Adam with betas=(0.9,0.99) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Clap
4.0612	0.6667	25	2.5741	0.1230
3.1723	1.3333	50	2.3362	0.0810
3.2968	2.0	75	2.7076	0.0155
3.423	2.6667	100	2.6853	0.0994
3.0688	3.3333	125	2.7360	0.1210

Framework versions

• PEFT 0.10.0
• Transformers 4.41.0
• Pytorch 2.9.1+cu128
• Datasets 4.4.1
• Tokenizers 0.19.1