| - solver=musicgen/musicgen_base_32khz | |
| - model/lm/model_scale=small | |
| - continue_from=//pretrained/facebook/musicgen-small | |
| - conditioner=text2music | |
| - dset=audio/folk | |
| - dataset.num_workers=2 | |
| - dataset.valid.num_samples=1 | |
| - dataset.batch_size=2 | |
| - schedule.cosine.warmup=8 | |
| - optim.optimizer=adamw | |
| - optim.lr=1e-4 | |
| - optim.epochs=10 | |
| - optim.updates_per_epoch=1000 | |
| - optim.adam.weight_decay=0.01 | |
| - generate.lm.prompted_samples=False | |
| - generate.lm.gen_gt_samples=True | |