okai-musiclang-content-t5-small_finetune

This model is a fine-tuned version of sandernotenbaert/okai-musiclang-content-t5-small_finetune on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.5041

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAFACTOR and the args are: No additional optimizer arguments
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_ratio: 0.02
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.7839	0.2226	500	1.6513
1.6576	0.4452	1000	1.6615
1.6396	0.6679	1500	1.6650
1.7168	0.8905	2000	1.6315
1.7366	1.1131	2500	1.6234
1.7171	1.3357	3000	1.6028
1.6238	1.5583	3500	1.6130
1.6217	1.7810	4000	1.6218
1.7077	2.0036	4500	1.5784
1.7034	2.2262	5000	1.5792
1.6049	2.4488	5500	1.5866
1.6018	2.6714	6000	1.5869
1.6628	2.8941	6500	1.5628
1.653	3.1171	7000	1.5606
1.6575	3.3397	7500	1.5381
1.64	3.5619	8000	1.5395
1.6455	3.7845	8500	1.5163
1.6308	4.0076	9000	1.5311
1.6324	4.2302	9500	1.5118
1.5481	4.4528	10000	1.5092
1.547	4.6754	10500	1.5109
1.5584	4.8981	11000	1.5041

Framework versions

• Transformers 4.55.0
• Pytorch 2.6.0+cu124
• Datasets 4.0.0
• Tokenizers 0.21.2