Model Card for midi-model-tv2o-medium

Midi event transformer for music generation.

Model Updates

  • new tv2o-medium model: "tvo" means MidiTokenizerV2 optimised.

Model Details

Model Description

  • Developed by: SkyTNT
  • Trained by: SkyTNT
  • Model type: Transformer
  • License: apache-2.0

Model Sources

Training Details

Training Data

Training Hyperparameters

  • config: tv2o-medium
  • max-len: 2048 then 4096
  • lr: 1e-4 then 2e-4
  • weight-decay: 0.01
  • batch: 2x2x2
  • bfloat16 precision

Loss

  • val loss: 0.2159

Files

  • model.ckpt: latest model checkpoint
  • onnx/*.onnx: onnx format model
Downloads last month
1,005
Safetensors
Model size
234M params
Tensor type
BF16
ยท
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no pipeline_tag.

Model tree for skytnt/midi-model-tv2o-medium

Adapters
2 models

Datasets used to train skytnt/midi-model-tv2o-medium

Spaces using skytnt/midi-model-tv2o-medium 11