musicxlstm

TristanBehrens/musicxlstm

musicxlstm is a machine learning model from TristanBehrens. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.

Downloads · 30 days

9% of all-time downloads

All-time downloads

Public

Parameters

101K

408 KB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors408 KB · 83%

At a glance

License: mit
Access: Public
Created: Aug 15, 2024
Updated: Aug 15, 2024
SHA: e9af0b35

Base models

None

Datasets

TristanBehrens/js-fakes-4bars

Model card

From the Hugging Face model README

musicxlstm - An xLSTM Model

Trained with Helibrunna by Dr. Tristan Behrens.

Configuration

training:
  model_name: musicxlstm
  batch_size: 256
  lr: 0.001
  lr_warmup_steps: 2000
  lr_decay_until_steps: 20000
  lr_decay_factor: 0.001
  weight_decay: 0.1
  amp_precision: bfloat16
  weight_precision: float32
  enable_mixed_precision: true
  num_epochs: 20
  output_dir: output/musicxlstm
  save_every_step: 100
  log_every_step: 10
  wandb_project: musicxlstm
  tensorboard_dir: runs
model:
  num_blocks: 2
  embedding_dim: 64
  mlstm_block:
    mlstm:
      num_heads: 1
  slstm_block:
    slstm:
      num_heads: 1
  slstm_at:
  - 1
  context_length: 256
  vocab_size: 119
dataset:
  hugging_face_id: TristanBehrens/js-fakes-4bars
tokenizer:
  type: whitespace

License: mit
Languages: en
Created: Aug 15, 2024
Updated: Aug 15, 2024