Downloads · 30 days
0
htdemucs-6s-webgpu-onnx
kramp/htdemucs-6s-webgpu-onnx
htdemucs-6s-webgpu-onnx is a machine learning model from kramp. 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.
A constant-folded copy of StemSplitio/htdemucs-6s-onnx (htdemucs6s: drums, bass, other, vocals, guitar, piano).
Access
Public
Updated Jun 30, 2026
Repo size
285 MB
Likes
1
Public
Repo makeup
Click a slice to open those files.
.onnx285 MB · 100%
At a glance
- License
- mit
- Access
- Public
- Created
- Jun 30, 2026
- Updated
- Jun 30, 2026
- SHA
- 0c850a01
Base models
Model card
From the Hugging Face model README
htdemucs_6s — WebGPU-compatible ONNX
A constant-folded copy of StemSplitio/htdemucs-6s-onnx
(htdemucs_6s: drums, bass, other, vocals, guitar, piano).
The only change: the original's embedded inverse-STFT contains a ConstantOfShape op that
onnxruntime-web's WebGPU backend cannot run. Because the input window is fixed
(mix [1,2,343980]), that node's output shape is static, so it was folded to a Constant
offline (ONNX Runtime ORT_ENABLE_BASIC). The result is bit-identical to the original
(same mix → stems [1,6,2,343980] I/O) but runs on WebGPU in the browser.
Used by the Audio Split Space. MIT, per the base model.
- License
- mit
- Created
- Jun 30, 2026
- Updated
- Jun 30, 2026