File size: 992 Bytes
98e91be | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 | ---
license: mit
library_name: webgpu
base_model: deepgrove/maple-preview
tags:
- webgpu
- browser
- text-generation
- quantized
---
# Maple-Preview WebGPU Pack
Lossless browser-streaming repack of
[`deepgrove/maple-preview-2bit-mlx`](https://huggingface.co/deepgrove/maple-preview-2bit-mlx)
for [`ProCreations/maple-webgpu`](https://huggingface.co/spaces/ProCreations/maple-webgpu).
No model values are changed and no additional quantization is applied. The
official ternary matrices are losslessly interleaved four rows at a time for
vectorized WebGPU dot products, then regrouped into aligned layer files.
- Source revision: `361db5da5e74ff6fcdd852d478e1f266ce11013a`
- Packed tensor bytes: `5.31 GB`
- Container format: `maple-webgpu-pack-v2`
- Ternary matrix layout: `tile4_rows`
- Original model and weights: MIT, © DeepGrove AI
This is a runtime artifact, not a separate model release. See `manifest.json`
for exact byte offsets, shapes, dtypes, and source tensor names.
|