Maple-Preview WebGPU Pack

Lossless browser-streaming repack of deepgrove/maple-preview-2bit-mlx for 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, ยฉ 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.

Downloads last month
16
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for ProCreations/maple-preview-webgpu-v2

Quantized
(16)
this model