ConvRot W4A4 kernels and reproducible benchmarks for Apple Silicon and MLX
-
Updated
Aug 24, 2026 - Python
#
w4a4
Here are 2 public repositories matching this topic...
Native mixed-precision quantization for ComfyUI models and text encoders W4A8, W4A4, INT8, FP8, streaming conversion, architecture-aware policies, and validation.
machine-learning cuda pytorch nvidia quantization int8 model-compression mixed-precision diffusion-models model-quantization text-encoder llm fp8 generative-ai safetensors comfyui qwen w4a8 w4a4
-
Updated
Sep 19, 2026 - Python
Add this topic to your repo
To associate your repository with the w4a4 topic, visit your repo's landing page and select "manage topics."