Popular repositories Loading

1. GLM-5.3-Flash-2x-DGX-Spark GLM-5.3-Flash-2x-DGX-Spark Public

   First GLM-5.3-Flash on DGX Spark (2x GB10): 24.7/30.3/19.6 tok/s with MTP-5. NoPE-MLA zero-pad mod + marlin MoE on the stock vLLM image. Tested, measured, honest numbers.

   Python 5

2. strix-halo-quant-lab strix-halo-quant-lab Public

   Run large LLMs locally on AMD Ryzen AI Max+ 395 (Strix Halo, gfx1151) with ROCmFP4 4-bit quantization. Measured benchmarks, build + serving recipes, and 55 ready-to-run GGUF models.

   Python 1

3. treebook treebook Public

   Ruby

4. ROCmFPX ROCmFPX Public

   Forked from charlie12345/ROCmFPX

   ROCmFPX Family for AMD Hardware and Processors. More quants and special agent quants

   C++