Skip to content

History / AWQ

Revisions

  • Sync wiki with main through eb90dbdd5 Covers merged PRs #3478, #3483, #3484, #3485: - #3478 ROCm RDNA4: gfx950/gfx120*/gfx125* fused GEMV+small-M GEMM with per-arch launch tables in utils/arch.py, ROCm>=10.1 floor, rocBLAS default on CDNA+RDNA4, AWQ GEMV/QQQ/VPTQ_CUDA/EXL3 explicit ROCm rejections, ParoQuant Triton + FP8 ROCm support, FP4 not advertised on ROCm, compute_device_filter honored by replicas and resolve_pack_device (Installation, Inference-Backends, Kernel-Backends-Reference, FP8-and-NVFP4, AWQ, QQQ, ParoQuant, VPTQ, EXL3, Memory-Offload, Troubleshooting) - #3485 Qwen-Omni transformers 5.22 thinker.model.language_model layout: dual module_tree variants + thinker_* helpers (Multimodal) - #3483 bounded remote image fetches: load_image_from_url timeouts, retries, byte/pixel caps, ImageFetchError, env overrides (Multimodal, Troubleshooting) - #3484 MoE early-stop: run_parallel skips empty batches instead of raising (Multimodal, Memory-Offload)

    @Qubitium Qubitium committed Nov 2, 2026
    d918e57

  • Sync wiki with main through 73cda73 Covers merged PRs #3459, #3469, #3479, #3480, #3481: - #3479 auto GPU peer access: new P2P section (env vars, probe diagnostics, benchmark script) in Memory-Offload-and-Multi-GPU - #3480 AVX512 CPU kernels now trainable: Kernel-Backends-Reference, AWQ, Inference-Backends (AUTO prefers trainable CPU kernels) - #3459/#3481 W4A staging: FP8-and-NVFP4 documents per-forward E4M3 operand, NVFP4 layout revalidation, format-2 bookkeeping - #3481 hf_pack_layer zero-point bookkeeping: HuggingFace-Integration, Packing-Formats-and-Compatibility (per-field v1/v2 shifts) - #3469 lazy kernel registry: Developer-Guide

    @Qubitium Qubitium committed Nov 2, 2026
    041ce2b

  • Rename Checkpoint-Formats page to Packing-Formats; add Complementary configs nav group - Checkpoint-Formats-and-Compatibility -> Packing-Formats-and-Compatibility (page file renamed; every inbound link retargeted and retitled). - Sidebar + Home: move Rotations/Smoothing/Norm Tuning and EoRA out of Quantization methods into a new Complementary configs group, with deep links to MSEConfig (mse=) and SmootherConfig in QuantizeConfig-Reference. These configs attach to multiple lifecycles rather than being methods.

    @Qubitium Qubitium committed Oct 29, 2026
    1ff37f6

  • Rename project references: GPTQModel -> GPT-QModel

    @Qubitium Qubitium committed Oct 28, 2026
    41b63e8

  • Sync wiki with main through c1976918

    @Qubitium Qubitium committed Oct 27, 2026
    3ec5a42

  • Add workflow SVGs and refresh documentation navigation and accuracy

    @Qubitium Qubitium committed Oct 26, 2026
    583550e

  • Full documentation site: concepts, all quant methods, config reference, lifecycle, backends, internals

    @Qubitium Qubitium committed Oct 26, 2026
    8b2e8a7