model-compression repositories

Trending repositories tagged model-compression, ranked by trending score. Updated daily.

  1. 1
    Scorp1o117/FIT-GGUF

    Fit-to-Size Intelligent Tensor Quantization for GGUF - deterministic, size-exact mixed quantization on llama.cpp presets. Turn any byte budget into a reproducible GGUF.

    GitHub · Python · 24 stars · 3 forks · +5 stars in 24 hours · Trending score 1.19

  2. 2
    angelolamonaca/PyTorch-Precision-Converter

    Convert PyTorch checkpoints, safetensors files and whole Hugging Face model folders to fp16, bf16, fp8 or ComfyUI fp8_scaled, int8_convrot and int8, turn FP8 models (DeepSeek, Qwen) back into BF16 on a CPU, merge or split sharded safetensors, and extract EMA weights, with one command and without loading the model into RAM.

    GitHub · Python · 15 stars · 0 forks · +3 stars in 24 hours · Trending score 0.92

  3. 3
    pinoox/neuromesh

    The Biomimetic Context Engine & Neural Runtime for AI Coding Assistants

    GitHub · Rust · 93 stars · 8 forks · +1 stars in 24 hours · Trending score 0.45

  4. 4
    shootthesound/torch-nvenc-compress

    Use the GPU's idle NVENC silicon to multiply the effective bandwidth of every PCIe wire on the board

    GitHub · Python · 32 stars · 5 forks · 0 stars in 24 hours · Trending score 0.42

  5. 5
    wangxb96/Awesome-EdgeAI

    Resources of our survey paper "Optimizing Edge AI: A Comprehensive Survey on Data, Model, and System Strategies" and "Cognitive Edge Computing: A Comprehensive Survey on Optimizing Large Models and AI Agents for Pervasive Deployment"

    GitHub · Python · 122 stars · 22 forks · 0 stars in 24 hours · Trending score 0.29

  6. 6
    rajveer43/VeloxQuant-MLX

    Fast KV-cache quantization for Apple Silicon (MLX) — 43 research-adapted compression methods with Metal kernels

    GitHub · Python · 36 stars · 10 forks · 0 stars in 24 hours · Trending score 0.23

  7. 7
    Zishan-Shao/lowrankarena

    First standardized benchmark for SVD-compressed LLMs up to 70B, with 5TiB+ reproduced open-sourced checkpoints (2,000+ GPU hours).

    GitHub · Python · 12 stars · 0 forks · 0 stars in 24 hours · Trending score 0.22

  8. 8
    pjmalandrino/llvq

    Tetra : Leech lattice vector quantization for 2-bit LLM weights, in Rust, with a fused multi-shell CUDA decoder.

    GitHub · Rust · 5 stars · 1 forks · 0 stars in 24 hours · Trending score 0.17

  9. 9
    Fazmin/iQuantize

    Desktop workspace for quantizing LLMs to GGUF, MLX and ONNX. Watch every tensor change, compare against the original, and publish to Hugging Face with an auto-written model card.

    GitHub · JavaScript · 16 stars · 1 forks · 0 stars in 24 hours · Trending score 0.14

  10. 10
    aymen94/NeuronDump

    Look inside any local LLM (GGUF from Ollama, LM Studio, Hugging Face, Jan, llama.cpp): layers, heads, vocabulary, chat template, weights. Explore it as a 3D graph, follow a token through every layer, edit it live (ablate, steer, ban, patch), diff two models, then bake or slim a new GGUF that still runs. CLI, web UI, browser demo. Runs 100% offline

    GitHub · Rust · 20 stars · 3 forks · 0 stars in 24 hours · Trending score 0.08

  11. 11
    peonist-ai/crucible

    MoE expert compression toolkit — prune, merge, and forge specialist models for local inference.

    GitHub · Python · 14 stars · 0 forks · 0 stars in 24 hours · Trending score 0.04

  12. 12
    Picovoice/picollm

    On-device LLM Inference Powered by X-Bit Quantization

    GitHub · Python · 318 stars · 26 forks · −1 stars in 24 hours · Trending score 0.03

  13. 13
    egesabanci/reap-mlx

    MLX-compatible REAP for pruning MoE models on Apple Silicon

    GitHub · Python · 43 stars · 0 forks · 0 stars in 24 hours · Trending score 0.00

  14. 14
    MessiX77/Awesome-Efficient-dLLMs

    [EMNLP 2026] 📚 A curated list of Awesome Efficient dLLMs Papers with Codes

    GitHub · 147 stars · 3 forks · 0 stars in 24 hours · Trending score 0.00

  15. 15
    CASE-Lab-UMD/Unified-MoE-Compression

    The official implementation of the paper "Towards Efficient Mixture of Experts: A Holistic Study of Compression Techniques (TMLR)".

    GitHub · Python · 93 stars · 7 forks · 0 stars in 24 hours · Trending score 0.00

  16. 16
    CASE-Lab-UMD/LLM-Drop

    The official implementation of the paper "Uncovering the Redundancy in Transformers via a Unified Study of Layer Dropping (TMLR)".

    GitHub · Python · 193 stars · 23 forks · 0 stars in 24 hours · Trending score 0.00

  17. 17
    nick7nlp/RAZOR

    RAZOR: Pruning Replaceable Experts in LLMs — training-free expert pruning for Mixture-of-Experts (MoE) models, with the MoECal calibration corpus

    GitHub · Python · 7 stars · 1 forks · 0 stars in 24 hours · Trending score 0.00