EADST

Quick Review: SpQR: A Sparse-Quantized Representation for Near-Lossless LLM Weight Compression

作者：XD / 发表： 2023年12月6日 23:57 / 更新： 2023年12月6日 23:57 / 科研学习 / 阅读量：2277

SpQR: A Sparse-Quantized Representation for Near-Lossless Large Language Model Weight Compression

Core Approach:

  • GPTQ without Outliers: Focuses on eliminating outliers during the GPTQ process, enabling more efficient and accurate weight compression for large language models.

本文作者：XD 转载请标明出处：http://www.eadst.com/blog/225

本站采用知识共享署名-非商业性使用-相同方式共享 4.0 国际许可协议进行许可。

上一篇
Quick Review: Optimize Weight Rounding via Signed Gradient Descent for the Quantization of LLMs
下一篇
Quick Review: QUIK: Towards End-to-end 4-Bit Inference on Generative Large Language Models
相关标签
LLM Quantization
About Me
XD
Goals determine what you are going to be.
Category
标签云
Interview Tiktoken CLAP Hilton Bin Proxy scipy Domain Excel git-lfs WAN 证件照 Github Plotly printf Pillow Baidu InvalidArgumentError Data SVR Jev VPN Disk Distillation Diagram Firewall ResNet-50 Card 云服务器 递归学习法 XGBoost API Safetensors 腾讯云 净利润 Numpy COCO C++ 音频 MD5 SQLite Git TSV Jupyter JSON Input Paper Python Magnet CSV LLAMA Anaconda CV DeepSeek IndexTTS2 公式 RAR uWSGI Hungarian ms-swift PIP TTS Qwen Michelin FP16 Quantize CAM GoogLeNet Video 报税 YOLO icon EXCEL Linux FlashAttention logger Vmess 算法题 GPT4 Random Rebuttal Web Miniforge Streamlit Breakpoint 图标 Conda TensorFlow Transformers Math 论文 Crawler UNIX Attention HaggingFace tar OpenCV Ptyhon VSCode Logo Cloudreve Hotel NLP Jetson PDF mmap Qwen2.5 SPIE CTC Heatmap Use Django llama.cpp LLM News API网关 Vim PyTorch 图形思考法 Website Knowledge SAM tqdm Pytorch QWEN 财报 CEIR GPTQ Password NLTK 版权 PDB ModelScope FastAPI XML 顶会 HuggingFace v2ray Harness git 第一性原理 Pandas Tensor 多进程 飞书 Review diffusers PyCharm Animate Food Freesound Windows UI AI FP64 Bitcoin Statistics Color transformers Search hf Llama Google LoRA LeetCode BeautifulSoup Image2Text WebCrawler Dataset Zip Template TensorRT RL torchinfo Bert SQL 多线程 BTC Agent VGG-16 BF16 域名 阿里云 CC Gemma Shortcut 论文速读 v0.dev Claude DeepStream Base64 Permission Qwen2 Docker Bipartite Datetime Quantization ONNX Sklearn CUDA Augmentation Nginx Plate Tracking 关于博主 强化学习 Mixtral Algorithm ChatGPT NameSilo Land GGML Ubuntu RGB OCR GIT Markdown Paddle Pickle uwsgi Clash Translation 搞笑 继承 LaTeX FP32 OpenAI 签证 FP8
站点统计

本站现有博文338篇,共被浏览966643次

本站已经建立2679天!

热门文章
文章归档
回到顶部