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：1875

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

本站现有博文323篇,共被浏览795409次

本站已经建立2493天!

热门文章
文章归档
回到顶部