AI & ML interests
Model Compression
Recent Activity
Organization Card
Welcome to AngelSlim 👋
AngelSlim, developed by Tencent, is a large language model compression toolkit engineered for enhanced usability, comprehensiveness, and efficiency. We will open-source compression algorithms including quantization, speculative decoding, pruning, and distillation. Supporting cutting-edge mainstream LLMs, the toolkit streamlines the complete end-to-end workflow from compression to deployment.
（AngelSlim是腾讯自研的，致力于打造更易用、更全面和更高效的大语言模型压缩工具包。我们将开源量化、投机采样、稀疏化和蒸馏等压缩算法。覆盖主流最前沿的大模型，并且端到端打通从压缩到部署的全流程。）
models 97
AngelSlim/Hy-MT2-1.8B-1.25Bit-GGUF
2B • Updated • 275
AngelSlim/Hy-MT2-1.8B-2Bit-GGUF
2B • Updated • 207
AngelSlim/Hy-MT1.5-1.8B-1.25bit-GGUF
Translation • 2B • Updated • 9.29k • 38
AngelSlim/Hy-MT1.5-1.8B-1.25bit
Translation • Updated • 18k • 181
AngelSlim/Hy-MT1.5-1.8B-2bit
Translation • 2B • Updated • 1.03k • 10
AngelSlim/Hy-MT1.5-1.8B-2bit-GGUF
Translation • 2B • Updated • 6.53k • 20
AngelSlim/Hy-MT-demo
Updated • 1
AngelSlim/Qwen2.5-VL-7B-Instruct-Selector
Updated • 39
AngelSlim/Qwen2.5-VL-3B-Instruct-Selector
Updated • 21
AngelSlim/Qwen2.5-VL-32B-Instruct-Selector
Updated • 42
datasets 0
None public yet