-
ModelBest
- HANGZHOU
Pinned Loading
-
Seedance2.0-Storyboard-Planner
Seedance2.0-Storyboard-Planner PublicSeedance 2.0 素材分镜规划器 — 从创意规划到视频生成的一站式前期制作工具
-
flash-attn-blackwell
flash-attn-blackwell PublicPre-built Flash Attention wheels for NVIDIA RTX 5090 / 5080 / 5070 Ti (Blackwell sm_120). pip install flash-attn doesn't ship sm_120 binaries — this repo fixes that.
Python 3
-
pocketclaw
pocketclaw Public🦀 Mobile-first AI agent framework. Your butler, in your pocket. OpenClaw alternative for 6.8B phone users.
Kotlin 14
-
turboquant-pytorch
turboquant-pytorch PublicForked from tonbistudio/turboquant-pytorch
From-scratch PyTorch implementation of Google's TurboQuant (ICLR 2026) for LLM KV cache compression. 5x compression at 3-bit with 99.5% attention fidelity.
Python 3
-
CallCenter-VoiceAgent
CallCenter-VoiceAgent PublicFull-Duplex Voice Agent — ASR+RAG+LLM+TTS with voice cloning
Python 1
-
RepEngvLLM
RepEngvLLM PublicRepEng × vLLM: Real-time LLM Personality Steering on MiniCPM4.1-8B — 5 orthogonal behavioral dimensions, INT4 compatible, vLLM-ready GGUF vectors
Python
If the problem persists, check the GitHub status page or contact support.
