zR
@zRdianjiao
Algorithm Engineer @Zai_org — GLM model research & OSS adaptation · Issues & PRs welcome
61 Following 3.7K Followers
So glad to see the community shipping work like GLM-5.2-Vision-NVFP4 — bolting Kimi's MoonViT onto GLM-5.2 with just a 49.5M projector and serving it on SGLang — this kind of dedication is exactly what makes the GLM ecosystem thrive 🚀
Show more
Here is the reference doc with an example of using Claude Code with Hugging Face Inference Providers:
GLM-5.2 is now selectable in Claude Code via Hugging Face🤗 Inference Providers + hf-claude.
Open models are becoming easier to plug directly into real developer workflows. 😀
Show more
In just days, the community made GLM-5.2 runnable almost anywhere — GGUF for llama.cpp & Mac, MLX for Apple Silicon, NVFP4 / W4AFP8 / MXFP4 for NVIDIA & AMD, plus REAP-pruned builds for tighter footprints.
Huge thanks to everyone quantizing, pruning & testing.❤️
Find the build that fits your hardware 👇
Show more
Stopped by the @huggingface Paris office during GOSIM.
Lowkey for the open-source AI leader — hidden in the city center. People writing code wherever they feel like sitting.
🤗 keeps shipping the best open work. This vibe is half the answer.
Show more
