Hugging Face's logo
gbhf's picture
šŸ”„ In a Training Loop

gbhf

gbuzhf
84 10 47
TatshSiow's profile picture Meinh's profile picture Kaytheist's profile picture
Ā·
https://github.com/gbgh1/context-governor

AI & ML interests

None yet

Recent Activity

reacted to danielhanchen's post with šŸ”„ about 3 hours ago
Run DeepSeek V4.1-Flash-1017 locally! šŸ³ We made 1-bit retain ~88.2% accuracy while shrinking size from 508GB to 250GB (āˆ’52%). Run on a 128GB+ RAM device. V4.1 Flash 1017 is the strongest open model to date Guide: https://unsloth.ai/docs/models/deepseek-v4.1-flash GGUF: https://huggingface.co/unsloth/DeepSeek-V4.1-Flash-1017-GGUF
reacted to AC-723's post with šŸ”„ 1 day ago
MixtureKit 1.1 is out šŸš€ Turn the fine-tuned models you already have on the Hub into one Mixture-of-Experts model. pip install mixturekit Give it a base model and its fine-tunes (math, code, a language...). MixtureKit merges them, trains the routers that pick an expert for every token, and shows you what they learned: šŸ§© 3 composition methods behind one YAML: Branch-Train-MiX, Mixtral-style MoE, Branch-Train-Stitch šŸ’» Merges stream the experts tensor by tensor: they run on a laptop šŸŽÆ Router training with a load balancing loss, optional LoRA šŸ” A routing visualizer: in the image, two Nile 4B experts split a mixed Egyptian sentence, Arabic script to one expert, Arabizi to the other šŸ“¦ Export to Mixtral / Qwen3-MoE, then GGUF for llama.cpp šŸ¤— mixturekit push uploads the result with a generated model card āœ… Llama 3, Qwen 2.5/3/3.5, Gemma 3/4 experts; tested on transformers 4.52 to 5.19 In the paper, a BTX model built with MixtureKit outperforms the dense baselines on multiple benchmarks. ā–¶ļø Build, train and visualize an MoE in 15 min on a free Colab GPU: https://colab.research.google.com/github/MBZUAI-Paris/MixtureKit/blob/main/examples/build_your_own_moe.ipynb šŸ“„ Paper: https://huggingface.co/papers/2512.12121 šŸ’» Code: https://github.com/MBZUAI-Paris/MixtureKit šŸ“š Docs: https://mbzuai-paris.github.io/MixtureKit/ Work with @guokan-shang and @oe2015 at @MBZUAI-Paris šŸ™Œ
new activity 4 days ago
nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB-GGUF:Now we need MiniCPM variant
View all activity

Organizations

None yet