view post Post 1012 After a VLM, StepFun dropped a new audio model: Step-Audio-R1.1, enabling thinking while speaking ๐ฅ stepfun-ai/Step-Audio-R1.1โจ Apache 2.0โจ Combines dual-brain architecture and acoustic-grounded reasoning to enable real-time dialogue with SOTA-level reasoning See translation 2 replies ยท ๐ 4 4 + Reply
meituan-longcat/LongCat-Flash-Thinking-2601 Text Generation โข 562B โข Updated about 20 hours ago โข 503 โข 76
view post Post 2214 New GRPO + TRL free Colab notebook out! ๐ฅFine-tune 7B+ models on T4 GPUs thanks to a ton of memory optimizations for GRPO 7B model uses only 9.2 GB VRAM (~7ร reduction) ๐คฏTry the notebook here ๐ https://colab.research.google.com/github/huggingface/trl/blob/main/examples/notebooks/grpo_trl_lora_qlora.ipynb See translation ๐ฅ 10 10 ๐ 5 5 + Reply
view post Post 2083 Happy birthday to me!!! See translation 2 replies ยท ๐ค 15 15 ๐ 7 7 ๐ 3 3 โค๏ธ 2 2 + Reply
Jamba2 Collection Jamba2 is a highly-efficient open source family of language models built for maximum reliability and steerability in the enterprise. โข 3 items โข Updated 14 days ago โข 5