3 18

André Leo Bjørningstad

LeoHyperlink

AI & ML interests

None yet

Recent Activity

liked a model 1 day ago

Qwen/Qwen3.5-122B-A10B

liked a model 1 day ago

Qwen/Qwen3.5-397B-A17B

liked a model 1 day ago

Qwen/Qwen3.5-35B-A3B

View all activity

Organizations

liked 4 models 1 day ago

upvoted an article 4 days ago

Article

GGML and llama.cpp join HF to ensure the long-term progress of Local AI

9 days ago

•

469

liked a model 15 days ago

Nanbeige/Nanbeige4.1-3B

Text Generation • 4B • Updated 2 days ago • 311k • • 900

reacted to danielhanchen's post with 🔥 17 days ago

Post

5153

We collaborated with Hugging Face to enable you to train MoE models 12× faster with 35% less VRAM via our new Triton kernels (no accuracy loss). 🤗

Train gpt-oss locally on 12.8GB VRAM with our free notebooks: https://unsloth.ai/docs/new/faster-moe

1 reply

liked 2 models 23 days ago

TeichAI/Qwen3-4B-Thinking-2507-Claude-4.5-Opus-High-Reasoning-Distill-GGUF

4B • Updated Dec 9, 2025 • 5.19k • 19

ruv/ruvltra-claude-code

Text Generation • 0.5B • Updated Jan 20 • 10.2k • 115

replied to danielhanchen's post 23 days ago

I got qwen3-coder-next up 24tps when i removed -nkvo and kept -kvu. Think i could increase it further if i took the time to compile llama.cpp and not use docker image.

upvoted a collection 23 days ago

[models] GTX 1660 Super 6gb

Collection

The best little card under 100 euros. Full Precision vs Quants not benchmarked. This card is so much better at running inference than you realize. • 11 items • Updated 6 days ago • 2

liked a dataset 23 days ago

Qwen/RationaleRM

Preview • Updated 23 days ago • 1.72k • 21

replied to danielhanchen's post 24 days ago

I run it on threadripper 3970x with 256gb system ram and offloading computation layers to a gtx 1660 6gb vram. Using llama.cpp with -nkvo -kvu and all MoE on CPU. With an amazing speed on 14/TpS generation speed using q8_0. I’m amazed

reacted to danielhanchen's post with 🤯🚀🔥 24 days ago

Post

3753

Qwen releases Qwen3-Coder-Next! 💜 Run the locally on 46GB RAM or less.

Thhe model excels at agentic coding & local use. With 256K context, it delivers similar performance to models with 10-20× more active parameters.

GGUF: unsloth/Qwen3-Coder-Next-GGUF
Guide: https://unsloth.ai/docs/models/qwen3-coder-next