Swift 1.5 27B Collection Swift 1.5 on Qwen3.8-27B: stronger than Swift 1.0 on agentic and coding tasks, with fewer thinking tokens. BF16 weights and every quant. • 11 items • Updated 8 days ago • 21
Running on Zero Agents Featured 264 Laya Demo 🎯 264 Fast System 1 decisions with calibrated probabilities
prism-ml/Ternary-Bonsai-2-27B-gguf Text Generation • 27B • Updated 7 days ago • 3.87M • 2.35k
Running Featured 147 Ternary Bonsai 2 WebGPU Kernels 🌳 147 Run Ternary-Bonsai-2-27B locally in your browser on WebGPU
litert-community/functiongemma-270m-ft-mobile-actions Text Generation • Updated Aug 31 • 1.15k • 247