#quantization
- Google DeepMind Releases Gemma 4 QAT Checkpoints: Sub-1 GB On-Device E2B Model Google DeepMind models-llm
- ViQ: Text-Aligned Visual Quantized Representations at Any Resolution (ECCV 2026) Tencent Hunyuan research
- The Mirage of Optimizing Training Policies: Monotonic Inference Policies as the Real Objective for LLM Reinforcement Learning Tianjin University / Alibaba research
- PrismML Releases Bonsai 27B: First 27B-Class Model to Run on iPhone PrismML models-llm
- PrismML ships Bonsai 2 27B: ternary weights near-lossless to the teacher PrismML models-llm
- LongLive-2.0: NVFP4 Parallel Infrastructure for Long Video Generation (NVIDIA, 1,220 HF upvotes) NVIDIA research
- Quantized Reasoning Models Think They Need to Think Longer, but They Do Not Meta research
- vLLM v0.24.0: Model Runner V2 Default, Rust Frontend, SM90 FP8 Speedups vLLM tools
- llama.cpp b9603: Qualcomm Adreno OpenCL Kernels for On-Device Inference ggml-org tools
- llama.cpp nightly burst b10668-b10679 includes Vulkan wrong-token fix llama.cpp tools