Tencent compresses 1.5TB model to 200GB with minimal benchmark loss via layer-wise calibration
ChrisGPT · x · 2026-08-29
Tencent demonstrated extreme model compression efficiency, reducing Hy4 preview from 1.5TB to 200GB with negligible impact on benchmarks. The core innovation involves using calibration data to dynamically determine compression intensity per layer: less sensitive layers are aggressively pushed down to 1.31 bit, while sensitive layers remain near 2 bit to prevent collapse. This layer-sensitivity-based differential quantization strategy yields massive efficiency gains.
More from Infra
- engy.ai processes 50B+ tokens in a day as open-model inference demand surges — const_reborn · 2026-10-16
- New inference engine 'e' hits 10,500 tok/s on a single DGX Station — brandon_galang · 2026-10-16
- SGLang ships Day-0 support for FLUX 3 Image Klein: 10 reference images, BF16/FP8/NVFP4 variants — ying11231 · 2026-10-16
- NVIDIA's Nemotron 3.5 Super VL hits CoreWeave: 12B-active multimodal MoE at $0.25/M input tokens — wandb · 2026-10-16
- Chip equipment beats bull case: 2026 WFE seen at $153B, $278B by 2030 — BenBajarin · 2026-10-16
- Steam now launches in 5.23s, beating Epic's 5.47s — and vLLM gets roasted for slow startup — AlpinDale · 2026-10-16