Unverified50% confidenceBenchmarkExact time
SSU-LSF converges within 3 to 5 epochs, and the GPU cost per forgetting request is 8.4 times lower than full retraining.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe mxfp scheme (stochastic rounding + Hadamard preprocessing) achieves 37.2K tokens/s/GPU, corresponding to 86.3% MFU, with a final training loss 2.11% higher than pure bfloat16.69% similarUnverifiedA new computer based on coupled oscillators consumes only one-thousandth the power of a traditional GPU, with approximately 6.7% performance loss, and is fully open source.68% similarUnverifiedThe core bottleneck in large model inference on GPUs is memory bandwidth; autoregressive generation of each token requires loading the full KV Cache from HBM, with compute utilization typically below 5%.67% similarUnverifiedSGLang using official NVFP4 weights with D-Spark speculative decoding achieved an average of approximately 173 tokens/second on Blackwell GPUs, approaching 200 for short tasks, while loading a 262k context window.66% similarUnverifiedWhen model weights and KV Cache cannot be fully loaded into GPU VRAM and some data must be offloaded to main memory, inference speed drops by one to two orders of magnitude.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/975724API
curl https://kongchang.com/api/v1/knowledge/claims/975724MCP
get_claim(id=975724)