Unverified50% confidenceBenchmarkExact time
With a configuration of 14 rollout GPUs plus 2 trainer GPUs, after 100 training steps, exact match on the held-out set rose from 14.5% to 47%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/20/2026
First Seen
Valid until: 1/18/2027
Sources
Related Entities
Related Claims
UnverifiedThe milestones publicly demonstrated by VKORE include: training a 46-million-parameter model from scratch, running on 32 inference-only NPU chips, with a loss correlation of 0.99 with a GPU training baseline.71% similarUnverifiedIn the controlled tests with step-by-step improvements, with the configuration of "both run on the GPU alternately", the mean first audio was 2.3 seconds (range 1.9-2.7), and the mean number of stutters per answer was 0.09 (range 0-1), better than "both run on the GPU simultaneously" at 3.9 seconds (3.8-4.1) and 2.84 times (9-17 times).70% similarUnverifiedUnder the synthetic settings evaluated, DASA provides a 3.6 to 4.9× speedup compared to GRADMM, with comparable peak GPU memory usage.68% similarUnverifiedThe source does not provide quantitative benchmark comparisons for VKORE-46M-NPU, including perplexity scores, downstream task accuracy, and training throughput in tokens per second.68% similarUnverified2012年Alex Krizhevsky等人使用AlexNet在ImageNet竞赛中将top-5错误率从26%降低到15.3%，该突破基于两块GTX 580显卡的CUDA加速训练实现67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1074534API
curl https://kongchang.com/api/v1/knowledge/claims/1074534MCP
get_claim(id=1074534)