Partially Verified65% confidenceOpinionExact time
VKORE aims to rebuild a CUDA-like abstraction layer on NPUs, providing a unified training and inference interface upward and adapting to different vendors' NPU instruction sets downward, so that data centers can swap accelerators like swapping storage or network cards without rewriting model code, thereby breaking single-vendor lock-in.
3
Sources
65%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedVIDRAFT is integrating the above capabilities into an NPU software layer called VKORE, and positions it by analogy to the ecosystem advantage Nvidia built through CUDA, describing it as "the CUDA of NPUs".80% similarUnverifiedThe goal of the VKORE software layer that VIDRAFT is building is to become "the CUDA of the NPU field," covering three capabilities: inference acceleration, full pretraining, and LoRA fine-tuning; what is released this time is a publicly verified artifact of the pretraining milestone.77% similarVerifiedCUDA is NVIDIA's proprietary platform, requiring explicit kernel function authoring; it has a steep learning curve but the greatest performance potential.72% similarUnverifiedcuML implements high-performance computing kernels in CUDA C++ at the backend and replicates scikit-learn's API interface at the Python layer, enabling zero-code migration70% similarUnverifiedPyTorch 通过 ATen 张量库将高层 API 转译为优化的 CUDA kernel，并整合 cuDNN、cuBLAS 等 NVIDIA 优化库70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1017232API
curl https://kongchang.com/api/v1/knowledge/claims/1017232MCP
get_claim(id=1017232)