Unverified50% confidenceFactExact time
VKORE-46M-NPU has a parameter scale of 46M, about one-third of GPT-2 Small (about 124 million parameters), and its significance lies in validating that the NPU training path is feasible, rather than in the quality of the model itself.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedTaking GPT-3's attention layers as an example, the original weight matrix is approximately 12,288×12,288 (about 150 million parameters), while a LoRA decomposition with rank r=8 requires training only about 200,000 parameters, a compression ratio exceeding 700×.66% similarUnverifiedThe open-source model MiniMax M2.6 Pro scored 46 points, while GLM 5.3 and Qwen 3.8 Max scored approximately 45 points, both close to GPT-6 Sol's score of 48 points.66% similarUnverifiedCRN v2 has approximately 34 million trainable parameters, equivalent to 0.73% of the 4.65 billion-parameter text module.65% similarUnverifiedLoRA使可训练参数量通常仅为全参微调的0.1%-1%，可在单张RTX 3090/4090（显存24GB）上微调7B乃至13B级别的模型64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1026580API
curl https://kongchang.com/api/v1/knowledge/claims/1026580MCP
get_claim(id=1026580)