Unverified50% confidenceFactExact time
模型量化（Quantization）是将深度学习模型参数从高精度浮点数（如FP32、FP16）压缩为低精度表示（如INT8、INT4）的技术，典型方案包括GPTQ、GGUF、AWQ
1
Sources
50%
Confidence
Long-term
Relevance
8/24/2026
First Seen
Sources
Related Entities
Related Claims
Unverified深度学习模型推理通常需要GPU或专用加速器，模型服务部署工具包括Seldon、KServe、BentoML、TorchServe68% similarUnverifiedThinking Level in DeepSeek models is a capability that allows callers to control how much reasoning the model performs before generating a response, representing a practical application of Test-Time Compute Scaling.66% similarUnverified精排层普遍使用深度学习模型如Wide&Deep、DeepFM或Transformer架构进行CTR预估和互动概率计算65% similarUnverified许多深度学习模型在权重量化到4位后仍能保持95%以上的推理精度65% similarUnverified本地ASR模块通常采用量化压缩的轻量级深度学习模型，参数量在数百万至数千万级，端侧延迟可控制在100ms以内64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/794269API
curl https://kongchang.com/api/v1/knowledge/claims/794269MCP
get_claim(id=794269)