Unverified50% confidenceFactExact time
经典DQN预处理后网络输入张量的shape为(84, 84, 4)，即灰度化和降采样至84×84后4个通道对应4个时间步
1
Sources
50%
Confidence
Long-term
Relevance
8/6/2026
First Seen
Sources
Related Claims
UnverifiedThe model ultimately outputs a tensor of 201×18×4, where 4 represents 4 lane lines, 18 represents key rows, and 201 represents 200 grid positions plus 1 non-existence class.59% similarUnverifiedINT4量化将原本由65536个不同值（FP16）表示的权重压缩到仅16个离散级别58% similarUnverifiedIn LOD generation testing, Mixar generated three levels—LOD0, LOD1, and LOD2—with textures reduced from 4K to 2K and 1K, creating a total of 52 new textures, with results meeting expectations.57% similarUnverified推理端优化技术包括量化、知识蒸馏和KV Cache，量化可将参数从32位或16位浮点压缩为8位甚至4位整数55% similarUnverifiedThe Q4_K_M quantization scheme averages approximately 4.8 bits per weight, while the Q8_0 quantization scheme uses 8 bits per weight.55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/693801API
curl https://kongchang.com/api/v1/knowledge/claims/693801MCP
get_claim(id=693801)