Unverified60% confidenceFactExact time
GQA allows multiple query heads to share the same set of KV heads, compressing the KV cache size to 1/4 to 1/8 of the original.
2
Sources
60%
Confidence
Long-term
Relevance
7/19/2026
First Seen
Sources
Related Claims
Unverified使用非GQA架构（KV头数等于查询头数）在相同条件下的KV缓存是GQA的4倍79% similarUnverifiedGQA groups attention heads to share KV projections, which can compress the KV cache to 1/4 or less of the original size with almost no loss in model quality.75% similarVerifiedQuantized KV caches (such as storing KV vectors in FP8 or INT4) can compress the cache size to 1/2 or even 1/4 of the original.74% similarVerifiedMLA compresses KV pairs into a low-dimensional latent space for caching and recovers the full Keys and Values via an up-projection matrix during inference, compressing the KV cache by several times.72% similarUnverifiedKV Cache memory usage is proportional to the number of layers × number of attention heads × context length × precision; GQA reduces cache size by having multiple query heads share the same set of Key-Value heads, and this design has been adopted by the Llama 2/3 and Qwen series.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/561355API
curl https://kongchang.com/api/v1/knowledge/claims/561355MCP
get_claim(id=561355)