[None][fix] Size the cross-attention KV cache from the encoder's decl… - #19516

[None][fix] Size the cross-attention KV cache from the encoder's decl…#19516

ravo42-maker wants to merge 3 commits into

NVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

ravo42-maker:fix/cross-kv-encoder-lengthravo42-maker/TensorRT-LLM:fix/cross-kv-encoder-lengthCopy head branch name to clipboard