Re:Jina AI Launches First Open-Source 8K Text Embedding, Riv...
by 3abiton on Thursday October 26, 2023 @03:55AM
(#38022947)
I wonder how the perfomance fair when context size is increased. Intuitively this should be higher, but some quantized models I've tested showed noticeably worst performance.
[ Parent
| on HN ]
Re:Jina AI Launches First Open-Source 8K Text Embedding, Riv...
by Kubuxu on Thursday October 26, 2023 @05:30AM
(#38023512)
Your KV cache size is linear with context size which might put you tight on memory. There is also increased cost of recalculating KV cache of context window when the window has to move but this is close to being solved with streaming LLMs.
Re:Jina AI Launches First Open-Source 8K Text Embedding, Riv...
by woadwarrior01 on Thursday October 26, 2023 @06:23AM
(#38023780)
BERT style encoder-only models, like the embedding model being discussed here, don't need a KV cache for inference. A KV cache is only needed for efficient inference with encoder-decoder and decoder-only (aka GPT) models.