ggml-org/llama.cpp

server: exit failure if `--embedding` is set with an incoherent `--ubatch-size`

已關閉

#6,263 建立於 2024年3月23日

 (8 則留言) (1 個反應) (0 位負責人)C++ (21,781 個分叉)batch import
enhancementgood first issuehelp wanted

倉庫指標

星標
 (124,194 顆星)
PR 合併指標
 (平均合併 6天 8小時) (30 天內合併 389 個 PR)

描述

Context

there is no advantage to increase n_batch above n_ubatch with embeddings models with pooling, because the entire batch must fit in a physical batch (ie. n_ubatch). n_batch is always >= n_ubatch.

Proposition

Exit failure if --embedding is set and --ubatch-size != --batch-size in the server example. Probably also in the retrieval example in #6193.

Aldo probably KV bert.context_size must be taken into account.

貢獻者指南