ggml-org/llama.cpp

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

Geschlossen

#6.263 geöffnet am 23.03.2024

 (8 Kommentare) (1 Reaktion) (0 zugewiesene Personen)C++ (21.825 Forks)batch import
enhancementgood first issuehelp wanted

Repository-Metriken

Stars
 (124.401 Sterne)
PR-Merge-Metriken
 (Durchschn. Merge 6T 8h) (389 gemergte PRs in 30 T)

Beschreibung

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.

Contributor Guide