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

[jakexcosme]

Note: This issue was copied from ggml-org#6263

Original Author: @phymbert
Original Issue Number: ggml-org#6263
Created: 2024-03-23T17:03:49Z

---

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.

• See @slaren comment in: server: docs: --threads and --threads, --ubatch-size, --log-disable ggml-org/llama.cpp#6254 (comment)

Proposition

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

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