From 59977eba7b0a3603d0017717d3beec7bde018f3c Mon Sep 17 00:00:00 2001 From: yifant-code Date: Tue, 16 Dec 2025 07:27:36 -0500 Subject: [PATCH] server: fix crash when batch > ubatch with embeddings (#17912) * server: fix crash when batch > ubatch with embeddings (#12836) Fixes #12836 where the server crashes with GGML_ASSERT failure when running with embeddings enabled and n_batch > n_ubatch. Root cause: Embeddings use non-causal attention which requires all tokens to be processed within a single ubatch. When n_batch > n_ubatch, the server attempts to split processing, causing assertion failure. Solution: - Add parameter validation in main() after common_params_parse() - When embeddings enabled and n_batch > n_ubatch: * Log warnings explaining the issue * Automatically set n_batch = n_ubatch * Prevent server crash This follows the approach suggested by @ggerganov in issue #12836. Note: This supersedes stalled PR #12940 which attempted a runtime fix in the old examples/server/server.cpp location. This implementation validates at startup in tools/server/server.cpp (current location). Testing: - Build: Compiles successfully - Validation triggers: Warns when -b > -ub with --embedding - Auto-correction works: Adjusts n_batch = n_ubatch - No false positives: Valid params don't trigger warnings - Verified on macOS M3 Pro with embedding model * Update tools/server/server.cpp --------- Co-authored-by: ytian218 Co-authored-by: Georgi Gerganov --- tools/server/server.cpp | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/tools/server/server.cpp b/tools/server/server.cpp index 235ae4e8c0..8538427f73 100644 --- a/tools/server/server.cpp +++ b/tools/server/server.cpp @@ -73,8 +73,18 @@ int main(int argc, char ** argv, char ** envp) { return 1; } + // validate batch size for embeddings + // embeddings require all tokens to be processed in a single ubatch + // see https://github.com/ggml-org/llama.cpp/issues/12836 + if (params.embedding && params.n_batch > params.n_ubatch) { + LOG_WRN("%s: embeddings enabled with n_batch (%d) > n_ubatch (%d)\n", __func__, params.n_batch, params.n_ubatch); + LOG_WRN("%s: setting n_batch = n_ubatch = %d to avoid assertion failure\n", __func__, params.n_ubatch); + params.n_batch = params.n_ubatch; + } + if (params.n_parallel < 0) { LOG_INF("%s: n_parallel is set to auto, using n_parallel = 4 and kv_unified = true\n", __func__); + params.n_parallel = 4; params.kv_unified = true; }