4 Commits

Author	SHA1	Message	Date
Nexesenex	3c9680fd3c	Fix Minimax M3 crash when -muge merges up/gate experts ... The graph builder for Minimax M3 (build_minimaxm3.cpp) was not passing model.layers[il].ffn_up_gate_exps to llm_build_std_moe_ffn, unlike Minimax M2 and all other MoE model graph builders. When -muge (merge_up_gate_experts) is enabled, the merge creates a single ffn_up_gate_exps tensor with ffn_up_exps and ffn_gate_exps as views. Only the parent merged tensor gets the split 'extra' pointer set. Without passing it as up_gate_exps parameter, the function sees null split pointers for up/gate (the views) while split_down_exps is valid, causing the assertion at llama-build-context.cpp:1453 to fail.	2026-06-15 15:00:32 +02:00
Kawrakow	c24d50dd88	Split mode graph for MiniMax-M3	2026-06-15 08:41:34 +00:00
Jun Yamog	c08d194edd	Use standard graph helpers for MiniMax-M3	2026-06-15 02:00:46 +00:00
Jun Yamog	0df00b3b94	Add preliminary MiniMax-M3 support	2026-06-14 12:23:20 +00:00