nicolas
/
ik_llama_opt
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
ik_llama_opt/src
History
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
..
graphs Fix Minimax M3 crash when -muge merges up/gate experts 2026-06-15 15:00:32 +02:00
CMakeLists.txt Add preliminary MiniMax-M3 support 2026-06-14 12:23:20 +00:00
llama-arch.cpp Add preliminary MiniMax-M3 support 2026-06-14 12:23:20 +00:00
llama-arch.h Add preliminary MiniMax-M3 support 2026-06-14 12:23:20 +00:00
llama-build-context.cpp Split mode graph for MiniMax-M3 2026-06-15 08:41:34 +00:00
llama-build-context.h Split mode graph for MiniMax-M3 2026-06-15 08:41:34 +00:00
llama-context.h Add extra nodes when dealing with MLA and amb (#1899) 2026-05-29 15:17:24 +03:00
llama-cparams.h MLA tensor parallelism under -sm graph (DEEPSEEK2/GLM_DSA/MISTRAL4) (#1821) 2026-05-19 08:36:17 +03:00
llama-delta-net.cpp MTP: faster recurrent state restore (#1791) 2026-05-13 11:00:24 +03:00
llama-delta-net.h MTP: faster recurrent state restore (#1791) 2026-05-13 11:00:24 +03:00
llama-expert-io.h Add --defer-experts flag to defer expert mmap residency on Linux (#1634) 2026-04-16 08:54:44 +02:00
llama-grammar.cpp Bug fixes (#1940) 2026-06-10 07:45:49 +02:00
llama-grammar.h llama : add token matching support to llama-grammar (#1220) 2026-02-03 07:57:17 +02:00
llama-hparams.cpp Add preliminary MiniMax-M3 support 2026-06-14 12:23:20 +00:00
llama-hparams.h model: add Poolside Laguna XS.2 support (#1911) 2026-06-08 18:33:12 +02:00
llama-impl.h Full graph parallel for Qwen3.5 (dense and MoE) (#1388) 2026-03-10 09:08:24 +01:00
llama-load-tensors.cpp Add preliminary MiniMax-M3 support 2026-06-14 12:23:20 +00:00
llama-mmap.cpp Add --defer-experts flag to defer expert mmap residency on Linux (#1634) 2026-04-16 08:54:44 +02:00
llama-mmap.h Add --defer-experts flag to defer expert mmap residency on Linux (#1634) 2026-04-16 08:54:44 +02:00
llama-model-loader.cpp fix: use int8_t for GGUF bool array loading instead of platform-dependent bool (#1648) 2026-04-17 07:25:07 +02:00
llama-model-loader.h MTP: option to use re-quantized output tensor for better TG performance (#1809) 2026-05-16 14:40:18 +03:00
llama-model.cpp Add preliminary MiniMax-M3 support 2026-06-14 12:23:20 +00:00
llama-model.h model: add Poolside Laguna XS.2 support (#1911) 2026-06-08 18:33:12 +02:00
llama-quantize.cpp quantize: add exception for Gemma4 (#1897) 2026-05-29 10:54:21 +03:00
llama-quantize.h Allow using -rtr and -muge together (#1444) 2026-03-16 18:26:26 +01:00
llama-sampling.cpp Bug fixes (#1940) 2026-06-10 07:45:49 +02:00
llama-sampling.h Add adaptive sampling clone and free functions to manage memory (#1851) 2026-05-21 08:11:17 +03:00
llama-spec-features.cpp Support for alternative Gemma4 assistant (#1937) 2026-06-09 09:30:12 +02:00
llama-spec-features.h Move embedding management to speculative (#1825) 2026-05-20 17:42:48 +03:00
llama-vocab.cpp model: add Cohere2-MoE North Mini Code support (#1945) 2026-06-10 15:28:27 +02:00
llama-vocab.h Add Mellum2 architecture support (#1919) 2026-06-04 14:28:02 +02:00
llama.cpp Add preliminary MiniMax-M3 support 2026-06-14 12:23:20 +00:00
unicode-data.cpp Merge mainline llama.cpp (#3) 2024-07-27 07:55:01 +02:00
unicode-data.h Merge mainline llama.cpp (#3) 2024-07-27 07:55:01 +02:00
unicode-script-data.cpp Add Unicode allowlist (#1597) 2026-04-10 18:22:57 +02:00
unicode.cpp Gemma4 tokenizer fixes (#1603) 2026-04-09 15:33:28 +02:00
unicode.h Add Unicode allowlist (#1597) 2026-04-10 18:22:57 +02:00