leejet b1f5a51a1e fix: chunk MiniMax-H3 QKV projections within backend buffer limits (#2286)

2026-11-08 00:04:37 +08:00

..

cpu

feat: support packed INT4 convrot models (#2276)

2026-11-06 00:35:58 +08:00

cuda

perf: accelerate packed INT4 CUDA inference with CUTLASS (#2278)

2026-11-06 01:59:05 +08:00

probe

fix: benchmark Vulkan direct convolution in auto mode (#2141)

2026-10-15 23:08:30 +08:00

vulkan

fix: chunk MiniMax-H3 QKV projections within backend buffer limits (#2286)

2026-11-08 00:04:37 +08:00

backend_fit.cpp

fix: prefer discrete GPUs in auto-fit device selection (#2259)

2026-11-02 01:09:36 +08:00

backend_fit.h

refactor: define VAE tile dimensions in image pixels (#2059)

2026-09-25 18:20:32 +08:00

compute_workspace.cpp

refactor: move model-specific operators and fusions into backend extensions (#2253)

2026-11-01 21:18:37 +08:00

compute_workspace.h

perf: optimize Vulkan kernels, graph fusion, and workspace reuse (#2251)

2026-10-31 14:55:58 +08:00

ggml_compat.cpp

refactor: move model-specific operators and fusions into backend extensions (#2253)

2026-11-01 21:18:37 +08:00

ggml_compat.h

refactor: move model-specific operators and fusions into backend extensions (#2253)

2026-11-01 21:18:37 +08:00

ggml_extend_backend.cpp

refactor: move model-specific operators and fusions into backend extensions (#2253)

2026-11-01 21:18:37 +08:00

ggml_extend_backend.h

perf: optimize H3 inference and segmented weight prefetch (#2240)

2026-10-30 00:38:14 +08:00

ggml_extend.cpp

perf: use native SiTU for MOSS with composite fallback (#2210)

2026-11-03 23:41:09 +08:00

ggml_extend.h

perf: use native SiTU for MOSS with composite fallback (#2210)

2026-11-03 23:41:09 +08:00

ggml_extensions.cpp

feat: support packed INT4 convrot models (#2276)

2026-11-06 00:35:58 +08:00

ggml_extensions.h

feat: support packed INT4 convrot models (#2276)

2026-11-06 00:35:58 +08:00

ggml_graph_cut.cpp

refactor: move model-specific operators and fusions into backend extensions (#2253)

2026-11-01 21:18:37 +08:00

ggml_graph_cut.h

perf: reduce CPU overhead in graph execution and sampling (#1997)

2026-09-19 18:32:58 +08:00

ggml_runner.cpp

fix: chunk MiniMax-H3 QKV projections within backend buffer limits (#2286)

2026-11-08 00:04:37 +08:00

ggml_runner.h

perf: optimize H3 inference and segmented weight prefetch (#2240)

2026-10-30 00:38:14 +08:00

ggml_tensor_utils.cpp

fix: open model files with UTF-8 paths on Windows (#2144)

2026-10-17 00:50:09 +08:00

ggml_tensor_utils.h

refactor: split ggml extensions and move implementations to cpp files (#1945)

2026-09-07 23:34:45 +08:00

layer_split_partition.cpp

feat: add P-H3-Image generation and editing support (#2272)

2026-11-04 23:24:51 +08:00

layer_split_partition.h

feat: drive layer split from graph-cut segments (#1762)

2026-07-07 23:16:52 +08:00

ordered_map.hpp

refactor: unify model source and weight lifecycle management (#1956)

2026-09-10 23:53:48 +08:00

parallel.cpp

perf: parallelize host tensor elementwise and broadcast ops (#1998)

2026-09-19 21:46:16 +08:00

parallel.h

perf: parallelize host tensor elementwise and broadcast ops (#1998)

2026-09-19 21:46:16 +08:00

pe_cache.cpp

perf: cache CPU position embeddings across DiT runners (#2211)

2026-10-25 05:18:30 +08:00

pe_cache.h

perf: cache CPU position embeddings across DiT runners (#2211)

2026-10-25 05:18:30 +08:00

regex.cpp

feat: support external Hugging Face tokenizer JSON files (#1973)

2026-09-15 00:27:39 +08:00

regex.h

feat: support external Hugging Face tokenizer JSON files (#1973)

2026-09-15 00:27:39 +08:00

rng_mt19937.hpp

feat: support Brownian tree noise in all noise injection samplers (#1899)

2026-09-15 02:37:31 +08:00

rng_philox.hpp

perf: reduce CPU overhead in graph execution and sampling (#1997)

2026-09-19 18:32:58 +08:00

rng.hpp

feat: support Brownian tree noise in all noise injection samplers (#1899)

2026-09-15 02:37:31 +08:00

runner_cache.cpp

perf: optimize H3 inference and segmented weight prefetch (#2240)

2026-10-30 00:38:14 +08:00

runner_cache.h

perf: optimize H3 inference and segmented weight prefetch (#2240)

2026-10-30 00:38:14 +08:00

segment_graph_bindings.cpp

chore: format code

2026-10-24 01:23:42 +08:00

segment_graph_bindings.h

refactor: unify runner lifecycles and weight residency (#1940)

2026-09-06 22:30:45 +08:00

segment_weight_pipeline.cpp

perf: optimize H3 inference and segmented weight prefetch (#2240)

2026-10-30 00:38:14 +08:00

segment_weight_pipeline.h

perf: optimize H3 inference and segmented weight prefetch (#2240)

2026-10-30 00:38:14 +08:00

tensor_ggml.hpp

fix: open model files with UTF-8 paths on Windows (#2144)

2026-10-17 00:50:09 +08:00

tensor.hpp

perf: parallelize host tensor elementwise and broadcast ops (#1998)

2026-09-19 21:46:16 +08:00

util.cpp

fix: correct Linux sysfs path for physical core detection (#2186)

2026-10-22 22:53:18 +08:00

util.h

refactor: escape newlines in prompt logs and clarify source separators (#2106)

2026-10-06 19:51:12 +08:00