 leejet
|
f0204a7eaa
|
refactor: preserve INT8 APIs through backend extensions
|
2026-11-01 01:55:51 +08:00 |
|
|
leejet
|
051b49751e
|
refactor: use custom RoPE Apply nodes exclusively
|
2026-11-01 01:11:17 +08:00 |
|
|
leejet
|
d07d48d466
|
refactor: separate CPU kernels from shared operator extensions
|
2026-11-01 00:44:28 +08:00 |
|
|
leejet
|
d11410eee7
|
refactor: share CUDA-style extension kernels across backends
|
2026-11-01 00:21:35 +08:00 |
|
|
leejet
|
7ab323ef25
|
refactor: move Vulkan RoPE kernels into extensions
|
2026-10-31 23:39:19 +08:00 |
|
|
leejet
|
5c5535d01d
|
refactor: register CUDA affine fusion extensions
|
2026-10-31 22:59:06 +08:00 |
|
|
leejet
|
7ad74e6490
|
refactor: register CUDA INT8 GEMM and epilogue extensions
|
2026-10-31 22:29:12 +08:00 |
|
|
leejet
|
d8964d0302
|
refactor: enable CUDA graph replay for extensions
|
2026-10-31 21:52:28 +08:00 |
|
|
leejet
|
ba2188624e
|
refactor: move CUDA ConvRot fusion into extensions
|
2026-10-31 21:18:40 +08:00 |
|
|
leejet
|
773b11caa7
|
refactor: move CUDA INT8 residual fusion into extensions
|
2026-10-31 20:55:44 +08:00 |
|
|
leejet
|
b71db9a314
|
feat: add CUDA execution for custom INT8 operators
|
2026-10-31 20:24:30 +08:00 |
|
|
leejet
|
2f68a92505
|
refactor: register CPU INT8 GEMM and ConvRot extensions
|
2026-10-31 20:10:56 +08:00 |
|
|
leejet
|
ab232a474b
|
refactor: register INT8 projection and attention fusion
|
2026-10-31 20:01:41 +08:00 |
|
|
leejet
|
1d2d31ffb6
|
refactor: register CUDA RoPE kernels and fusion in sd.cpp
|
2026-10-31 19:18:40 +08:00 |
|
|
leejet
|
4b274821e9
|
refactor: register CuTe attention and fusion in sd.cpp
|
2026-10-31 19:00:24 +08:00 |
|
|
leejet
|
c6a4666289
|
build: pin ggml operator extension support
|
2026-10-31 18:18:54 +08:00 |
|
|
leejet
|
f4a1902061
|
refactor: register RoPE kernels and fusion outside ggml
|
2026-10-31 18:18:33 +08:00 |
|
|
leejet
|
02e664d12d
|
perf(vulkan): update ggml for H3 inference optimizations
|
2026-10-31 17:33:06 +08:00 |
|
|
leejet
|
0a067120b3
|
perf: optimize Vulkan kernels, graph fusion, and workspace reuse (#2251)
|
2026-10-31 14:55:58 +08:00 |
|
|
leejet
|
e239e4b0fd
|
refactor: replace manual gated activations with native ggml GLU (#2250)
|
2026-10-30 01:32:46 +08:00 |
|
|
leejet
|
37bc88b5e9
|
perf: optimize H3 inference and segmented weight prefetch (#2240)
|
2026-10-30 00:38:14 +08:00 |
|
|
leejet
|
77b72a2efe
|
perf: accelerate INT8 diffusion inference with CUDA fusions (#2235)
|
2026-10-28 22:02:46 +08:00 |
|
|
leejet
|
687915518e
|
ci: complete Node 24 migration for workflows and Docker builds
|
2026-10-28 01:53:21 +08:00 |
|
|
leejet
|
be486498ea
|
fix: enable segmented execution for FLUX.3 (#2224)
|
2026-10-27 00:32:18 +08:00 |
|
|
leejet
|
bb6b128b2d
|
fix: avoid repeated seeks when skipping small GGUF metadata values (#2223)
|
2026-10-27 00:24:43 +08:00 |
|
|
leejet
|
e41fbcf455
|
feat(server): add recipe constraints and update generation workspace (#2217)
|
2026-10-26 01:01:43 +08:00 |
|
|
leejet
|
012c7f8866
|
ci: trigger builds on recipe JSON changes
|
2026-10-25 19:39:32 +08:00 |
|
|
leejet
|
a5e48c66aa
|
chore: reduce large recipe defaults to 1024x1024 except PiD
|
2026-10-25 19:37:41 +08:00 |
|
|
leejet
|
b29962db7a
|
sync: update frontend
|
2026-10-25 19:30:23 +08:00 |
|
|
leejet
|
bb6a0ccfa3
|
feat: support FLUX.3 Diffusers LoRA tensor names (#2214)
|
2026-10-25 16:55:45 +08:00 |
|
|
leejet
|
ea209dbfcf
|
perf: cache CPU position embeddings across DiT runners (#2211)
|
2026-10-25 05:18:30 +08:00 |
|
|
leejet
|
a85db96aa5
|
feat: return generation metadata with core results (#2206)
|
2026-10-25 03:18:03 +08:00 |
|
|
leejet
|
95ff762308
|
fix: update ggml to fix SageAttention builds on older GPUs (#2199)
|
2026-10-24 01:30:55 +08:00 |
|
|
leejet
|
e0ed724e28
|
chore: format code
|
2026-10-24 01:23:42 +08:00 |
|
|
leejet
|
cd7f986058
|
refactor: centralize model factories and defaults (#2198)
|
2026-10-24 01:03:05 +08:00 |
|
|
leejet
|
fd21ecfcee
|
refactor: use one placeholder token per image in LLM inputs (#2191)
|
2026-10-23 00:45:33 +08:00 |
|
|
leejet
|
e779d4f149
|
feat(server): support cancellation of running generation jobs (#2190)
|
2026-10-23 00:24:56 +08:00 |
|
|
leejet
|
39bc9d0b4b
|
fix: update ggml to avoid small BAR heaps on Vulkan (#2181)
|
2026-10-22 22:58:25 +08:00 |
|
|
leejet
|
a77d362318
|
fix: recheck memory capacity when eviction candidates are unchanged (#2183)
|
2026-10-22 00:03:58 +08:00 |
|
|
leejet
|
b15c8aac39
|
fix: include RPC devices in --list-devices (#2182)
|
2026-10-21 23:47:50 +08:00 |
|
|
leejet
|
2e2ba3faec
|
feat: add Fibo 2.0 Dev and Edit Turbo support (#2180)
|
2026-10-21 23:32:19 +08:00 |
|
|
leejet
|
0518b34374
|
feat(server): add recipe-based model management (#2176)
|
2026-10-21 00:17:30 +08:00 |
|
|
leejet
|
b13a48219d
|
docs: add flux3 gguf links
|
2026-10-18 18:12:17 +08:00 |
|
|
leejet
|
d55cc0f75f
|
docs: update MOSS-Image docs
|
2026-10-18 16:22:31 +08:00 |
|
|
leejet
|
95b3090664
|
feat: support MOSS split QKV weights and batch host loading (#2160)
|
2026-10-18 16:02:28 +08:00 |
|
|
leejet
|
ab8d5ea8c3
|
feat: support Diffusers-format MiniMax-H3 LoRAs (#2159)
|
2026-10-18 14:50:36 +08:00 |
|
|
leejet
|
f0aac524d3
|
ci: align MUSA Docker image with upstream SDK 5.2.0 and mp_31
|
2026-10-17 14:14:29 +08:00 |
|
|
leejet
|
68285df319
|
feat: add moss-image support (#2148)
|
2026-10-17 03:38:38 +08:00 |
|
|
leejet
|
af0fc72c96
|
chore: sync ggml and remove obsolete conv1d workarounds (#2147)
|
2026-10-17 02:39:51 +08:00 |
|
|
leejet
|
8cd731e7b4
|
feat: add FLUX.3 Image Klein and Klein Base support (#2143)
|
2026-10-16 02:42:07 +08:00 |
|
|
leejet
|
e47426c044
|
fix: benchmark Vulkan direct convolution in auto mode (#2141)
|
2026-10-15 23:08:30 +08:00 |
|
|
leejet
|
1d0a58a350
|
chore: move parameter and tokenizer diagnostics to debug (#2135)
|
2026-10-13 23:37:58 +08:00 |
|
|
leejet
|
46362fc4c0
|
feat: add automatic direct convolution selection (#2134)
|
2026-10-13 23:25:19 +08:00 |
|
|
leejet
|
d763dae3b8
|
feat: add per-module automatic flash attention selection (#2132)
|
2026-10-13 22:58:45 +08:00 |
|
|
leejet
|
d637cfb935
|
feat: add Qwen Image 2.1 Turbo sigma preset (#2130)
|
2026-10-12 23:09:42 +08:00 |
|
|
leejet
|
2988060a43
|
feat: add TAEQI2.1 support for Qwen Image 2.1 (#2123)
|
2026-10-11 00:01:18 +08:00 |
|
|
leejet
|
0e1d733b7a
|
docs: restrict README UI listings to open-source projects
|
2026-10-10 22:40:06 +08:00 |
|
|
leejet
|
f89d9b13d7
|
feat: add Iris-3B text-to-image support (#2117)
|
2026-10-10 01:39:49 +08:00 |
|
|
leejet
|
f2f168fa70
|
fix: preserve singleton output dimension when merging LoRA (#2116)
|
2026-10-10 01:12:03 +08:00 |
|
|
leejet
|
7867f6da4a
|
fix: bump RPC protocol patch version for rope apply op (#2115)
|
2026-10-10 00:21:25 +08:00 |
|
|
leejet
|
228c707fde
|
perf: use ggml rope apply op on supported backends (#2113)
|
2026-10-09 04:32:36 +08:00 |
|
|
leejet
|
a1ded76da5
|
chore: resolve MSVC warnings in ggml extensions
|
2026-10-06 20:10:36 +08:00 |
|
|
leejet
|
8cc5ad8f1c
|
sync: update frontend
|
2026-10-06 20:04:34 +08:00 |
|
|
leejet
|
4a892293ac
|
docs: document video reference image support and constraints
|
2026-10-06 20:03:52 +08:00 |
|
|
leejet
|
fd43b542a6
|
refactor: escape newlines in prompt logs and clarify source separators (#2106)
|
2026-10-06 19:51:12 +08:00 |
|
|
leejet
|
fb8ee71f55
|
refactor: unify circular and non-circular tiling (#2105)
|
2026-10-06 19:20:44 +08:00 |
|
|
leejet
|
203a882a28
|
refactor: shorten log levels and move source locations to the end (#2104)
|
2026-10-06 18:26:49 +08:00 |
|
|
leejet
|
3f8527a46c
|
feat: enable hip INT8 tensorwise matmul and convrot (#2071)
|
2026-09-27 23:24:07 +08:00 |
|
|
leejet
|
42ab1c12f1
|
fix: keep scaled INT8 convrot matmuls on Vulkan (#2070)
|
2026-09-27 20:43:06 +08:00 |
|
|
leejet
|
27f7c430c0
|
fix: update ggml to fix INT8 convrot backend fallback (#2069)
|
2026-09-27 20:42:31 +08:00 |
|
|
leejet
|
ede32a621e
|
fix: load INT8 convrot LLM embeddings correctly (#2067)
|
2026-09-27 19:46:21 +08:00 |
|
|
leejet
|
47e83d7136
|
sync: update ggml
|
2026-09-27 18:40:09 +08:00 |
|
|
leejet
|
f8890b9db3
|
feat: add Ming-Image Design support (#2063)
|
2026-09-27 17:24:13 +08:00 |
|
|
leejet
|
2f886889e6
|
refactor: align PixArt weights with upstream layout (#2061)
|
2026-09-26 02:03:05 +08:00 |
|
|
leejet
|
19bbbca1c7
|
refactor: define VAE tile dimensions in image pixels (#2059)
|
2026-09-25 18:20:32 +08:00 |
|
|
leejet
|
4c3cf7543d
|
fix: load safetensors index shards without recursion (#2058)
|
2026-09-25 16:40:20 +08:00 |
|
|
leejet
|
510bccf330
|
fix: map Qwen Image 2.1 LoRAs to fused MLP weights (#2057)
|
2026-09-25 16:30:34 +08:00 |
|
|
leejet
|
0a9340c599
|
fix: scale Qwen Image 2.1 VAE convolutions (#2054)
|
2026-09-25 15:52:35 +08:00 |
|
|
leejet
|
b167b942f7
|
fix: align Qwen Image 2.1 flow schedule with official defaults (#2048)
|
2026-09-25 02:43:10 +08:00 |
|
|
leejet
|
1a2330de68
|
fix: reserve 128 MiB headroom when selecting monolithic execution (#2046)
|
2026-09-25 01:53:11 +08:00 |
|
|
leejet
|
740c7ae193
|
feat: add configurable Qwen cache types and early cache scheduling (#2045)
|
2026-09-25 01:51:01 +08:00 |
|
|
leejet
|
88411ef1e0
|
refactor: centralize circular RoPE and extend image model support (#2039)
|
2026-09-24 01:45:14 +08:00 |
|
|
leejet
|
2a4ebba818
|
ci: automatically close PRs from organization-owned forks
|
2026-09-24 00:40:42 +08:00 |
|
|
leejet
|
2dc7f5408a
|
feat: add Qwen Image 2.1 prefix KV cache (#2035)
|
2026-09-23 23:08:21 +08:00 |
|
|
leejet
|
e6281b6318
|
feat: add configurable conditioning cache for all models (#2034)
|
2026-09-23 22:25:42 +08:00 |
|
|
leejet
|
c92d73c408
|
fix: preserve alpha when upscaling RGBA images with ESRGAN (#2029)
|
2026-09-23 02:23:10 +08:00 |
|
|
leejet
|
28b454bda1
|
feat: add configurable image input preprocessing (#2028)
|
2026-09-23 02:12:49 +08:00 |
|
|
leejet
|
ac45422a05
|
fix: honor reference image resize settings in OpenAI edits (#2025)
|
2026-09-22 22:02:21 +08:00 |
|
|
leejet
|
6dcb5bbd42
|
fix: handle GPU memory reports and LLM encoding failures (#2020)
|
2026-09-22 00:28:52 +08:00 |
|
|
leejet
|
97d932b8f8
|
fix: restrict VAE tiling retries to allocation failures (#2019)
|
2026-09-21 23:40:05 +08:00 |
|
|
leejet
|
2726dd35c2
|
fix: update ggml to prevent permute metadata truncation (#2018)
|
2026-09-21 22:19:27 +08:00 |
|
|
leejet
|
74988b290e
|
fix: reject video models in image generation (#2017)
|
2026-09-21 21:48:42 +08:00 |
|
|
leejet
|
c678dfe704
|
docs: update CONTRIBUTING.md
|
2026-09-21 01:30:25 +08:00 |
|
|
leejet
|
b56c68617d
|
fix: avoid narrowing conversion in SigVQ patch embedding and format code
|
2026-09-21 01:27:31 +08:00 |
|
|
leejet
|
187b2561ea
|
feat: add native CUDA SageAttention support (#2005)
|
2026-09-21 01:16:59 +08:00 |
|
|
leejet
|
137f7409bb
|
feat: add Qwen Image 2.1 support (#1994)
|
2026-09-20 22:51:21 +08:00 |
|
|
leejet
|
1330cebae8
|
feat: support building with upstream ggml (#1999)
|
2026-09-19 22:25:01 +08:00 |
|
|
leejet
|
17860c0e45
|
perf: parallelize host tensor elementwise and broadcast ops (#1998)
|
2026-09-19 21:46:16 +08:00 |
|
|
leejet
|
275ab58e01
|
perf: reduce CPU overhead in graph execution and sampling (#1997)
|
2026-09-19 18:32:58 +08:00 |
|
|
leejet
|
9982c9caae
|
fix: propagate CUDA driver dependency to shared library consumers
|
2026-09-19 18:04:59 +08:00 |
|