yindo/rocm-stable-diffusion.cpp
1
0
Fork 0
mirror of https://github.com/BillyOutlast/rocm-stable-diffusion.cpp.git synced 2026-02-04 19:21:19 +01:00
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • 6ad46bb700 sync: update ggml leejet 2025-09-25 21:57:43 +08:00
  • 1ba30ce005 sync: update ggml leejet 2025-09-25 00:38:38 +08:00
  • 2abe9451c4 fix: optimize the handling of CLIP embedding weight (#840) leejet 2025-09-25 00:28:20 +08:00
  • f3140eadbb fix: tensor loading thread count (#854) Wagner Bruna 2025-09-24 13:26:38 -03:00
  • 98ba155fc6 docs: HipBLAS / ROCm build instruction fix (#843) Stefan-Olt 2025-09-24 18:03:05 +02:00
  • 513f36d495 docs: include Vulkan compatibility for LoRA quants (#845) Wagner Bruna 2025-09-24 13:01:10 -03:00
  • 1e0d2821bb fix: correct tensor deduplication logic (#844) rmatif 2025-09-24 17:22:40 +02:00
  • fd693ac6a2 refactor: remove unused --normalize-input parameter (#835) leejet 2025-09-18 00:12:53 +08:00
  • 171b2222a5 fix: avoid segfault for pix2pix models without reference images (#766) Wagner Bruna 2025-09-17 13:11:38 -03:00
  • 567f9f14f0 fix: avoid multithreading issues in the model loader leejet 2025-09-18 00:00:15 +08:00
  • 1e5f207006 chore: fix workflow (#836) leejet 2025-09-17 22:11:55 +08:00
  • 79426d578e chore: set release tag by commit count leejet 2025-09-16 23:24:36 +08:00
  • 97ad3e7ff9 refactor: simplify DPM++ (2S) Ancestral (#667) vmobilis 2025-09-16 18:05:25 +03:00
  • 8909523e92 refactor: move tiling cacl and debug print into the tiling code branch (#833) Erik Scholz 2025-09-16 16:46:56 +02:00
  • 8376dfba2a feat: add sgm_uniform scheduler, simple scheduler, and support for NitroFusion (#675) rmatif 2025-09-16 16:42:09 +02:00
  • 0ebe6fe118 refactor: simplify the logic of pm id image loading (#827) leejet 2025-09-14 22:50:21 +08:00
  • 55c2e05d98 feat: optimize tensor loading time (#790) rmatif 2025-09-14 16:48:35 +02:00
  • 52a97b3ac1 feat: add vace support (#819) leejet 2025-09-14 16:57:33 +08:00
  • 2c9b1e2594 feat: add VAE encoding tiling support and adaptive overlap (#484) stduhpf 2025-09-14 10:00:29 +02:00
  • 288e2d63c0 docs: update docs leejet 2025-09-14 14:24:24 +08:00
  • dc46993b55 feat: increase work_ctx memory buffer size (#814) leejet 2025-09-14 13:19:20 +08:00
  • a6a8569ea0 feat: Add SYCL Dockerfile (#651) Richard Palethorpe 2025-09-14 06:02:59 +01:00
  • 9e7befa320 fix: harden for large files (#643) Erik Scholz 2025-09-14 06:44:19 +02:00
  • c607fc3ed4 feat: use Euler sampling by default for SD3 and Flux (#753) Wagner Bruna 2025-09-14 01:34:41 -03:00
  • b54bec3f18 fix: do not force VAE type to f32 on SDXL (#716) Wagner Bruna 2025-09-14 01:19:59 -03:00
  • 5869987fe4 fix: make weight override more robust against ggml changes (#760) Wagner Bruna 2025-09-14 01:15:53 -03:00
  • 48956ffb87 feat: reduce CLIP memory usage with no embeddings (#768) Wagner Bruna 2025-09-14 01:08:00 -03:00
  • ddc4a18b92 fix: make tiled VAE reuse the compute buffer (#821) Wagner Bruna 2025-09-14 00:41:50 -03:00
  • fce6afcc6a feat: add sd3 flash attn support (#815) leejet 2025-09-11 23:24:29 +08:00
  • 49d6570c43 feat: add SmoothStep Scheduler (#813) Erik Scholz 2025-09-11 17:17:46 +02:00
  • 6bbaf161ad chore: add install() support in CMakeLists.txt (#540) clibdev 2025-09-11 17:24:16 +03:00
  • 87cdbd5978 feat: use log_printf to print ggml logs (#545) clibdev 2025-09-11 17:16:05 +03:00
  • b017918106 chore: remove sd3 flash attention warn (#812) leejet 2025-09-10 22:21:02 +08:00
  • ac5a215998 fix: use {} for params init instead of memset (#781) Wagner Bruna 2025-09-10 10:49:29 -03:00
  • abb36d66b5 chore: update flash attention warnings (#805) Wagner Bruna 2025-09-10 10:38:21 -03:00
  • ff4fdbb88d fix: accept NULL in sd_img_gen_params_t::input_id_images_path (#809) Wagner Bruna 2025-09-10 10:22:55 -03:00
  • abb115cd02 fix: clarify lora quant support and small fixes (#792) Markus Hartung 2025-09-08 16:39:25 +02:00
  • c648001030 feat: add detailed tensor loading time stat (#793) leejet 2025-09-07 22:51:44 +08:00
  • c587a43c99 feat: support incrementing ref image index (omni-kontext) (#755) stduhpf 2025-09-07 16:35:16 +02:00
  • f8fe4e7db9 fix: add flash attn support check (#803) leejet 2025-09-07 21:29:06 +08:00
  • 1c07fb6fb1 docs: update docs/wan.md leejet 2025-09-07 12:07:20 +08:00
  • 675208dcb6 chore: update to c++17 leejet 2025-09-07 12:04:17 +08:00
  • d7f430cd69 docs: update docs and help message leejet 2025-09-07 02:26:44 +08:00
  • 141a4b4113 feat: add flow shift parameter (for SD3 and Wan) (#780) stduhpf 2025-09-06 20:16:59 +02:00
  • 21ce9fe2cf feat: add support for timestep boundary based automatic expert routing in Wan MoE (#779) stduhpf 2025-09-06 19:44:10 +02:00
  • cb1d975e96 feat: add wan2.1/2.2 support (#778) leejet 2025-09-06 18:08:03 +08:00
  • 2eb3845df5 fix: typo in the verbose long flag (#783) Wagner Bruna 2025-09-03 13:49:01 -03:00
  • 4c6475f917 feat: show usage on unknown arg (#767) stduhpf 2025-09-01 15:38:34 +02:00
  • f0fa7ddc40 docs: add compile option needed by Ninja (#770) SmallAndSoft 2025-09-01 16:35:25 +03:00
  • a7c7905c6d docs: add missing dash to docs/chroma.md (#771) SmallAndSoft 2025-09-01 16:34:34 +03:00
  • eea77cbad9 feat: throttle model loading progress updates (#782) Wagner Bruna 2025-09-01 10:32:01 -03:00
  • 0e86d90ee4 chore: add Nvidia 30 series (cuda arch 86) to build NekopenDev 2025-09-01 08:21:34 -05:00
  • 5900ef6605 sync: update ggml, make cuda im2col a little faster leejet 2025-08-03 01:29:40 +08:00
  • 5b8996f74a Conv2D direct support (#744) Daniele 2025-08-02 17:25:17 +00:00
  • f7f05fb185 chore: avoid setting GGML_MAX_NAME when building against external ggml (#751) Wagner Bruna 2025-08-02 14:24:40 -03:00
  • 6167e2927a feat: support build against system installed GGML library (#749) Seas0 2025-08-02 11:03:18 +08:00
  • f6b9aa1a43 refector: optimize the usage of tensor_types leejet 2025-07-28 23:18:29 +08:00
  • 7eb30d00e5 feat: add missing models and parameters to image metadata (#743) Wagner Bruna 2025-07-28 11:00:27 -03:00
  • 59080d3ce1 feat: change image dimensions requirement for DiT models (#742) stduhpf 2025-07-28 15:58:17 +02:00
  • 8c3c788f31 feat: upgrade musa sdk to rc4.2.0 (#732) R0CKSTAR 2025-07-28 21:51:11 +08:00
  • f54524f620 sync: update ggml leejet 2025-07-28 21:50:12 +08:00
  • eed97a5e1d sync: update ggml leejet 2025-07-24 23:03:42 +08:00
  • fb86bf4cb0 docs: add LocalAI to README's UIs (#741) Ettore Di Giacinto 2025-07-24 16:39:26 +02:00
  • bd1eaef93e fix: convert f64 to f32 and i64 to i32 when loading weights leejet 2025-07-24 00:59:38 +08:00
  • ab835f7d39 fix: correct head dim check and L_k padding of flash attention (#736) Erik Scholz 2025-07-23 18:57:45 +02:00
  • 26f3f61d37 docs: add sd.cpp-webui as an available frontend (#738) Daniele 2025-07-23 15:51:57 +00:00
  • 1896b28ef2 fix: make --taesd work (#731) Oleg Skutte 2025-07-14 20:45:22 +04:00
  • 0739361bfe fix: avoid macOS build failed leejet 2025-07-13 20:18:10 +08:00
  • ca0bd9396e refactor: update c api (#728) leejet 2025-07-13 18:48:42 +08:00
  • a772dca27a feat: add Instruct-Pix2pix/CosXL-Edit support (#679) stduhpf 2025-07-12 09:36:45 +02:00
  • 6d84a30c66 feat: overriding quant types for specific tensors on model conversion (#724) Wagner Bruna 2025-07-07 13:11:38 -03:00
  • dafc32d0dd feat: add support for f64/i64 and clip_g diffusers model (#681) stduhpf 2025-07-06 17:24:55 +02:00
  • 225162f270 fix: mark encoder.embed_tokens.weight as unused tensor (#721) idostyle 2025-07-06 17:10:10 +02:00
  • b9e4718fac fix: correct --chroma-enable-t5-mask argument leejet 2025-07-06 11:11:47 +08:00
  • 1ce1c1adca feat: make lora graph size variable leejet 2025-07-05 22:44:22 +08:00
  • 19fbfd8639 feat: override text encoders for unet models (#682) stduhpf 2025-07-04 16:19:47 +02:00
  • 76c72628b1 fix: fix a few typos on cli help and error messages (#714) Wagner Bruna 2025-07-04 11:15:41 -03:00
  • 3bae667f3d fix: break the line after skipping tensors in VAE (#591) vmobilis 2025-07-03 17:50:42 +03:00
  • 8d0819c548 fix: actually use embeddings with SDXL (#657) stduhpf 2025-07-03 16:39:57 +02:00
  • 7a8ff2e819 docs: add golang cgo bindings to README (#635) Binozo 2025-07-02 17:19:49 +02:00
  • 0927e8e322 docs: add Android app to README (#647) rmatif 2025-07-02 17:18:16 +02:00
  • 83ef4e44ce feat: add T5 with llama.cpp naming convention support (#654) stduhpf 2025-07-02 17:13:00 +02:00
  • 7dac89ad75 refector: reuse some code leejet 2025-07-01 23:33:50 +08:00
  • 9251756086 feat: add CosXL support (#683) stduhpf 2025-07-01 17:13:04 +02:00
  • ecf5db97ae chore: fix windows build and release leejet 2025-07-01 23:05:48 +08:00
  • ea46fd6948 fix: force zero-initialize output of tiling (#703) stduhpf 2025-07-01 17:01:29 +02:00
  • 23de7fc44a chore: avoid warnings when building on linux leejet 2025-06-30 23:49:52 +08:00
  • d42fd59464 feat: add OpenCL backend support (#680) rmatif 2025-06-30 17:32:23 +02:00
  • 0d8b39f0ba fix: avoid crash on sdxl loras (#658) Wagner Bruna 2025-06-30 12:29:32 -03:00
  • 539b5b9374 fix: fix musa docker build (#662) R0CKSTAR 2025-06-30 23:27:40 +08:00
  • b1fc16b504 fix: allow resetting clip_skip to its default value (#697) Wagner Bruna 2025-06-30 12:23:21 -03:00
  • d6c87dce5c docs: add chroma doc leejet 2025-06-29 23:58:15 +08:00
  • a28d04dd81 fix: fix the issue in parsing --chroma-disable-dit-mask leejet 2025-06-29 23:52:36 +08:00
  • 45d0ebb30c style: format code leejet 2025-06-29 23:40:55 +08:00
  • b1cc40c35c feat: add Chroma support (#696) stduhpf 2025-06-29 17:36:42 +02:00
  • 884e23eeeb docs: add kontext doc leejet 2025-06-29 10:35:31 +08:00
  • c9b5735116 feat: add FLUX.1 Kontext dev support (#707) stduhpf 2025-06-29 04:08:53 +02:00
  • 10c6501bd0 fix missing argument in prototype of stbi_write_jpg (#613) vmobilis 2025-03-09 07:30:10 +03:00
  • 10feacf031 fix: correct img2img time (#616) vmobilis 2025-03-09 07:29:08 +03:00
  • 655f8a5169 fix: clang complains about needless braces (#618) vmobilis 2025-03-09 07:26:41 +03:00