github/AMD-SHARK-Studio
1
1
Fork 0
mirror of https://github.com/nod-ai/AMD-SHARK-Studio.git synced 2026-02-19 11:56:43 -05:00
Code Issues Packages Projects Releases Wiki Activity
976 Commits 66 Branches 705 Tags
enable_caching_alloc
Commit Graph

279 Commits

Author SHA1 Message Date
Prashant Kumar
3595b4aaff Incorporate latest changes in the shark_dynamo backend. 2023-02-08 20:37:30 +05:30
Quinn Dawkins
fbd6224b04 Revert "Revert pipelines (#948)" (#951) 
This reverts commit 8115b26079.
Additionally fixes img2col by adding detach elementwise from named op
passes.
 2023-02-04 22:44:08 -05:00
powderluv
8115b26079 Revert pipelines (#948) 
* Revert "[SD] Modify the flags to use --iree-preprocessing-pass-pipeline (#914)"

This reverts commit a783c089a9.

* Revert "Fix iree flags due to the change in shark-runtime (#944)"

This reverts commit 1d38d49162.
 2023-02-04 07:09:51 -08:00
yzhang93
1d38d49162 Fix iree flags due to the change in shark-runtime (#944) 2023-02-03 21:34:02 -08:00
powderluv
e26a70aa4f Drop old cli and webui (#911) 2023-02-01 13:13:46 -08:00
Daniel Garvey
6a32a4e26c move ci sd stuff to apps (#912) 
Co-authored-by: dan <dan@nod-labs.com>
Co-authored-by: powderluv <powderluv@users.noreply.github.com>
 2023-02-01 12:15:07 -08:00
powderluv
e853abf98b Update stable_diffusion_amd.md 2023-02-01 11:11:58 -08:00
Daniel Garvey
e374074013 Windows test (#896) 
* add generate_sharktank for stable_diffusion model defaults

* add windows test for sd

---------

Co-authored-by: dan <dan@nod-labs.com>
 2023-02-01 12:03:54 -06:00
Evan Guan
8cafe56eb4 Added flags for metadata information. (#894) 2023-02-01 05:16:11 -08:00
powderluv
1a37675435 Revert "move beta to release (#898)" (#905) 
This reverts commit 7edcaf5a06.
 2023-01-31 20:31:41 -08:00
Ean Garvey
a504903dd5 Fix formatting issues. (#903) 2023-02-01 09:12:45 +05:30
Daniel Garvey
7edcaf5a06 move beta to release (#898) 
Co-authored-by: dan <dan@nod-labs.com>
 2023-01-31 17:14:08 -06:00
Gaurav Shukla
c124b76328 [SD] Reorganize the stable diffusion model. (#806) 
The stable diffusion codebase has been reorganized to make it more
modular so that the same script can be used for web as well as cli,
instead of duplicating the whole codebase.

Signed-off-by: Gaurav Shukla <gaurav@nod-labs.com>
 2023-01-31 14:42:41 -08:00
Daniel Garvey
a4634632ba add generate_sharktank for stable_diffusion model defaults (#742) 
Co-authored-by: dan <dan@nod-labs.com>
Co-authored-by: powderluv <powderluv@users.noreply.github.com>
 2023-01-31 09:44:54 -08:00
Abhishek Varma
d17e8dc5ad [NFC] Rename SD negative_prompts flag 
-- This commit renames SD `negative-prompts` -> `negative_prompts` flag.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-31 21:38:59 +05:30
Abhishek Varma
fcd62513cf [SD-CLI] Add support for .safetensors + Use diffusers pipeline to load SD 
-- This commit uses `load_pipeline_from_original_stable_diffusion_ckpt`
   as exposed due to [Diffusers PR](https://github.com/huggingface/diffusers/pull/2019).
-- It also adds a support for the end users to use `.safetensors` along
   with `.ckpt` file.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-31 00:00:37 +05:30
Abhishek Varma
c3c701e654 Update requirements.txt + README.md of SD 
-- This commit includes two python modules as part of requirements.txt.
-- It also updates README.md to also inclue `--no-use_tuned` for users to
   be able to try `hf_model_id` or `ckpt_loc` without any issue.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-30 14:12:54 +05:30
yzhang93
9644e78545 Fix CUDA tuned model annotation (#880) 2023-01-27 11:35:18 -08:00
dymil
c911189ef0 Add note about latest RDNA3 driver support (#881) 
Also tweak other wording
 2023-01-27 09:39:19 -08:00
Abhishek Varma
1118b4b651 [SD-CLI] Clean up vmfbs if a retry method fails 
-- This commit cleans up vmfb files generated as a result of retry method.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-27 21:55:36 +05:30
PhaneeshB
4be75d4418 fix seed values in SD json and filename 2023-01-27 18:40:26 +05:30
yzhang93
fee73b0b63 Add SD model annotation on fly (#869) 
* Add SD model annotation on fly

* Move tuned_compile_through_fx to utils

* Fix SD compilation flags
 2023-01-26 11:46:36 -08:00
powderluv
9bbffa519e Add an option to respect LLPC env var (#875) 
Also add OSX paths
 2023-01-25 13:56:55 -08:00
yzhang93
aafe7c4701 Add more cuda devices to use tuned model (#868) 2023-01-25 06:36:17 -08:00
Abhishek Varma
9a0b082cf8 [SD-CLI] Add batch_size command-line arg + prompt processing 
-- This commit adds `batch_size` command-line arg.
-- It also involves replicating the prompt `batch_size` no. of times.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-25 19:21:25 +05:30
powderluv
8265e34a29 Add SHARK SD CLI tool (#870) 2023-01-24 23:14:32 -08:00
powderluv
8ef8ae097f Update to build 469 2023-01-24 22:16:13 -08:00
Abhishek Varma
6ed02f70ec [SD-CLI] Make using ckpt_loc and hf_model_id easier 
-- Currently we require users to specify the base model on which the custom
   model (.ckpt) is tuned on. Even for running a HuggingFace repo-id, we
   require the users to go a tedious way of adding things to variants.json.

-- This commit aims to address the above issues and will be treated as a
   starting point for a series of design changes which makes using SHARK's SD
   easier.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-24 23:03:46 +05:30
Prashant Kumar
cb78cd8ac0 Add the support for the batch size parameter. 2023-01-24 22:33:13 +05:30
Abhishek Varma
e3e4ea5443 Update README.md 
-- Make usage of `hf_model_id` clearer.
 2023-01-23 23:25:23 +05:30
Prashant Kumar
a3e4ea3228 Remove the dependency of the torchvision. (#858) 
Remove the dependency of torchvision library for the conversion
of tensor layout format to what PIL library expects.
 2023-01-23 08:49:57 -08:00
powderluv
56f16d6baf Update SD readme 2023-01-23 06:51:54 -08:00
Abhishek Varma
7a55ab900e [SD-CLI] Fix CKPT script + add more variants + update README.md 
-- This commit fixes CKPT script to rely on the previous CKPT to Diffusers
   script.
   TODO: Let go of the script once the CKPT is included in next release
         of diffusers.
-- It also adds many variants as part of `variants.json` and updates
   `README.md` to reflect change in default `hf_model_id`.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-23 18:34:24 +05:30
Abhishek Varma
137643fe72 [SD-CLI] Update README.md of custom models to include hf_model_id 2023-01-23 11:37:13 +05:30
Erkin Alp Güney
8259f08864 Collapsibles for Win10 and Linux users (#851) 
Co-authored-by: powderluv <powderluv@users.noreply.github.com>
 2023-01-22 09:50:33 -08:00
Prashant Kumar
b3ab0a1843 Add width and height support for the scheduler. 2023-01-22 23:16:50 +05:30
Daniel Garvey
e842c8c19b add main.py testing for sdiff (#836) 
Co-authored-by: dan <dan@nod-labs.com>
Co-authored-by: powderluv <powderluv@users.noreply.github.com>
 2023-01-22 01:16:17 -08:00
powderluv
f6c3112d44 Revert "potential fix to pre-load DLL dir for torch-mlir (#848)" (#852) 
This reverts commit 6c470d8131.
 2023-01-22 00:09:35 -08:00
yzhang93
7059610632 Modify the default for --hf_model_id flag 2023-01-21 11:21:47 +05:30
powderluv
2d272930d9 Update to signed build 455 2023-01-20 16:50:42 -08:00
powderluv
6c470d8131 potential fix to pre-load DLL dir for torch-mlir (#848) 
Doesn't regress the main.py script but system already pre-loaded
the DLL so needs more testing.
 2023-01-20 14:48:45 -08:00
stanley
c4a9365aa1 [Shark][Training] Refresh SharkTrainer to latest APIs. 2023-01-19 20:30:15 +00:00
Prashant Kumar
9d3af37104 bugfix related to the height width params. 2023-01-20 00:21:44 +05:30
Prashant Kumar
7b3d57cff7 Add height and width as args. 2023-01-19 23:43:29 +05:30
Abhishek Varma
a802270da9 [SD-CLI] Update README.md about variants.json 2023-01-19 22:46:54 +05:30
Abhishek Varma
dd194a8758 [SD-CLI] Reorder loading of opt_params when needed 
-- This commit reorders loading of opt_params when `import_mlir` is not used.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-19 22:02:51 +05:30
Abhishek Varma
6de02de221 [SD-CLI] Make using custom models easier 
-- This commit makes using custom models easier using a combination of
   `import_mlir`, `ckpt_loc` and `hf_model_id`.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-19 22:02:36 +05:30
Abhishek Varma
85259750bf [SD-CLI] Fix variants.json mapping 
-- This commit fixes variants.json's mapping.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-19 22:02:36 +05:30
Prashant Kumar
1249f0007d Remove args.variant and args.version with args.custom_model. 2023-01-19 19:55:12 +05:30
Abhishek Varma
db0514d3fa [SD-CLI] Fix get_model_configuration to use max_length 
-- This commit fixes `get_model_configuration` to use `max_length`.

Signed-off-by: Abhishek Varma <abhishek@nod-labs.com>
 2023-01-19 19:10:04 +05:30