Commit Graph

  • 44a8f2f8db Include VAE & LoRA data into PNG metadata (#1573) Nelson Sharpe 2023-06-22 19:05:54 -04:00
  • 8822b9acd7 added ability to use config file to shard vicuna (#1565) Eliasj42 2023-06-22 15:40:35 -07:00
  • 0ca3b9fce3 fix some mmap and vicuna bugs (#1576) Daniel Garvey 2023-06-22 17:39:55 -05:00
  • 045f2bb147 Add dispatch-level config file generator for manual annotation (#1566) Nithin Meganathan 2023-06-22 15:11:41 -07:00
  • a811b867b9 Add shark_eager mode. Prashant Kumar 2023-06-22 15:38:22 +00:00
  • cdd505e2dd [SharkInference-SharkRuntime] Adds capability to mmap vmfbs Abhishek Varma 2023-06-15 13:01:58 +00:00
  • 1b0f39107c Move torch_mlir import to the top (#1574) powderluv 2023-06-21 22:31:35 -07:00
  • b9b8955f74 exclude vulkan on macos powderluv 2023-06-21 22:22:27 -07:00
  • 6f7a85eee3 switch to metal backend for CI powderluv 2023-06-21 22:17:11 -07:00
  • 18c8e9e51e Metal typo fix (#1572) 20230621.780 Ranvir Singh Virk 2023-06-21 21:56:11 -07:00
  • a202bb466a fp16 fixes for webui (#1571) Daniel Garvey 2023-06-21 22:24:02 -05:00
  • 07c1e1d712 Adding metal_utils for iree_utils (#1561) Ranvir Singh Virk 2023-06-21 19:09:03 -07:00
  • 18daec78c8 Added check for python version (#1570) Ranvir Singh Virk 2023-06-21 18:56:47 -07:00
  • dd40a3fafe Add simple OPT tuning script. ean-opt-tuner Ean Garvey 2023-06-21 20:02:42 -05:00
  • 1a8e2024d6 Exclude non-square sizes from use_tuned on rdna2 (#1568) Ean Garvey 2023-06-21 11:36:55 -05:00
  • bf6fcc353a fix formatting Ean Garvey 2023-06-21 09:26:33 -05:00
  • 918eba6524 Exclude non-square sizes from use_tuned on rdna2 Ean Garvey 2023-06-21 08:42:19 -05:00
  • d61b6641fb Rest API: Resolved Generator Object not Subscripatable error (#1556) 20230620.779 AyaanShah2204 2023-06-20 19:27:41 -07:00
  • 88cc2423cc Enable Vicuna fp16 cpu (#1562) Phaneesh Barwaria 2023-06-21 00:13:21 +05:30
  • ccf944c1bd Enable tuner for upscaler unet. (#1563) Ean Garvey 2023-06-20 13:40:13 -05:00
  • 0def74f520 [SD] Update unet in_channels API and add PIL metadata to spec. (#1560) Ean Garvey 2023-06-20 12:26:36 -05:00
  • 92320d0b3d Merge branch 'main' into monorimet-patch-2 20230620.778 Ean Garvey 2023-06-20 12:16:16 -05:00
  • 4a1780a080 Include PIL metadata instead of hidden imports in SD spec. Ean Garvey 2023-06-20 12:10:24 -05:00
  • 3fb72e192e Add patch for making compile API work for both MEGABYTE and MiniGPT4 (#1559) Abhishek Varma 2023-06-20 22:34:17 +05:30
  • 458e52d230 Fix deprecation warning for unet config. Ean Garvey 2023-06-20 10:40:36 -05:00
  • 855435ee24 Fix for the user input for Falcon pipeline Vivek Khandelwal 2023-06-20 06:22:42 +00:00
  • 6f9f868fc0 fixed a bug where designating device for vicuna didn't work Elias Joseph 2023-06-19 20:12:29 -04:00
  • fb865f1b99 Move to checkout@v3 20230619.777 powderluv 2023-06-19 18:44:36 -07:00
  • 3e5c50f07b changes for tuning (#1542) rprasad2 2023-06-19 13:29:08 -07:00
  • a544f30a8f Move mega to the shark examples (#1555) powderluv 2023-06-19 11:10:51 -07:00
  • 1fe56d460a [MEGABYTE] Add script to compile MEGABYTE through SHARK (#1553) Abhishek Varma 2023-06-19 23:30:35 +05:30
  • fafd713141 Minor change to falcon pipeline Vivek Khandelwal 2023-06-19 17:03:45 +00:00
  • 015d0132c3 Modify falcon pipeline to add fp16 support (#1551) Vivek Khandelwal 2023-06-19 22:27:13 +05:30
  • 20ddd96ef7 unpin diffusers (#1550) powderluv 2023-06-18 13:45:55 -07:00
  • ee33cfd2d1 Add PIL in main index.py (#1549) 20230618.775 powderluv 2023-06-18 11:51:44 -07:00
  • a3cba21d5b Fix load of unet512 vmfb fail on get of iree opts (#1546) 20230618.774 Stefan Kapusniak 2023-06-18 14:42:20 +01:00
  • a7b6ec4095 Fix unet512 always being used when --max_length=77 (#1547) Stefan Kapusniak 2023-06-18 14:41:25 +01:00
  • d80b087d95 Add PIL hidden imports to sd spec. (#1544) Ean Garvey 2023-06-18 08:39:08 -05:00
  • 297a209608 Remove workarounds for gradio tempfile bugs (#1548) 20230617.773 Stefan Kapusniak 2023-06-18 03:50:36 +01:00
  • b204113563 Add UNet512 (#1504) gpetters94 2023-06-17 03:46:25 -04:00
  • f60ab1f4fa Add Deberta to stablehlo in shark tank (#1545) 20230616.772 Chi_Liu 2023-06-16 13:24:44 -07:00
  • 288602ad1d Add PIL hidden imports to sd spec. 20230616.771 Ean Garvey 2023-06-16 11:04:22 -05:00
  • b203779462 Added Adreno target triples to vulkan_utils (#1543) 20230615.770 Surya Jasper 2023-06-15 16:42:59 -07:00
  • 493f776253 add pretrain_models.csv ean-dynamo-test Ean Garvey 2023-06-15 14:52:18 -05:00
  • 6a82667778 Add pretraining pytest script with SHARK dynamo. Ean Garvey 2023-06-15 11:59:35 -05:00
  • 38570a9bbb Some Fixes for update to gradio 3.34.0 (#1538) Stefan Kapusniak 2023-06-15 09:10:36 +01:00
  • a5c882f296 Bump gradio from 3.15.0 to 3.34.0 (#1518) 20230614.769 dependabot[bot] 2023-06-14 18:13:48 -07:00
  • eb6d11cfed Change mlir dialects for tf tests to stablehlo. (#1535) Ean Garvey 2023-06-14 12:43:49 -05:00
  • 46184a81ac Add Falcon pipeline (#1534) 20230614.768 Vivek Khandelwal 2023-06-14 22:09:16 +05:30
  • 149165a2f0 add multi-device mutli-precision vmfb names PhaneeshB 2023-06-14 20:01:07 +05:30
  • bec82a665f mega vicuna merge dan 2023-06-12 19:26:37 -05:00
  • 9551490341 Remove deprecared --iree-mhlo-demote-164-to-132 flag usage. (#1533) 20230613.767 Ean Garvey 2023-06-13 22:40:47 -05:00
  • 49b3ecdbca (pytest) don't run redundant tests in cpu suite (#1532) Ean Garvey 2023-06-13 22:40:33 -05:00
  • f53e3594c3 OPT Refactor (#1516) Ean Garvey 2023-06-13 22:40:07 -05:00
  • 5562d1dfda Fix xfails for cpu pytest cases (#1527) Ean Garvey 2023-06-13 19:01:51 -05:00
  • c7b0c2961e UI/Web Improve output gallery temp file handling (#1531) Stefan Kapusniak 2023-06-13 22:25:37 +01:00
  • 44273b0791 Fix conditional in transform_fx() (#1530) Ean Garvey 2023-06-13 16:24:53 -05:00
  • 0a4c8fcb3e Minor changes in the fx transforms. 20230613.766 Prashant Kumar 2023-06-13 15:22:40 +00:00
  • 2fec3c8169 re-indents add_upcast in shark importer (#1523) Stefan Kapusniak 2023-06-12 20:41:10 +01:00
  • 5e7d5930dd [vicuna] Add device and precision propagation in vicuna (#1520) Gaurav Shukla 2023-06-12 22:44:43 +05:30
  • b6dbd20250 Modify the fx transforms. (#1521) Prashant Kumar 2023-06-12 22:10:14 +05:30
  • 34f1295349 Add a model config generator (#1511) Nithin Meganathan 2023-06-09 15:32:00 -07:00
  • 1980d7b2c3 Cpu device map (#1515) Phaneesh Barwaria 2023-06-09 21:57:02 +05:30
  • 2cfacc5051 fix osx torch_mlir (#1513) powderluv 2023-06-09 00:57:26 -07:00
  • 436f58ddc4 cli using generate and mem fixes (#1509) 20230608.761 Phaneesh Barwaria 2023-06-08 23:43:32 +05:30
  • 6b29bd17c8 Enable compilation vicuna (#1507) 20230607.760 Phaneesh Barwaria 2023-06-08 01:38:22 +05:30
  • 2c3485ca3e Add standalone OPT sentence completion script. (#1506) Ean Garvey 2023-06-07 12:58:03 -05:00
  • f206ecc635 reenable compilation in vicuna pipeline, add flags (#1505) Daniel Garvey 2023-06-07 11:49:27 -05:00
  • a187e05ae6 Prevent having no cuda devices breaking the UI (#1503) 20230606.759 Stefan Kapusniak 2023-06-06 19:41:16 +01:00
  • 8c21960486 [vicuna] Set only cuda devices in vicuna UI for now Gaurav Shukla 2023-06-06 22:04:20 +05:30
  • be62fce676 [vicuna] Fix vicuna chatbot (#1499) Gaurav Shukla 2023-06-06 21:53:32 +05:30
  • f23b778a6c remove old vicuna scripts PhaneeshB 2023-06-06 21:20:24 +05:30
  • 436edf900d add vic sharded pipeline PhaneeshB 2023-06-05 17:53:55 +05:30
  • ed58c2553f [vicuna] Integrate vicuna in shark studio Gaurav Shukla 2023-06-06 16:42:34 +05:30
  • f2ca58e844 Add .csv and .json param info to output gallery (#1495) Stefan Kapusniak 2023-06-06 15:08:34 +01:00
  • 1dbcc736eb [SD] (RDNA2) Enable new tuning for sd1.4 (#1498) Ean Garvey 2023-06-06 08:48:58 -05:00
  • a83808ddc5 Vicuna cuda on A100 40G (#1496) Phaneesh Barwaria 2023-06-06 15:10:33 +05:30
  • a07fe80530 Update OPT, ResNet example scripts. (#1492) 20230605.757 20230605.756 Ean Garvey 2023-06-05 22:19:35 -05:00
  • d0ba3ef8fa disable use_tuned on SD1.4 for rdna2 (#1490) Ean Garvey 2023-06-05 19:46:16 -05:00
  • 8400529c2c Fix output gallery not using shark_tmp (#1493) Stefan Kapusniak 2023-06-05 22:23:49 +01:00
  • 7eaee9c242 update SHARK to nodai SHARK 20230605.755 powderluv 2023-06-05 00:44:49 -07:00
  • 8230eebce5 Switch to CPU torch builds for shark.whl 20230605.754 powderluv 2023-06-05 00:36:03 -07:00
  • 6296ea4be9 fix config handling for sd1.4 on rdna2 (#1489) Ean Garvey 2023-06-05 02:02:30 -05:00
  • 4151ec3a8f (pytest) tag efficientnet, mobilenet as xfails on vulkan (#1488) Ean Garvey 2023-06-05 01:22:32 -05:00
  • a2467e8d43 Enable SHARK whl packages powderluv 2023-06-04 23:21:22 -07:00
  • e677178bcc Replace RDNA2 SD lowering configs. (#1486) Ean Garvey 2023-06-05 00:57:43 -05:00
  • 7ef1bea953 XFAIL some macos tests 20230604.752 Anush Elangovan 2023-06-04 15:27:03 -07:00
  • ad89bb1413 Add distilgpt2 to stablehlo in shark tank (#1481) 20230603.751 20230602.750 Chi_Liu 2023-06-02 14:44:46 -07:00
  • 218ed78c40 Change instances of input_type='mhlo' to 'auto' (#1482) Ean Garvey 2023-06-02 16:43:47 -05:00
  • 6046f36ab6 UI/Web: Fix upscaler stop button (mostly) (#1479) Stefan Kapusniak 2023-06-02 06:26:55 +01:00
  • 5915bf7de3 Add to and tweak vulkan configuration environments. (#1475) Foxlum 2023-06-02 00:25:20 -05:00
  • f0a4e59758 LLM Pipeline Wrapper (#1477) 20230601.749 20230531.748 20230531.747 Phaneesh Barwaria 2023-05-31 22:47:20 +05:30
  • 1ddef26af5 Web/UI: Add an Output Gallery tab for SD (#1470) 20230531.746 Stefan Kapusniak 2023-05-30 21:47:48 +01:00
  • ba8eddb12f Add GPT3/OPT to Stablehlo in shark tank (#1468) 20230530.745 Chi_Liu 2023-05-29 21:58:39 -07:00
  • 47b346d428 Modify the lowering config format for SPIRVMatmulPromoteVectorize pipeline (#1471) yzhang93 2023-05-29 21:53:48 -07:00
  • 1b4f4f5f4d Fix download path for SD1.4 Unet. (#1469) 20230528.744 20230527.743 20230526.742 Ean Garvey 2023-05-26 13:59:51 -05:00
  • 73cd7e8320 added full vicuna to vicuna.py Elias Joseph 2023-05-26 05:13:40 +00:00
  • 19c0ae3702 Cleanup SD pipeline utils (#1466) 20230525.741 Ean Garvey 2023-05-25 12:50:11 -05:00
  • 54e57f7771 Revive SD downloads from shark_tank. (#1465) Ean Garvey 2023-05-25 12:03:21 -05:00
  • 6d64b8e273 vic and slm common generation base PhaneeshB 2023-05-17 21:50:45 +05:30