Comments (7)
several hours? this is a known driver issue that usually resolves in 8-20mins.
Can you share more info about hardware/versions so we can try to recreate this? Thanks
from shark.
@monorimet
Yes. Latest AMD driver to date: 23.8.1
OS: Windows 11
GPU: Radeon 7900 XTX
CPI: AMD 7950X
from shark.
Here are the logs, which appear after I press the button "Generate Image(s)":
Found device AMD Radeon RX 7900 XTX. Using target triple rdna3-7900-windows.
Using tuned models for stabilityai/stable-diffusion-2-1(fp16) on device vulkan://00000000-0300-0000-0000-000000000000.
No internet connection. Using the model already present in the tank.
Model artifacts for euler_scale_model_input_fp16 found at C:\Users\f1am3d\.local/shark_tank/...
No internet connection. Using the model already present in the tank.
Verifying that model artifacts were downloaded successfully to C:\Users\f1am3d\.local/shark_tank/euler_scale_model_input_fp16_torch\euler_scale_model_input_fp16_torch.mlir...
loading existing vmfb from: C:\Users\f1am3d\Downloads\stable-diffusion\shark\euler_scale_model_input_fp16.vmfb
WARNING: [Loader Message] Code 0 : windows_read_data_files_in_registry: Registry lookup failed to get layer manifest files.
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\euler_scale_model_input_fp16.vmfb...
No internet connection. Using the model already present in the tank.
Model artifacts for euler_step_fp16 found at C:\Users\f1am3d\.local/shark_tank/...
No internet connection. Using the model already present in the tank.
Verifying that model artifacts were downloaded successfully to C:\Users\f1am3d\.local/shark_tank/euler_step_fp16_torch\euler_step_fp16_torch.mlir...
loading existing vmfb from: C:\Users\f1am3d\Downloads\stable-diffusion\shark\euler_step_fp16.vmfb
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\euler_step_fp16.vmfb...
use_tuned? sharkify: True
_1_64_512_512_fp16_tuned_stable-diffusion-2-1-base
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
No internet connection. Using the model already present in the tank.
Model artifacts for clip_1_64_512_512_fp16_stable-diffusion-2-1-base_vulkan found at C:\Users\f1am3d\.local/shark_tank/...
No internet connection. Using the model already present in the tank.
Verifying that model artifacts were downloaded successfully to C:\Users\f1am3d\.local/shark_tank/clip_1_64_512_512_fp16_stable-diffusion-2-1-base_vulkan\clip_1_64_512_512_fp16_stable-diffusion-2-1-base_vulkan_torch.mlir...
loading existing vmfb from: C:\Users\f1am3d\Downloads\stable-diffusion\shark\clip_1_64_512_512_fp16_stable-diffusion-2-1-base_vulkan.vmfb
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\clip_1_64_512_512_fp16_stable-diffusion-2-1-base_vulkan.vmfb...
stablediffusion/tuned/stablediffusion/v2_1base/unet/fp16/length_64/tuned is not present in the models database
download pipeline failed, falling back to import_mlir
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\unet_1_64_512_512_fp16_tuned_stable-diffusion-2-1-base_vulkan.vmfb...
::: Detailed report (took longer than 2.5s):
+1.039266586303711ms: get_iree_runtime_config
+3.540515899658203ms: mmap C:\Users\f1am3d\Downloads\stable-diffusion\shark\unet_1_64_512_512_fp16_tuned_stable-diffusion-2-1-base_vulkan.vmfb
+3.540515899658203ms: ireert.SystemContext created
+12555.148601531982ms: module initialized
50it [00:03, 16.37it/s]
stablediffusion/tuned/stablediffusion/v2_1base/vae/fp16/length_77/tuned is not present in the models database
download pipeline failed, falling back to import_mlir
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\vae_1_64_512_512_fp16_tuned_stable-diffusion-2-1-base_vulkan.vmfb...
from shark.
This message happens to be the last one that prints before the generated image is returned. Does it ever generate an image? Also, would you mind trying another model to see if you experience the issue there as well?
from shark.
This message happens to be the last one that prints before the generated image is returned. Does it ever generate an image? Also, would you mind trying another model to see if you experience the issue there as well?
Yes, it generates some mess.
I will try other models and will let you know.
from shark.
@monorimet Okay, here are some results so far:
- anything v3.0 (not working)
- stable-diffustion-2-1-base (not working)
- stable-diffusion-2-1 (not working)
- open journey (working!)
Working log:
To create a public link, set `share=True` in `launch()`.
Found device AMD Radeon RX 7900 XTX. Using target triple rdna3-7900-windows.
Using tuned models for prompthero/openjourney(fp16) on device vulkan://00000000-0300-0000-0000-000000000000.
loading existing vmfb from: C:\Users\f1am3d\Downloads\stable-diffusion\shark\euler_scale_model_input_1_512_512_vulkan_fp16.vmfb
WARNING: [Loader Message] Code 0 : windows_read_data_files_in_registry: Registry lookup failed to get layer manifest files.
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\euler_scale_model_input_1_512_512_vulkan_fp16.vmfb...
loading existing vmfb from: C:\Users\f1am3d\Downloads\stable-diffusion\shark\euler_step_1_512_512_vulkan_fp16.vmfb
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\euler_step_1_512_512_vulkan_fp16.vmfb...
use_tuned? sharkify: True
_1_64_512_512_fp16_tuned_openjourney
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\clip_1_64_512_512_fp16_tuned_openjourney_vulkan.vmfb...
Downloading (…)ch_model.safetensors: 100%|████████████████████████████████████████| 3.44G/3.44G [02:20<00:00, 24.5MB/s]
huggingface_hub\file_download.py:133: UserWarning: `huggingface_hub` cache-system uses symlinks by default to efficiently store duplicated files but your machine does not support them in C:\Users\f1am3d\.cache\huggingface\hub. Caching files will still work but in a degraded version that might require more space on your disk. This warning can be disabled by setting the `HF_HUB_DISABLE_SYMLINKS_WARNING` environment variable. For more details, see https://huggingface.co/docs/huggingface_hub/how-to-cache#limitations.
To support symlinks on Windows, you either need to activate Developer Mode or to run Python as an administrator. In order to see activate developer mode, see this article: https://docs.microsoft.com/en-us/windows/apps/get-started/enable-your-device-for-development
mat1 and mat2 shapes cannot be multiplied (128x1024 and 768x320)
Retrying with a different base model configuration
Loading Winograd config file from C:\Users\f1am3d\.local/shark_tank/configs\unet_winograd_vulkan.json
100%|█████████████████████████████████████████████████████████████████████████████████| 107/107 [00:00<00:00, 9.30kB/s]
Loading lowering config file from C:\Users\f1am3d\.local/shark_tank/configs\unet_v1_4_64_fp16_vulkan_rdna3_512x512.json
100%|█████████████████████████████████████████████████████████████████████████████| 24.2k/24.2k [00:00<00:00, 43.6kB/s]
Applying tuned configs on unet_1_64_512_512_fp16_tuned_openjourney_vulkan
No vmfb found. Compiling and saving to C:\Users\f1am3d\Downloads\stable-diffusion\shark\unet_1_64_512_512_fp16_tuned_openjourney_vulkan.vmfb
Configuring for device:vulkan://00000000-0300-0000-0000-000000000000
Using target triple -iree-vulkan-target-triple=rdna3-7900-windows from command line args
Saved vmfb in C:\Users\f1am3d\Downloads\stable-diffusion\shark\unet_1_64_512_512_fp16_tuned_openjourney_vulkan.vmfb.
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\unet_1_64_512_512_fp16_tuned_openjourney_vulkan.vmfb...
::: Detailed report (took longer than 2.5s):
+2.9997825622558594ms: get_iree_runtime_config
+13.107061386108398ms: mmap C:\Users\f1am3d\Downloads\stable-diffusion\shark\unet_1_64_512_512_fp16_tuned_openjourney_vulkan.vmfb
+17.12322235107422ms: ireert.SystemContext created
+12111.94920539856ms: module initialized
50it [00:03, 13.52it/s]
Downloading (…)main/vae/config.json: 100%|█████████████████████████████████████████████| 547/547 [00:00<00:00, 547kB/s]
Downloading (…)ch_model.safetensors: 100%|██████████████████████████████████████████| 335M/335M [00:35<00:00, 9.32MB/s]
Loading Winograd config file from C:\Users\f1am3d\.local/shark_tank/configs\vae_winograd_vulkan.json
100%|███████████████████████████████████████████████████████████████████████████████| 33.0/33.0 [00:00<00:00, 3.28kB/s]
Loading lowering config file from C:\Users\f1am3d\.local/shark_tank/configs\vae_fp16_vulkan_rdna3.json
100%|██████████████████████████████████████████████████████████████████████████████| 2.08k/2.08k [00:00<00:00, 193kB/s]
Applying tuned configs on vae_1_64_512_512_fp16_tuned_openjourney_vulkan
No vmfb found. Compiling and saving to C:\Users\f1am3d\Downloads\stable-diffusion\shark\vae_1_64_512_512_fp16_tuned_openjourney_vulkan.vmfb
Configuring for device:vulkan://00000000-0300-0000-0000-000000000000
Using target triple -iree-vulkan-target-triple=rdna3-7900-windows from command line args
Saved vmfb in C:\Users\f1am3d\Downloads\stable-diffusion\shark\vae_1_64_512_512_fp16_tuned_openjourney_vulkan.vmfb.
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\vae_1_64_512_512_fp16_tuned_openjourney_vulkan.vmfb...
50it [00:02, 16.67it/s]
from shark.
Okay, I figured this out. Now it simply writes the message
Loading module C:\Users\f1am3d\Downloads\stable-diffusion\shark\vae_1_64_512_512_fp16_tuned_stable-diffusion-2-1-base_vulkan.vmfb...
But the module is actually loaded. It's confusing, since it doesn't output any message after that, like "done" or "completed".
Anyway, most of the models are working, however all basic stable diffusion models are producing weird results, as I mentioned here.
from shark.
Related Issues (20)
- RESOURCE_EXHAUSTED when trying to use hires fix
- SDXL Error
- Sharkify Model UnetModel.forward() Arguments Error
- Input image masks on Image2Image does not seem to do anything HOT 1
- SHARK won't start or run without internet
- Error while installing rx570 don't know if it is because of the card or software issue please fix it
- LoRA models not showing up in dropdown
- System interrupts gets very high CPU followed by complete computer shutdown HOT 3
- Document for the compilation flow ?
- Enable llama2 benchmarking with Turbine HOT 1
- FP8 Support HOT 10
- (Shark-1.0) setup_venv.ps1 specifies a torch-mlir version that has dropped off the end of package index
- RuntimeError: duplicate registrations HOT 3
- Gradio Button has no attr "label"
- Gradio JSON has no option interactive
- problem with Torch-mlir on Ubuntu, latest en previous Shark version
- (Studio2) Centralize and minimize prompt handling for LLMs
- (Studio2) (Windows) 'rocm devices are not available'
- Drastic difference in images generated in 984 vs 1091
- Can't run SDXL Turbo at all. My 32GB of RAM at 100% and fails. HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from shark.