I have 6GB memory in gtx2060. trying to use a 1024x659 image. used smallest brush with just one stroke getting this memory error

RuntimeError: CUDA out of memory. Tried to allocate 16.00 MiB (GPU 0; 6.00 GiB total capacity; 4.02 GiB already allocated; 0 bytes free; 4.38 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

is it that memory taxing or there is an error?

I understand that the LDM model is intensive but I think something peculiar is happening. It always OOM's on the 3rd stroke. I tried lowering the steps but it doesn't matter. No matter what the step size is or the stroke size is, the model crashes on the 3rd stroke with an OOM.

Wondering if there might be a fix for that? I'm on a RTX3080.

I have installed cuda, but the WARNING-- “Did not find CUDA device on your computer, fallback to cpu tasks.py:86 Start as desktop app? [y/n] (y):” so，How to use GPU to run this program

I'd love it if there was a manual "Clean up" button, rather than having to select everything within one selection.

For example, I'd mask several different items, and then click on Clean up, rather than mask the first one -> wait for clean up > mask the second one > wait for clean up, and so on...

Also, sometimes when zoomed in, I can't reach and mask all of the object so I have to let go of the mouse click which automatically starts the clean up process, rather than letting me continue it and then manually deciding when to clean up

Hope this makes sense :)

This patch brings in a massive number of changes to the frontend of the application. Please feel free to discuss the proposed changes with me at any time.

Implemented Recoil as a state management system.

Why Recoil? It is a robust library built by developers at Facebook for state management. It has an extremely simple API for implementation that is in sync with React syntax compared to any other state management system out there and works amazingly well. While the official release status is beta as it becomes fully featured, the library is already used in various systems at Facebook and is very stable for the use cases of this application.

Why global state management? One of the major issues I saw with the current file structure is that there is minimal code splitting and it makes further development of the frontend a cumbersome task. I have broken down the frontend into various easy to access components isolating the GUI from the logic. To avoid prop drilling, we need global state management to handle the necessary tasks. This will also facilitate the addition of any new features greatly.

Code Splitting. Majority of the components that can be isolated in the application have now been done so.

All New Custom CSS & Removal of Tailwind

While Tailwind is a great way to deploy beautiful interfaces quickly, anyone trying to stylize the application further needs to be familiar with Tailwind which makes it harder for more people to work on it. Not to mention, I am not a particular fan of flooding JSX elements with inline CSS classes. That makes reading the code extremely hard and bloats up component code drastically.

As a replacement to Tailwind, I implemented a custom styling system using SCSS as a developer dependency.

In the new system, all the general and shared styles are in the styles folder and all the component styles are in the same folder as the component for easy access. The _index.scss file now acts as a central import for every other stylesheet that needs to be loaded.

What Changed?

• The entire application looks and feels like the current implementation with minimal changes.
• The green (#bdff01) highlight used in the application has now been changed to a bright yellow (rgb(255, 190, 0)) because I felt it better suited the new Dark Mode (see below).
• The swipe bar for comparing before and after images has now been removed and instead the comparison is a smooth fade effect. I felt this was better to analyze image changes rather than a swiper. // Can add the swipe back if needed.

Dark Mode

A brand new Dark Mode has been added for the application. Users can enable and disable by tapping the button in the header or by using the Shift + D hotkey.

Other New Features

• When the editor image is now zoomed out to its default size, the image now also gets centered back.
• Various minor bug and QoL fixes

TODO - Stuff I might work on next

• The currently used react-zoom-pinch-pan module is not mobile friendly. It does not allow brush strokes. Need to figure out a way to fix this.
• Further optimization of the frontend code with better code splitting and performance.
• When using the LaMa model, the first stroke has a delayed response from the backend but the ones that follow are almost immediate. I believe this is happening because of the initialization of the model on the first stroke. I wonder if either of us can look at it and see if this can somehow be preloaded so the user experience is smooth from the first stroke.
• Enable threading for the desktop application mode so flaskwebgui does not block the main applications Python console.

When used as a local tool, most of the time you may only need to process a single photo, and it could be a bit more convenient if it could support passing in the image path at startup, eliminating the need to manually select the path. Let's say the following.

python3 main.py --device=cuda --port=8080 --model=lama --image path/example.jpg

python3 main.py --device=cuda --port=8080 --model=ldm --ldm-steps=50 --image path/example.jpg

I updated from V0.14.0 to V0.15.0 and try to run it using: python main.py

(lama015) C:\Users\username\Documents\Ai_Ml\lama-cleaner-0.15>python main.py OMP: Error #15: Initializing libiomp5md.dll, but found libiomp5md.dll already initialized. OMP: Hint This means that multiple copies of the OpenMP runtime have been linked into the program. That is dangerous, since it can degrade performance or cause incorrect results. The best thing to do is to ensure that only a single OpenMP runtime is linked into the process, e.g. by avoiding static linking of the OpenMP runtime in any library. As an unsafe, unsupported, undocumented workaround you can set the environment variable KMP_DUPLICATE_LIB_OK=TRUE to allow the program to continue to execute, but that may cause crashes or silently produce incorrect results. For more information, please see http://www.intel.com/software/products/support/.

I installed it using conda new env and installed pytorch conda install pytorch==1.11.0 torchvision==0.12.0 torchaudio==0.11.0 cudatoolkit=11.3 -c pytorch

I wax using 0.14.0 and it was working perfictly with almost 0 issues something had changed cusing this.

It seems to error out on larger images when editing the original size. 3840x2880 3648x5472 2299x3065

I've attached the error outputs: err2.txt err1.txt

I'm on CUDA 11.1 (I've tried 11.3/11.5 too) and have also tried torch 1.9-1.10, with the same errors occurring on my RTX 3080 GPU. I'm using the Docker setup but with the torch replaced with the cuda 11.1+ version. Selecting 2k or 1080p usually results in it working properly (I believe sometimes 2k will throw an error, but then will work)

https://user-images.githubusercontent.com/71787427/163864876-0ab01a33-4d0c-4d2f-9c0e-ee69606ab2a3.mp4

The image resizes itself auto zooms out when the mouse is released.Previously it was keeping the zoom level. It is doing it with multi stroke as well.

and this is how thee old version acts:

https://user-images.githubusercontent.com/71787427/163869321-ed254ddf-7229-4c47-9b75-d2046d5645f9.mp4

I just installed yarn.

When trying to run the frontend using "yarn start" or trying to do "yarn build" on Windows 10, I get an error:

D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app>"C:\Program Files\nodejs\node_modules\corepack\shims\yarn" start i ｢wds｣: Project is running at http://192.168.56.1/ i ｢wds｣: webpack output is served from i ｢wds｣: Content not from webpack is served from D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\public i ｢wds｣: 404s will fallback to / Starting the development server...

D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\node_modules\react-scripts\scripts\start.js:19 throw err; ^

Error: error:0308010C:digital envelope routines::unsupported at new Hash (node:internal/crypto/hash:71:19) at Object.createHash (node:crypto:133:10) at module.exports (D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\node_modules\webpack\lib\util\createHash.js:135:53) at NormalModule._initBuildHash (D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\node_modules\webpack\lib\NormalModule.js:417:16) at D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\node_modules\webpack\lib\NormalModule.js:452:10 at D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\node_modules\webpack\lib\NormalModule.js:323:13 at D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\node_modules\loader-runner\lib\LoaderRunner.js:367:11 at D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\node_modules\loader-runner\lib\LoaderRunner.js:233:18 at context.callback (D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\node_modules\loader-runner\lib\LoaderRunner.js:111:13) at D:\py\lama-cleaner\lama-cleaner-main\lama-cleaner-main\lama_cleaner\app\node_modules\babel-loader\lib\index.js:59:103 { opensslErrorStack: [ 'error:03000086:digital envelope routines::initialization error' ], library: 'digital envelope routines', reason: 'unsupported', code: 'ERR_OSSL_EVP_UNSUPPORTED' }

Node.js v18.12.1

I want to modify the server code to not be a server and to run on automatically modified images with automatically generated masks, the only reason I'm trying to run the frontend is to try to figure out some default/good values for some of the config variables:

config = Config( ldm_steps=form["ldmSteps"], ldm_sampler=form["ldmSampler"], hd_strategy=form["hdStrategy"], zits_wireframe=form["zitsWireframe"], hd_strategy_crop_margin=form["hdStrategyCropMargin"], hd_strategy_crop_trigger_size=form["hdStrategyCropTrigerSize"], hd_strategy_resize_limit=form["hdStrategyResizeLimit"], prompt=form["prompt"], use_croper=form["useCroper"], croper_x=form["croperX"], croper_y=form["croperY"], croper_height=form["croperHeight"], croper_width=form["croperWidth"], sd_mask_blur=form["sdMaskBlur"], sd_strength=form["sdStrength"], sd_steps=form["sdSteps"], sd_guidance_scale=form["sdGuidanceScale"], sd_sampler=form["sdSampler"], sd_seed=form["sdSeed"], cv2_flag=form["cv2Flag"], cv2_radius=form['cv2Radius'] )

Hi:

Is it possible to use lama-cleaner to do the opposite of inpainting, that is, to keep the object selected with a mask?

I have tried substracting the inpainted image to the original, but that process leaves a lot of noise. Since the model is trained to remove structures, what is needed to train another model which leaves them (and removes the background)?

I would be very grateful for any hint on how to solve this problem.

(PD: I don't know if this is the appropiate space for this kind of question, I apologize in advance if that is not the case)

Thanks in advance!

Thanks for looking this question ! When I click the download icon , the cleaned image can be downloaded successfully. But I want get the image output url. I check in js files like '2.ee9dcc6c.chunk.js' and 'main.5cf6948e.chunk.js' and try to find the cleaned image download url, but I failed. Anybody know how can do it? Thanks again!

Hi, I am only getting a black blob where the mask is drawn instead of an image when using the SD1.5 and Paint by Example tools. I have the dreaded MSI 1660 Super GPU with 6gb VRAM. I added the --sd-disable-nsfw command line argument. Attached is the error message I get. Any solutions?

Thanks

Hello, whenever I try to paint then either the software just blackens the marked area or transforms it, but it never painted the things I asked it to. What can I do? I have an RTX 3070 Ti 8GB