Git Product home page Git Product logo

civitai's Introduction

Contributors Forks Stargazers Issues Apache License 2.0 Discord


Table of Content

About the Project

Our goal with this project is to create a platform where people can share their stable diffusion models (textual inversions, hypernetworks, aesthetic gradients, VAEs, and any other crazy stuff people do to customize their AI generations), collaborate with others to improve them, and learn from each other's work. The platform allows users to create an account, upload their models, and browse models that have been shared by others. Users can also leave comments and feedback on each other's models to facilitate collaboration and knowledge sharing.

Tech Stack

We've built this project using a combination of modern web technologies, including Next.js for the frontend, TRPC for the API, and Prisma + Postgres for the database. By leveraging these tools, we've been able to create a scalable and maintainable platform that is both user-friendly and powerful.

  • DB: Prisma + Postgres
  • API: tRPC
  • Front-end + Back-end: NextJS
  • UI Kit: Mantine
  • Storage: Cloudflare

Getting Started

To get a local copy up and running follow these simple example steps.

Prerequisites

First, make sure that you have the following installed on your machine:

  • Node.js (version 18 or later)
  • Docker (for running the database)

We recommend you have installed nvm in order to set the right node version to run this project

curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.2/install.sh | bash

Installation

  1. Clone the repository to your local machine.
  2. Run npm install in the project directory to install the necessary dependencies.
  3. Spin up required services with docker-compose up -d
    • Note: In addition to postgres and redis, this will also run maildev for email and minio for s3 storage with all necessary buckets automatically created, minio and maildev are not strictly needed but are preferred for testing and development purposes.
  4. Create your .env by making a copy of the contents from .env-example file.
    • Most default values are configured to work with the docker-compose setup, with the exception of the S3 upload key and secret. To generate those, navigate to the minio web interface at http://localhost:9000 with the default username and password minioadmin, and then navigate to the "Access Keys" tab. Click "Create Access Key" and copy the generated key and secret into the .env file.
    • Set WEBHOOK_TOKEN to a random string of your choice. This will be used to authenticate requests to the webhook endpoint.
  5. Run npm run db:migrate to run all database migrations.
  6. Run npm run db:generate to generate the prisma client.
  7. Start the development server by running npm run dev.
  8. Visit the page http://localhost:3000/api/webhooks/run-jobs?token=WEBHOOK_TOKEN&run=update-metrics to start the metrics update job (make sure to substitute WEBHOOK_TOKEN)
  9. Finally, visit http://localhost:3000 to see the website.
    • Note that account creation will run emails through maildev, which can be accessed at http://localhost:1080.
    • Also note that Cloudflare credentials are necessary in order for image uploads to work.

Important Scripts

docker-compose up -d # Spin up db, redis, maildev, and minio

npm run dev # Start the dev environment

npm run db:migrate -- --name migration-name # Create a database migration with prisma after updating the schema

npm run db:generate # Generates local prisma client

npm run db:ui # Start Prisma Studio to manage the database content

npm run build # Build the NextJS project

Contributing

Any contributions you make are greatly appreciated.

If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement". Don't forget to give the project a star! Thanks again!

  1. Fork the repository to your own GitHub account.
  2. Create a new branch for your changes.
  3. Make your changes to the code.
  4. Commit your changes and push the branch to your forked repository.
  5. Open a pull request on our repository.

Sponsors

Support this project by becoming a sponsor. Your logo will show up here with a link to your website.

License

Apache License 2.0 - Please have a look at the LICENSE for more details.

civitai's People

Contributors

bkdiehl avatar blafking avatar charleprr avatar cmoody avatar eltociear avatar eurotaku avatar flipperbw avatar garfield550 avatar justmaier avatar koenbeuk avatar lrojas94 avatar lychees avatar manuelurenah avatar peterlorre avatar rickbau5 avatar robintail avatar rockerboo avatar ssube avatar zaidhaan avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

civitai's Issues

[Bug]: GitHub sign in redirects to localhost

What happened?

Tried to sign in using GitHub, got redirected to http://localhost:3000/api/auth/callback/github?error=redirect_uri_mismatch&error_description=The+redirect_uri+MUST+match+the+registered+callback+URL+for+this+application.&error_uri=https%3A%2F%2Fdocs.github.com%2Fapps%2Fmanaging-oauth-apps%2Ftroubleshooting-authorization-request-errors%2F%23redirect-uri-mismatch&state=REDACTED

Steps to reproduce the problem

  1. Go to https://civitai.com/login
  2. Press "GitHub"

What should have happened?

A successful sign-in.

What platforms do you use to access the site?

No response

What browsers do you use to access the site?

No response

Additional information, context and logs

No response

[Bug]: Not possible to use api to download "early access" models via api

What happened?

I signed up to be a member today, and noticed my code to the API bonked with the new membership thing. Is there and API update that let's me log in as a member to download the "early access" models via the API?

Steps to reproduce the problem

trying to download the model api link yields an html file asking to be a member, I am a member but have no way of telling the api that.

What should have happened?

the early access model should be downloaded

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

No response

Additional information, context and logs

No response

[Bug]: Uploading model files to published models. Prisma update Unique constraint failed on the fields: (`modelVersionId`,`type`,`format`)

What happened?

I was adding a model file to a model version that was already created and set as published. A pruned model. But getting the following after saving the model.

Could not save model
Something went wrong: An error occurred while saving the model: Invalid prisma.model.update() invocation: Unique constraint failed on the fields: (modelVersionId,type,format)

Steps to reproduce the problem

  1. On a published model, upload a new model using +Add model under the model file on the edit model page.
  2. Wait for upload to finish and press save.

What should have happened?

New added model file is uploaded and saved to the model and update the model version with the new pruned file.

What platforms do you use to access the site?

Linux

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

I additionally tried to change the name of the model version but it did not change it.

Screenshot 2023-01-13 at 16-59-38 Civitai Share your models

[Bug]: Page element layout

What happened?

截屏2023-02-27 17 40 07

This screenshot is from https://civitai.com/models/13256/boneless-pizza. The carousel component on the left is not vertically centered, and the controller is not vertically centered either. These make the whole page look unbalanced.

Steps to reproduce the problem

https://civitai.com/models/13256/boneless-pizza

This occurs when the aspect ratio of the preview images are all greater than one.

What should have happened?

截屏2023-02-27 17 45 42

This layout may be better.

What platforms do you use to access the site?

Windows, MacOS

What browsers do you use to access the site?

Google Chrome

Additional information, context and logs

No response

[Bug]: Incredibly slow downloads

What happened?

Downloads have been incredibly slow since yesterday, despite a stable gigabit connection. Taking 12 hours+ with some models (depending on the size).

Is there some kind of 24 hour bandwidth cap? If so, I couldn’t find it documented anywhere.

Steps to reproduce the problem

  1. Go to any model
  2. Press download
  3. Observe download speeds

What should have happened?

A quick and easy download at high speeds.

What platforms do you use to access the site?

Windows, Linux, MacOS

What browsers do you use to access the site?

Google Chrome

Additional information, context and logs

If this is working as intended, thats fine, just wasn’t documented or notified anywhere I could see. :)

Notification system to know when a model has a new version

Discussed in https://github.com/orgs/civitai/discussions/53

Originally posted by JustMaier November 30, 2022
This one comes from DarkLocust on Discord.

So you can follow your favorite models and get notified for the updates

Now that we have a favorite/liked model system. It'd be great for us to add a notification system so that we can notify you when new versions of those models are available. I could also see this being used to follow specific creators so you can be notified any time they release a new model.


Here's a rough outline of how this can work:
HackMD

And a rough mockup:
Notifications mockup

[Bug]: The upload of the picture failed, the preview code found that the sub -thread started failed

What happened?

Drag Images Here or CLICK to Select Files, I have always uploaded no error information output.

Steps to reproduce the problem

Drag Images Here or CLICK to Select Files, I have always uploaded no error information output
Picture upload failed

What should have happened?

Normal should be displayed or successful

What platforms do you use to access the site?

No response

What browsers do you use to access the site?

No response

Additional information, context and logs

企业微信截图_16777604506893

[Bug]: add a new model fails "file _.safetensors could not be found" on Firefox

What happened?

Uploading a lora file always fails when clicking save. Website is unable to find the model I uploaded when using Firefox

Steps to reproduce the problem

  1. click upload a model
  2. Select LORA
  3. Create a new version
  4. Upload the LORA .safetensors file
  5. Add required fields for the LORA
  6. Click Save
  7. Get error

What should have happened?

Model uploads just fine and I can save / publish a draft

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

Error message

"Could not save model

An error occurred while saving the model: File NameOfLORA.safetensors could not be found.

Please re-upload."

Worked when I upload on Brave browser instead of Firefox.

[Bug]: db:migrate failed

What happened?

db error: ERROR: cannot alter type of a column used by a view or rule
DETAIL: rule _RETURN on view "ImageResourceHelper" depends on column "name"
   0: sql_migration_connector::validate_migrations
           with namespaces=None
             at migration-engine\connectors\sql-migration-connector\src\lib.rs:301
   1: migration_core::state::DevDiagnostic
             at migration-engine\core\src\state.rs:269

Happening after a fresh clone of main branch and following the steps described in the readme.
Step 5 is where it fails. Ive never worked with this db framework so no idea whats going wrong here.
Is there a faulty config/file on my side or something with repo?

Steps to reproduce the problem

Follow the readme steps till step 5.

What should have happened?

Succesful migration.

What platforms do you use to access the site?

No response

What browsers do you use to access the site?

No response

Additional information, context and logs

No response

[Bug]: Reporting broken

What happened?

Reporting is broken

Unable to send report
Invalid `prisma.commentReport.create()` invocation: An operation failed because it depends on one or more records that were required but not found. No 'Comment' record(s) (needed to inline the relation on 'CommentReport' record(s)) was found for a nested connect on one-to-many relation 'CommentToCommentReport'.

Steps to reproduce the problem

  1. Go to comment/model/etc (specifically image comment)
  2. Report it

What should have happened?

Report content

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

[Bug]: Epoch limited to 1000 for textual inversions

What happened?

Inputting epoch of 1478 into the epoch of a textual inversion. It reverts it to 1000.

Steps to reproduce the problem

Enter 1478 into the epoch of a textual inversion

What should have happened?

Should keep the epoch of 1478.

What platforms do you use to access the site?

Linux

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

There may be a sane limit for this field, but for my case it's too low. If you do batch gradient descent (of the whole training set) each step would be a full epoch. This contributed to the epoch (1478) being large at 4,476 steps, with steps per epoch starting at 1 then to 3 steps per epoch later.

[Bug]: "Please verify your username and agree to the terms of service"

What happened?

I get that message at every page that I visit. The whole form that asks me to enter my nickname and agree to the TOS.

I registered with my google account if that can help.

Steps to reproduce the problem

Register with google.
Confirm nickname
Refresh page

What should have happened?

not showing this at every refresh of the page

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

[Bug]: All models filtered out when browsing while logged in

What happened?

Yesterday, I noticed that all models are filtered out when I am browsing while logged in. This is true on the main page, as well as my liked models page and any user pages. The exception is the hidden models page, where I can still see the (faded) list of hidden models.

This issue affects me when logged in on multiple browsers, but not when logged out. If I have the model page link, I can still go to the page and interact with it, only search is impaired.

Steps to reproduce the problem

  1. I logged in, and went to the main page, removing all filters. I get the no results found icon.
  2. I log out, and the search works fine.
  3. I made a burner account to see if it was a system/browser issue, but I the new account works fine.
  4. I attempted other troubleshooting steps as well (clearing cookies, disabling browser extensions, etc..)

What should have happened?

I just want to restore normal account behavior. It's strange that it seems to be account specific. My guess is that I triggered a bug when switching a few user settings in rapid combination that corrupted some user setting. I'm okay if you need to reset my account: I can rebuild my likes/hidden settings, but just want to keep the supporter badge.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox, Google Chrome

Additional information, context and logs

No response

[Bug]: Editing a comment clears out mention linking

What happened?

When editing a comment, the link to the mentioned user is removed.

Steps to reproduce the problem

  1. Make a comment and mention someone
  2. Edit comment
  3. Save edit comment

What should have happened?

Link should remain after editing

What platforms do you use to access the site?

Linux

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

Just noticed it and know it's a new feature but wanted to mention it.

[Bug]: Values on frontpage do not align with values on model page

What happened?

Example: ChilloutMix (model page downloads: 50000+, frontpage: 13000)

Also applies to review count and likes, e. g. the Ulzzang-6500 embedding has three reviews but the counter on the frontpage says 2. Likes are the same story, around half the number is shown.

image

Doesn't seem to be a local caching issue, it shows the same when ignoring it.

Steps to reproduce the problem

  1. Go to frontpage, check numbers (sorted by likes/day)
  2. Go to the model page and compare

What should have happened?

It should show the accurate values when reloading the page, or atleast do a regular update so that the gaps aren't as big.

What platforms do you use to access the site?

Linux

What browsers do you use to access the site?

Google Chrome

Additional information, context and logs

No response

When saving a review images save in the wrong order

What happened?

I made a review and put my images in a specific order and they were in a different order after I saved

Steps to reproduce the problem

  1. Make a review
  2. Save
  3. Check review image order

What should have happened?

The order should stay what I made it

What platforms do you use to access the site?

Windows, iOS

What browsers do you use to access the site?

Apple Safari, Microsoft Edge

Additional information, context and logs

It works fine when editing, just not with the initial save

[Bug]: stripping username underscore characters - malforming user profile paths

What happened?

If you click a user in the 'Uploader' field on a model page that has a underscore in their name, it will strip the underscore in the redirected path

Steps to reproduce the problem

  1. Click user in the 'Uploader' field on a model page that has a underscore in their name (eg, cool_person)
  2. cool_person will resolve to https://civitai.com/user/coolperson which shows a "No Results Found" page.

--

If you explicitly search the user in the search bar (@cool_person) it will work correctly.

What should have happened?

Dunno, are you intentionally stripping underscores for some reason?

What platforms do you use to access the site?

No response

What browsers do you use to access the site?

No response

Additional information, context and logs

No response

[Bug]: Inpainting models aren't working because filename is renamed without -inpainting suffix

What happened?

Inpainting models aren't working right because the filename is changed when downloaded (or uploaded), which removes the -inpainting suffix at the end of the filename. The filename has to end with -inpainting.ckpt in order to work as an inpainting model in Auto1111. There is a PR for Auto1111 that changes this behavior that will recognize inpainting models without the suffix, but I don't think it has been implemented yet. Does Civitai rename the files when they are uploaded or downloaded?

Steps to reproduce the problem

  1. Download an inpainting model
  2. The model won't load in Auto1111 and throws errors
  3. The model filename has to be renamed with -inpainting at the end, before the .ckpt extension, to get it to work

What should have happened?

The filename has to keep the -inpainting suffix at the end in order for Auto1111 to recognize the model as an inpainting model.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Microsoft Edge

Additional information, context and logs

No response

[Bug]: Paged Image Galleries Broken

What happened?

After the most recent outage, it seems like the ability to page through a gallery of images from either the model page itself or from one of its reviews is gone. Only one image can be viewed at a time, to view another one must exit that image and click another. There is no "page indicator" at the bottom, left/right arrows don't work,

Steps to reproduce the problem

-Click on an image from any model or model review with multiple images.
-Attempt to scroll left or right through the image gallery

What should have happened?

When viewing an image fullscreen, it should have the page indicator dots and it should show clickable left/right arrows as well as respond to left/right arrow key inputs.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

not found: Can't resolve 'use-sync-external-store/shim'

What happened?

odule not found: Can't resolve 'use-sync-external-store/shim'

Import trace for requested module:
./node_modules/_use-sync-external-store@1.2.0@use-sync-external-store/shim/with-selector.js
./node_modules/_zustand@4.3.3@zustand/esm/index.mjs
./src/store/ClientHistoryStore.tsx
./src/pages/_app.tsx

Steps to reproduce the problem

  1. Go to ....
  2. Press ....
  3. ...

What should have happened?

npm run dev error : not found: Can't resolve 'use-sync-external-store/shim'

What platforms do you use to access the site?

No response

What browsers do you use to access the site?

No response

Additional information, context and logs

No response

[Bug]: Show full model description

What happened?

I don't think model description should be automatically hidden. No one is reading the descriptions of my models, including example prompts for the very questions they routinely ask.

Description, representation, usage restrictions, etc, in any case, should be right there on the main page of a model, not hidden away at any point without the user manually doing it -- especially when we can't actually select licenses to store and distribute with models (which is technically required to share 1.4/1.5/2.0/2.1 models and derivatives, and isn't being honored)

Steps to reproduce the problem

Visit any model page

What should have happened?

All model authors information presented to the user

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

[Bug]: Page over burdened

What happened?

Suddenly Civitai is overburdened with a lot of jank time, and slow loading. Additionally, scrolling is no longer smooth. It's very jerky/stuttery.

Checking online, the site gets a performance score of C or D (and fluctuating, whether server performance or scripting performance), which ain't good.

One of the biggest factors is the image delivery. It's using cookies, which is unnecessary traffic as the server will ignore them. Need a proper CDN or a subdomain distributor that is cookie-free.

Setup:
Core i9 13900KF (Up to 5.8Ghz Boost)
32GB 5600MHz DDR5
SN570 NVMe 1TB (OS Drive)

Steps to reproduce the problem

  1. Visit the site

What should have happened?

Site should be much more responsive and smooth.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

Thins were smooth yesterday, despite all the downtime. 🤷

[Bug]: Model information jumps location in prompt information modal

What happened?

The location of the model used in a generation changes location between images seemingly by random between images.

I guess differences in template files or something?

Screenshot_8
Screenshot_9

Steps to reproduce the problem

  1. Go to model and view image prompt details
  2. Certain images, the model data location changes

What should have happened?

A consistent layout design for referencing data between images.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

Just a minor design issue

Can't enable mature content

What happened?

When I'm going into the settings I can enable mature content. But once the option is enable it will not save it, so each time I reload or close the page I will always get a home page without nsfw.

Steps to reproduce the problem

  1. Go to settings
  2. Enable mature content
  3. Reload the page

What should have happened?

It must save what I've select

What platforms do you use to access the site?

Windows, Linux, MacOS, iOS, Android, Other/Cloud

What browsers do you use to access the site?

Mozilla Firefox, Brave, Apple Safari

Additional information, context and logs

No response

[Bug]: Error 404 on all model pages

What happened?

Screenshot 2022-12-29 at 02-54-10 Page Not Found

Steps to reproduce the problem

  1. Go to home page
  2. Click on Any model
  3. Get 404 Error

What should have happened?

Show Model Details

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

[Bug]: Animated svg heart on top left takes a lot of cpu time (firefox, linux)

What happened?

My laptop likes to spin its fan up when browsing civitai, really seems to be the heaviest page I browse regularly.

Narrowed down the issue to the animated svg heart on top left, since doing simple css style tweak seems to bring cpu use down to regular levels:

svg, svg path {
        animation: none !important;
}

Steps to reproduce the problem

  1. Open any civitai.com page
  2. Observe cpu usage (~60-120% in my case)

What should have happened?

Don't eat so much cpu plz.

What platforms do you use to access the site?

Linux

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

These might be related firefox bugs:
https://bugzilla.mozilla.org/show_bug.cgi?id=1812394
https://bugzilla.mozilla.org/show_bug.cgi?id=1409486

[Bug]: LoRA filenames are not preserved on download links

What happened?

I've noticed that on a LoRA I've made the downloaded files do not have the same name as the one's that I've uploaded. For example, I have two versions of a LoRA that I made for compatibility purposes, and AnythingV3 version and AOM3 version. The AnythingV3 was uploaded with the name "zoroark_AT3.safetensors" but the downloaded file has the title "zoroarkLora_zoroarkLoraAnythingV3.safetensors." Even stranger is that the "zoroark_AOM3.safetensors" version is only named "zoroarkLora_zoroarkLora.safetensors" with no mention of AOM3 in the filename. This is bad since there's now no way to tell what the second file is based on without context (or reading the metadata). I can only confirm that this occurs on LoRA's, I do not have any models to upload.

I think there should either be an option to use the original file name or enter a new one for downloaded files, and the published name shouldn't be prepended to the downloaded files since that's redundant.

Here's the upload in question: https://civitai.com/models/25485

Steps to reproduce the problem

  1. Create a LoRA.
  2. Upload LoRA file(s).
  3. Publish.
  4. Download LoRA from published page and compare filenames.

What should have happened?

Downloaded files should have had the original filenames from the uploaded files.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Google Chrome

Additional information, context and logs

No response

[Bug]: Copy Generation data button gets pushed to the bottom of the screen due to comments

What happened?

When a user adds a comment to the Full preview mode after clicking an image on the model, IF there is a comment on the picture, the Copy Generation Data button vanishes to the bottom.
image

Steps to reproduce the problem

  1. Go to ....
  2. Press ....
  3. ...

What should have happened?

The Copy Generation data button should not be pushed to the bottom

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Google Chrome

Additional information, context and logs

No response

[Bug]: Download Total Random

What happened?

The total downloads are seemingly random, and do not reflect the project.

For example, I have 743 total downloads, but the total only reads a random 293

https://civitai.com/models/20793/was-node-suites-comfyui

Steps to reproduce the problem

  1. Go to model with multiple version
  2. Count the total downloads of all versions
  3. Refer to the total downloads at the top of the page. It doesn't relate to any one version, or the total.

What should have happened?

Simply count up all versions downloads. No funny business.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

Issues with notifications creats a bad user experience on the site

Discussed in https://github.com/orgs/civitai/discussions/143

Originally posted by a-l-e-x-d-s-9 January 12, 2023

  1. People are adding responses to comments on my model, and I don't get any notification about it. I would like to have an option for notifications regarding all comments.
  2. When people use a username or @ + username, it doesn't send me a notification.
  3. When I leave a response to a comment, I don't get notifications about new comments on this thread.
    All these notification problems create a lot of broken conversations, miscommunication, and misunderstandings. That leads to a generally bad experience on the site.
    It would be great if you can address these issues, and fix them.
    Thanks!

[Bug]: Wrong data in search result

What happened?

The data of some objects are not correctly shown in search result. As shown in the following screenshots:

incorrect data in search result
correct data

This issue is not only about this object, but also can be found on other objects.

Steps to reproduce the problem

  1. Search "zyd232's Chinese Girl LORA" on civitai.com
  2. Check the data about downloads, likes and comments on the card.
  3. Click the card to enter the detail page. Check the data about downloads, likes and comments in the detail page.
  4. You can find the data in step 2 and step 3 are not the same.

What should have happened?

The data in search result should be the same as the data in detail page.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Microsoft Edge

Additional information, context and logs

No response

[Bug]: Discussion section is disappearing on scroll after expanding a long model description

What happened?

The discussions and reviews is disappearing on scroll after expanding a long model description, it seems like there's a problem with Intersection Observer. For more details, watch video below

2303270946171581.mp4

Steps to reproduce the problem

  1. Go to any model page that have a long model description (like this one)
  2. Press Show More in model description
  3. Scroll down to discussion/review section

What should have happened?

Discussion/review section should be visible all the time regardless on how long the model description is

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Google Chrome

Additional information, context and logs

No response

[Bug]: OpenGraph description preview has no spaces for new lines, results appear aweful.

What happened?

The OpenGraph data for models and such is pretty bad. It doesn't have spaces for new lines, so they are back to back with a often different paragraph.

Screenshot_2

Perhaps new lines should be prefixed with | or just a space, though a space may confuse people when it is a new paragraph/subject.

Steps to reproduce the problem

  1. Post content from CivitAI anywhere with OpenGraph support, with a formatted description area.

What should have happened?

N/A

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

[Bug]: API by-hash is returning 404's for valid hashes.

What happened?

The API is returning a 404 error for valid hashes. Tried with blake3 and sha256 copied from model pages.

Steps to reproduce the problem

For example, using the blake3 hash copied from Dreamshaper's model page for the 3.31 bake vae version:

https://civitai.com/models/4384/dreamshaper

Which is: 30BE3842DAE0A525F6BF47A1E1B47098AEA72500553B77541885F497C1D03723

$ curl --head -X GET https://civitai.com/api/v1/models-versions/by-hash/30BE3842DAE0A525F6BF47A1E1B47098AEA72500553B77541885F497C1D03723
HTTP/2 404
date: Sun, 12 Mar 2023 03:06:30 GMT
content-type: text/html; charset=utf-8
x-powered-by: Next.js
cache-control: private, no-cache, no-store, max-age=0, must-revalidate
vary: Accept-Encoding
strict-transport-security: max-age=15724800; includeSubDomains
cf-cache-status: DYNAMIC
server-timing: cf-q-config;dur=6.9999950937927e-06
report-to: {"endpoints":[{"url":"https:\/\/a.nel.cloudflare.com\/report\/v3?s=YBenzlVU0DdR4pPlK%2Bvxn%2BeDxCuVxZOZei%2B1Z%2FlpSrF9o4sLoS7hBHr8gWIIL1f0wbV1b5qrbMGImqG49hE2poe%2BU4Tgruvyw6EwIx05iXp0Fq7KDbNnLcA6dOmAXscpD7LtjA4HrqsLzg%3D%3D"}],"group":"cf-nel","max_age":604800}
nel: {"success_fraction":0,"report_to":"cf-nel","max_age":604800}
server: cloudflare
cf-ray: 7a68d2552bb89836-SJC
alt-svc: h3=":443"; ma=86400, h3-29=":443"; ma=86400

What should have happened?

A valid model JSON response should be sent.

What platforms do you use to access the site?

MacOS

What browsers do you use to access the site?

Apple Safari

Additional information, context and logs

No response

Not remembering color preferences

What happened?

After choosing dark mode, a day or so later when I came back, it was light mode again

Steps to reproduce the problem

  1. Switch to dark mode
  2. Wait a day
  3. Come back and it’s light again…

What should have happened?

It should remember my setting for ever. Ideally it would be for my entire account on all of my devices.

What platforms do you use to access the site?

Windows, iOS

What browsers do you use to access the site?

Apple Safari, Microsoft Edge

Additional information, context and logs

No response

[Bug]: Landscape format images have blurry thumbnails

What happened?

I uploaded images for model reviews and all 16:9 ones get blurry thumbnails.
From looking at other images it seems the wider the image, the blurrier the thumbnail.

Steps to reproduce the problem

Check the reviews on this model's page for example : https://civitai.com/models/7241/mix-pro-v3
Look closely, some thumbnails are sharp, others are a bit or quite blurry.

What should have happened?

--

What platforms do you use to access the site?

No response

What browsers do you use to access the site?

No response

Additional information, context and logs

No response

[Bug]: Adult Models Vanished Across the entire Site

What happened?

There are no adult models visible on Civitai, even when you have adult models active. On my model list, if adult models are active, none of my models show. If I disable it, only my safe models show. But with it active, everything of mine is hidden with:

No results found
Try adjusting your search or filters to find what you're looking for

Steps to reproduce the problem

  1. Browser civitai or
  2. Visit your profile and personal models (if you have NSFW models)
  3. Notice you have no results found.
  4. Disable NSFW and only your safe models show, but no longer receive no results found error.

What should have happened?

I can see adult models site-wide with my appropriate settings.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

[Bug]: Deployment problem

What happened?

After the deployment process is completed, visit Civitai, and try to report an error The table public.User DOES NOT Exist in the Current DataBase through the mailbox. There are other documents to refer to the configuration.env configuration file

Steps to reproduce the problem

The mail configuration is as follows
image
But I can't receive a registered email, and the back -end log to report an error message:
The table public.User does not exist in the current database.

What should have happened?

Normally, it should be able to successfully register the email normally

What platforms do you use to access the site?

No response

What browsers do you use to access the site?

No response

Additional information, context and logs

No response

[Bug]: website is down

What happened?

cloudflare reports the website is down.

Steps to reproduce the problem

according to the website

What should have happened?

the website should work.

What platforms do you use to access the site?

No response

What browsers do you use to access the site?

No response

Additional information, context and logs

No response

[Bug]: There are many models that are not showing up in the api response when sorting by new

What happened?

If I send a request for the newest 100 uploads (https://civitai.com/api/v1/models?limit=100&sort=Newest) I do not see several uploads that should be in the list.

For example at the time of this writing none of the models from:
https://civitai.com/user/bloodsplash09
https://civitai.com/user/Puffin

show up in the api response

Steps to reproduce the problem

go here: (https://civitai.com/api/v1/models?limit=100&sort=Newest

you should see models from:
https://civitai.com/user/bloodsplash09
https://civitai.com/user/Puffin

but you don't see any of these models in the api response, even though they are showing on the civitai site.

What should have happened?

new models should be showing up in the api call for the newest 100 uploads, but there are several missing and they look to be missing as a function of the person uploading them, that is that the missing models from the api call are not random they are from certain uploaders

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Google Chrome

Additional information, context and logs

I use the api a lot, and have never seen this happen before.

[Bug]: API: Models Type parameter doesn't work

What happened?

Calling the /models endpoint with the type parameter doesn't seem to make a difference regardless of the enum type chosen.

Steps to reproduce the problem

The response from the following url will return 3 Checkpoints:

https://civitai.com/api/v1/models?limit=3&type=TextualInversion

That's a link taken from the wiki and it also exhibits this problem in the documented response example.

This seems to be happening with the other options as well.

What should have happened?

The Type enum values should return their corresponding model types.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

[Bug]: 'prisma invocation: Unique constraint failed' when verifying email after registering with Reddit

What happened?

When signing up to Civitai using a Reddit account, an error is thrown if the email address provided for Civitai registration is the same as the email that was used to register for the Reddit account.

Steps to reproduce the problem

-Click burger menu
-Sign in/sign up
-Register with Reddit account
-When validating username/email use same email as for Reddit account

What should have happened?

Username/Email validation form allows you to use the same email address that you used to register your SSO account

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Mozilla Firefox

Additional information, context and logs

No response

[Bug]: Keep getting logged out

What happened?

I keep getting logged out when I open a new tab or refresh and have to go to https://civitai.com/models/create to snap it back into thinking im logged in. Pressing sign in while "logged out" does nothing. The website functions as if I'm logged out, blurring nsfw models and showing a follow button on my profile.

Steps to reproduce the problem

  1. Refresh/Open new tab
  2. Ghost logged out on all tabs
  3. Go to https://civitai.com/models/create to "log back in"

What should have happened?

It remains logged in.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Google Chrome

Additional information, context and logs

[email protected]

[Bug]: Connecting to the API isn't working anymore

What happened?

My code was running as usual, but around 11:40 today it stopped working. Upon trying to get a response back form Civitai I am getting this error: schannel: next InitializeSecurityContext failed:

Steps to reproduce the problem

(https://civitai.com/api/v1/models?limit=100&sort=Newest) trying to read from this link for example yields the above error.

What should have happened?

I should have gotten a response from the server.

What platforms do you use to access the site?

Windows

What browsers do you use to access the site?

Google Chrome

Additional information, context and logs

Everything had been working fine for a while, but now I am getting this error.

[Bug]: Looks like some of the images don't upload correct

What happened?

I've occasionally seen some images that failed to upload or something:
image

Steps to reproduce the problem

  1. Go to this page

What should have happened?

I guess some how we need to make sure that the image uploads were successful so that we don't save temp blob urls to the db.

What platforms do you use to access the site?

No response

What browsers do you use to access the site?

No response

Additional information, context and logs

No response

Change models url with slug for better SEO

Discussed in https://github.com/orgs/civitai/discussions/29

Originally posted by manuelurenah November 16, 2022
Right now the model url structure is as follow /models/{id} which is not great for SEO since it will generate the following /models/1234. To improve this, it'd be best to slugify the model title and set that as the url /models/{slug} (e.g.: /model/arcane-diffusion).

Also, we should strip any html tags from the meta description

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo 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.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.