Just get back to my pc after 3 weeks, do some update and the text-generation UI stops working, I am using oobabooga. Can someone take a look and tell how to fix this?
AttributeError: 'LlamaLikeModel' object has no attribute 'layers'
Output generated in 0.49 seconds (0.00 tokens/s, 0 tokens, context 75, seed 514002040)
Are there any other text-generation UI, hopefully lighter than it?
Edit: it seems possible to install a version in the past(thanks to github) now I just have to find the one that worked for me.
Just get back to my pc after 3 weeks, do some update and the text-generation UI stops working, I am using oobabooga. Can someone take a look and tell how to fix this?
Are there any other text-generation UI, hopefully lighter than it?
Edit: it seems possible to install a version in the past(thanks to github) now I just have to find the one that worked for me.
This thread is about Stable Diffusion image generation, not LLM oobabooga etc.
The short answer to your actual question:
Oobabooga and the underlying libraries are in a constant state of churn and breakage. seems like something updated that needs another package updated.
1st thing to try would be the update_webui.bat (or whatever it is called now)
2nd would be move your downloaded models out to a temp folder (and anything else you want to keep like character defs, system prompts). Then delete the whole thing and reinstall fresh.
been making depth map poses from images found on net. is there a way to get rid of the background areas and just keep the figure? cause when u use the pose it renders whatever was in the map too. i just want the model depth pose created
example:
modine2021 The depth map is just an image. You could download it, use an image editor to remove the background (replace it with pure black) and then re-upload as the controlnet depth mask.
The details of downloading/saving and re-uploading depend on the tool you are using. No guarantee that online AI generators allow that functionality. Local install of comfyui etc will definitely allow the setup though.
note that the AI will fill the black background parts with whatever is in your prompt, but is unlikely to do anything "close" there because of the black -> far away meaning.
modine2021 The depth map is just an image. You could download it, use an image editor to remove the background (replace it with pure black) and then re-upload as the controlnet depth mask.
The details of downloading/saving and re-uploading depend on the tool you are using. No guarantee that online AI generators allow that functionality. Local install of comfyui etc will definitely allow the setup though.
note that the AI will fill the black background parts with whatever is in your prompt, but is unlikely to do anything "close" there because of the black -> far away meaning.
I've not used that plugin before, I assumed you were using ControlNet Depth, which has the functionality I mentioned to upload your own source image for depth maps.
I've not used that plugin before, I assumed you were using ControlNet Depth, which has the functionality I mentioned to upload your own source image for depth maps.
This is primarily for deadshots2842 who requested it, but may be of interest to anyone who wants to replicate the artist SunsetRiders7's work (as found in Something Unlimited).
I've included the prompts etc. separately as deadshots2842 is using ComfyUI and I'm not sure how the PNGInfo of an image generated in A1111 would 'port across'.
Main Checkpoint I use (for SD1.5 - I can't get XL to work due to technical issues) is 'Anything 4.5 (VAE-Swapped)' which no longer seems to be available. Apparently that's the Checkpoint the main LoRA I use was trained on, so I've uploaded it to
You must be registered to see the links
- N.B. it's 3.9GB! Short hash should be a504b5b137
Please exercise appropriate caution downloading a file that's been uploaded by some random stranger on the Internet. I know it's safe, but then I would say that, wouldn't I?
However,any comics / cartoon / anime-focused Checkpoint is worth a try, they can give great results. Anything v3.0 or v5.0 especially.
The main LoRA I use is
You must be registered to see the links
, created some time ago now by ablackcatman
There are also
You must be registered to see the links
which, although they're not as consistent and can produce 'Anime-like' results, they can come in handy if weird stuff is happening with the first one.
There's a
You must be registered to see the links
for PonyXL, but I have no 1st hand experience with it.
As an example image, here's one of the waitresses from the Glamor Slam who decided to get some of that sweet, sweet tips money:
If you run into problems converting the above info from A1111 to ComfyUI, you'll need to ask one of the resident ComfyUI gurus, I haven't even tried it yet - been too busy having heart surgery recently!
This is primarily for deadshots2842 who requested it, but may be of interest to anyone who wants to replicate the artist SunsetRiders7's work (as found in Something Unlimited).
I've included the prompts etc. separately as deadshots2842 is using ComfyUI and I'm not sure how the PNGInfo of an image generated in A1111 would 'port across'.
Main Checkpoint I use (for SD1.5 - I can't get XL to work due to technical issues) is 'Anything 4.5 (VAE-Swapped)' which no longer seems to be available. Apparently that's the Checkpoint the main LoRA I use was trained on, so I've uploaded it to
You must be registered to see the links
- N.B. it's 3.9GB! Short hash should be a504b5b137
Please exercise appropriate caution downloading a file that's been uploaded by some random stranger on the Internet. I know it's safe, but then I would say that, wouldn't I?
However,any comics / cartoon / anime-focused Checkpoint is worth a try, they can give great results. Anything v3.0 or v5.0 especially.
The main LoRA I use is
You must be registered to see the links
, created some time ago now by ablackcatman
There are also
You must be registered to see the links
which, although they're not as consistent and can produce 'Anime-like' results, they can come in handy if weird stuff is happening with the first one.
There's a
You must be registered to see the links
for PonyXL, but I have no 1st hand experience with it.
As an example image, here's one of the waitresses from the Glamor Slam who decided to get some of that sweet, sweet tips money:
If you run into problems converting the above info from A1111 to ComfyUI, you'll need to ask one of the resident ComfyUI gurus, I haven't even tried it yet - been too busy having heart surgery recently!
This is primarily for deadshots2842 who requested it, but may be of interest to anyone who wants to replicate the artist SunsetRiders7's work (as found in Something Unlimited).
I've included the prompts etc. separately as deadshots2842 is using ComfyUI and I'm not sure how the PNGInfo of an image generated in A1111 would 'port across'.
Main Checkpoint I use (for SD1.5 - I can't get XL to work due to technical issues) is 'Anything 4.5 (VAE-Swapped)' which no longer seems to be available. Apparently that's the Checkpoint the main LoRA I use was trained on, so I've uploaded it to
You must be registered to see the links
- N.B. it's 3.9GB! Short hash should be a504b5b137
Please exercise appropriate caution downloading a file that's been uploaded by some random stranger on the Internet. I know it's safe, but then I would say that, wouldn't I?
However,any comics / cartoon / anime-focused Checkpoint is worth a try, they can give great results. Anything v3.0 or v5.0 especially.
The main LoRA I use is
You must be registered to see the links
, created some time ago now by ablackcatman
There are also
You must be registered to see the links
which, although they're not as consistent and can produce 'Anime-like' results, they can come in handy if weird stuff is happening with the first one.
There's a
You must be registered to see the links
for PonyXL, but I have no 1st hand experience with it.
As an example image, here's one of the waitresses from the Glamor Slam who decided to get some of that sweet, sweet tips money:
If you run into problems converting the above info from A1111 to ComfyUI, you'll need to ask one of the resident ComfyUI gurus, I haven't even tried it yet - been too busy having heart surgery recently!
Comfy is a lot better if you have a gfx card with 8-16gig ram, it only didn't work for me because my 4090 was quick enough through auto1111 and I loath the flowchart method of creating with comfy
Comfy is a lot better if you have a gfx card with 8-16gig ram, it only didn't work for me because my 4090 was quick enough through auto1111 and I loath the flowchart method of creating with comfy
I don't have 4090 but I got 3060 12 gb I'll try to install auto. I heard someone saying that comfy can also give less results compared to auto is that true?
I don't have 4090 but I got 3060 12 gb I'll try to install auto. I heard someone saying that comfy can also give less results compared to auto is that true?
ComfyUI and Automatic1111 are different user interfaces to the same underlying diffusion model processing.
Automatic1111 is the OG public open way for non-experts to run stable diffusion models at home. The first wave of community extensions (sd1.5 controlnet and the like) were all built around it and broke new ground. It is no longer the only game in town, and the pace of development seems to have slowed a bit. Some of the energy has been siphoned off into making commercialized hosted versions of webui.
ComfyUI can potentially have more flexibility in how you can arrange inputs and operations on the pipeline. But given how hard it is to understand what is going on anyway, few people seem to have a good grasp on how to use it effectively, and thus just imitate other people's working flows. cargo cult stuff. The flow based mechanism is similar to the game development patterns of unreal engine or blender's 3d shaders, so some people are familiar with that style of "coding/configuration". I think it's a bit of the mental divide though - flow style appeals to some people, others like the explicitness of a traditional UI.
Two ways of getting to know how other people's prompts are put together - there are other ways but this is a good start:
1- civitai : check the examples on models or in the comments section - each image on there has the metadata available via the "i" icon
2- some images posted in the various image subforums of this site (like the Summertime saga fan art thread: https://f95zone.to/threads/summertime-saga-fan-art.4744/page-478 ) include the metadata that will show the prompt used (and all other settings for the render). If you have Automatic1111 installed, use the pnginfo tab to check the image. When downloading, be careful to click through to get the actual uploaded image, not the resized/compressed/CDN served image thumbnail that is in the post.
As a hint: sometimes ai image creators use LORAs that are specific to certain image contents (like blowjob etc) to encourage the diffusion to end up with certain content. This will be listed in the meta data separately to the prompt.
