Adding model from HugginFace failed | Invoke | Page 1

lone verge May 12, 2024, 2:56 PM

#

Hi, never tried this before so it might be a beginner error.
I wanted to try this https://huggingface.co/prs-eth/marigold-v1-0 so I copied the model name prs-eth/marigold-v1-0 to the "HugginFace Repo ID" field of Invoke and clicked "Install Repo". It was added to the install queue, progressed to 100% then displayed "FAILED", terminal output:
[ModelInstallService]::INFO --> Model download complete: prs-eth/marigold-v1-0 [ModelInstallService]::INFO --> Model install started: prs-eth/marigold-v1-0 [ModelInstallService]::ERROR --> Model install error: prs-eth/marigold-v1-0 InvalidModelConfigException: Unable to determine model type for /home/timker/InvokeAI/models/tmpinstall_f5wdt_ax/marigold-v1-0; class MarigoldPipeline is not one of the supported classes [StableDiffusionPipeline, StableDiffusionInpaintPipeline, StableDiffusionXLPipeline, StableDiffusionXLImg2ImgPipeline, StableDiffusionXLInpaintPipeline, LatentConsistencyModelPipeline, AutoencoderKL, AutoencoderTiny, ControlNetModel, CLIPVisionModelWithProjection, T2IAdapter]
Is this not how it works?

prs-eth/marigold-v1-0 · Hugging Face

rain lionBOT May 12, 2024, 2:56 PM

#

Hi @lone verge. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

#

It seems like the model you're trying to install is not compatible with the InvokeAI platform. The error message indicates that the model class MarigoldPipeline is not one of the supported classes. InvokeAI currently supports a specific set of model classes, including StableDiffusionPipeline, StableDiffusionInpaintPipeline, StableDiffusionXLPipeline, StableDiffusionXLImg2ImgPipeline, StableDiffusionXLInpaintPipeline, LatentConsistencyModelPipeline, AutoencoderKL, AutoencoderTiny, ControlNetModel, CLIPVisionModelWithProjection, T2IAdapter.

The model you're trying to install, prs-eth/marigold-v1-0, might not be compatible with these classes. It's important to note that not all models available on HuggingFace can be used with InvokeAI.

You can try installing other models that are compatible with InvokeAI. For instance, InvokeAI provides access to a series of ControlNet models that provide different effects or styles in your generated images. You can install these models using the InvokeAI model installer application. You can also use the "Add Model" function of the model manager to enter the HuggingFace Repo ID of the ControlNet[^1^].

If you're still facing issues, it might be a bug similar to the one reported on GitHub[^4^][^7^]. In such cases, it's recommended to report the issue to the InvokeAI team for further assistance.

^1^]: [InvokeAI ControlNet Installation (https://invoke-ai.github.io/InvokeAI/features/CONTROLNET#installation)
^4^]: [Github Issue: Model installation doesn't proceed after download finishes (https://github.com/invoke-ai/InvokeAI/issues/5807)
^7^]: [Github Issue: can't install invoke AI (https://github.com/invoke-ai/InvokeAI/issues/5009)

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

#

@latent shuttle kapa.ai is still learning and improving, please let us know how it did by reacting below

lone verge May 12, 2024, 2:58 PM

#

@rain lion How to identify compatible models on HugginFace?

rain lionBOT May 12, 2024, 2:58 PM

#

Let me check the knowledge base for you. One moment! :books:

#

Identifying compatible models on HuggingFace for InvokeAI involves understanding the types of models that InvokeAI supports. Currently, InvokeAI supports the following types of models:

Stable Diffusion XL (SDXL)
Stable Diffusion 1.5 (SD1.5)
ControlNet models for both SD1.5 and SDXL[^8^]

InvokeAI prefers the diffusers format for models. These are distributed as directories containing multiple subfolders, each of which contains a different aspect of the model[^6^].

When you're browsing models on HuggingFace, you should look for models that are in the diffusers format and are of the types SDXL, SD1.5, or ControlNet. The ID of these models is in the format "author/repoName".

Once you've identified a model, you can install it using the InvokeAI model installer application or the "Add Model" function of the model manager[^1^].

However, please note that not all models available on HuggingFace can be used with InvokeAI. The model class must be one of the supported classes by InvokeAI. If you're unsure about a model's compatibility, it's recommended to refer to the model's documentation or contact the model's author for more information.

^1^]: [InvokeAI ControlNet Installation (https://invoke-ai.github.io/InvokeAI/features/CONTROLNET#installation)
^6^]: [InvokeAI Checkpoint and Diffusers Models (https://invoke-ai.github.io/InvokeAI/installation/050_INSTALLING_MODELS#checkpoint-and-diffusers-models)
^8^]: [InvokeAI Supported Models (https://support.invoke.ai/support/solutions/articles/151000170961-supported-models)

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

#

@latent shuttle kapa.ai is still learning and improving, please let us know how it did by reacting below

lone verge May 12, 2024, 3:12 PM

#

@rain lion how to filter hugginface for compatible models

rain lionBOT May 12, 2024, 3:12 PM

#

Let me check the knowledge base for you. One moment! :books:

#

HuggingFace doesn't provide a direct way to filter models based on compatibility with specific platforms like InvokeAI. However, you can use the following steps as a guide to find models that are likely to be compatible:

Visit the HuggingFace Models (https://huggingface.co/models) page.
Use the search bar to look for specific types of models that InvokeAI supports, such as "Stable Diffusion", "ControlNet", or "diffusers".
Click on a model to view its details. Check the model's documentation and files to see if it matches the types of models that InvokeAI supports. For instance, InvokeAI prefers the diffusers format for models. These are distributed as directories containing multiple subfolders, each of which contains a different aspect of the model[^6^].
Once you've identified a model, you can install it using the InvokeAI model installer application or the "Add Model" function of the model manager[^1^].

Remember, not all models available on HuggingFace can be used with InvokeAI. The model class must be one of the supported classes by InvokeAI. If you're unsure about a model's compatibility, it's recommended to refer to the model's documentation or contact the model's author for more information.

^1^]: [InvokeAI ControlNet Installation (https://invoke-ai.github.io/InvokeAI/features/CONTROLNET#installation)
^6^]: [InvokeAI Checkpoint and Diffusers Models (https://invoke-ai.github.io/InvokeAI/installation/050_INSTALLING_MODELS#checkpoint-and-diffusers-models)

#Adding model from HugginFace failed

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

^1^]: [InvokeAI ControlNet Installation (https://invoke-ai.github.io/InvokeAI/features/CONTROLNET#installation) ^6^]: [InvokeAI Checkpoint and Diffusers Models (https://invoke-ai.github.io/InvokeAI/installation/050_INSTALLING_MODELS#checkpoint-and-diffusers-models)

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

^1^]: [InvokeAI ControlNet Installation (https://invoke-ai.github.io/InvokeAI/features/CONTROLNET#installation)
^6^]: [InvokeAI Checkpoint and Diffusers Models (https://invoke-ai.github.io/InvokeAI/installation/050_INSTALLING_MODELS#checkpoint-and-diffusers-models)