Skip to content

[bug]: IP Adapter always triggers re-install (and fails) #7589

Open
@Tobias-Reich

Description

Is there an existing issue for this problem?

  • I have searched the existing issues

Operating system

Windows

GPU vendor

Nvidia (CUDA)

GPU model

RTX4080

GPU VRAM

17GB

Version number

5.2 -5.6

Browser

Firefox

Python dependencies

No response

What happened

Using any IP Adapter always triggers a reinstall and fails.

  1. Installed the latest version.
  2. adding an IP Adapter with a reference image
  3. strat generation

Now the log states:
[InvokeAI]::WARNING --> The image encoder required by this IP Adapter (IP Adapter SD1.5 Image Encoder) is not installed. Downloading and installing now. This may take a while.

The IP Adapter is downloaded and tried to install to invokeai\models\any\clip_vision in the folder "IP Adapter SD1"
However once it is downloaded I get the error message there is a model already installed (even though it was just downloaded and put there).

The logs look like that:

[InvokeAI]::ERROR --> Error while fetching CLIP Vision Image Encoder
[InvokeAI]::ERROR --> Error while invoking session 1f5ce095-1582-46ac-9b45-79f51a69b976, invocation deb9ca93-64b3-4ea3-a982-0f43a23118e0 (ip_adapter):
[InvokeAI]::ERROR --> Traceback (most recent call last):
  File "...\invokeai\.venv\lib\site-packages\invokeai\app\services\session_processor\session_processor_default.py", line 129, in run_node
    output = invocation.invoke_internal(context=context, services=self._services)
  File "...\invokeai\.venv\lib\site-packages\invokeai\app\invocations\baseinvocation.py", line 300, in invoke_internal
    output = self.invoke(context)
  File "...\invokeai\.venv\lib\site-packages\invokeai\app\invocations\ip_adapter.py", line 129, in invoke
    image_encoder_model = self.get_clip_image_encoder(context, image_encoder_model_id, image_encoder_model_name)
  File "...\invokeai\.venv\lib\site-packages\invokeai\app\invocations\ip_adapter.py", line 190, in get_clip_image_encoder
    assert len(image_encoder_models) == 1
AssertionError

Even a subsecuent invocation triggers the same with the same result. Deleting this directory again does not help either.
This seems only happen to IP Adapters. Without them (i.e. just a normal generation) works fine.

I first noticed this issue from version 5.2 on.

What you expected to happen

Either a normal installation of a missing model or recognition of the existing model.

How to reproduce the problem

No response

Additional context

Might be related to these issues:

#6352
#7398

Discord username

No response

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions