Catch generic model errors

[d8ahazard]

Prevent the app from dying on invalid models.

What type of PR is this? (check all applicable)

• Refactor
• Feature
• [X ] Bug Fix
• Optimization
• Documentation Update
• Community Node Submission

Have you discussed this change with the InvokeAI team?

• Yes
• [ X] No, because: Just discovered it, seemed like a simple fix.

Have you updated all relevant documentation?

• [X ] Yes
• No

Description

Fix multiple potential issues when scanning models on first load, specifically with an instructPix2Pix model and some non-safetensor files.

Related Tickets & Documents

• Related Issue #
• Closes #

QA Instructions, Screenshots, Recordings

Added/updated tests?

• Not Applicable
• Yes
• No : please replace this line with details on why tests
  have not been included

[optional] Are there any post deployment tasks we need to perform?

Nope.

[RyanJDick]

Seems reasonable to me. We're already tolerating a bunch of model load errors - might as well tolerate them all.

We should tidy our error handling in this block, but that can be left for a future PR.

Thanks for this contribution!

[d8ahazard]

Seems reasonable to me. We're already tolerating a bunch of model load errors - might as well tolerate them all.

We should tidy our error handling in this block, but that can be left for a future PR.

Thanks for this contribution!

My pleasure! Love the project, doubly so for the new features you guys have been adding.

FWIW, it may also be useful to filter "non-checkpoint" models when scanning, I'm getting a lot of errors for custom metadata files I have alongside my models: