fix probing for ip_adapter folders (#4669)

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

- [X] Bug Fix
- [ ] Optimizatio

## Have you discussed this change with the InvokeAI team?
- [ ] Yes
- [X] Np

      
## Have you updated all relevant documentation?
- [ ] Yes
- [X] No


## Description

ip_adapter models live in a folder containing the file
`image_encoder.txt` and a safetensors file. The load-time probe for new
models was detecting the files contained within the folder rather than
the folder itself, and so models.yaml was not getting correctly updated.
This fixes the issue.

## Added/updated tests?

- [ ] 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?

invokeai/backend/model_management/model_search.py

@@ -71,7 +71,13 @@ class ModelSearch(ABC):
                 if any(
                     [
                         (path / x).exists()
-                        for x in {"config.json", "model_index.json", "learned_embeds.bin", "pytorch_lora_weights.bin"}
+                        for x in {
+                            "config.json",
+                            "model_index.json",
+                            "learned_embeds.bin",
+                            "pytorch_lora_weights.bin",
+                            "image_encoder.txt",
+                        }
                     ]
                 ):
                     try: