Problem updating reranking model: 'NoneType' object has no attribute 'get'

[chrisoutwright]

Bug Report

Description

Bug Summary:
When attempting to update the reranking model from llm-blender/PairRM to llm-blender/PairRM in Open WebUI, an AttributeError occurs. This error suggests a possible issue with model metadata not being properly initialized or handled, specifically the NoneType object not having a get attribute.

Steps to Reproduce:

1. Launch Open WebUI.
2. Navigate to the reranking model update section via the UI.
3. Trigger a model update by selecting the option to update from llm-blender/PairRM to llm-blender/PairRM.
4. Observe the error in the system logs and the failure notification in the UI.

Expected Behavior:
The model update should complete successfully without errors, allowing the new reranking model to be utilized for enhanced performance.

Actual Behavior:
The update fails, resulting in a 500 Internal Server Error. The logs indicate an AttributeError due to trying to access a get method on a NoneType object.

Environment

• Open WebUI Version: v0.1.124 (latest)
• Ollama Version: 0.1.36
• Operating System: Windows 11 with WSL2 and Docker for Desktop
• Browser: Chrome

Logs Details

INFO:apps.rag.main:Updating reranking model: ibm/re2g-reranker-nq to llm-blender/PairRM
Fetching 11 files: 100%|
ERROR:apps.rag.main:Problem updating reranking model: 'NoneType' object has no attribute 'get'
Traceback (most recent call last):
  File "C:\Users\Chris\OpenWebUI\open-webui\backend\apps\rag\main.py", line 298, in update_reranking_config
    update_reranking_model(app.state.RAG_RERANKING_MODEL, True)
  File "C:\Users\Chris\OpenWebUI\open-webui\backend\apps\rag\main.py", line 151, in update_reranking_model
    app.state.sentence_transformer_rf = sentence_transformers.CrossEncoder(
                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Chris\AppData\Local\Programs\Python\Python312\Lib\site-packages\sentence_transformers\cross_encoder\CrossEncoder.py", line 81, in __init__
    self.model = AutoModelForSequenceClassification.from_pretrained(
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Chris\AppData\Local\Programs\Python\Python312\Lib\site-packages\transformers\models\auto\auto_factory.py", line 563, in from_pretrained
    return model_class.from_pretrained(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Chris\AppData\Local\Programs\Python\Python312\Lib\site-packages\transformers\modeling_utils.py", line 3313, in from_pretrained
    if metadata.get("format") == "pt":
       ^^^^^^^^^^^^

Confirmation:

• I have read and followed all the instructions provided in the README.md.
• I am on the latest version of both Open WebUI and Ollama.
• I have included the browser console logs.
• I have included the Docker container logs.