Embeddings model from website model explorer can't be instantiated in Python bindings

[AtR1an]

Bug Report

Instantiating the embeddings model all-MiniLM-L6-v2-f16.gguf (https://gpt4all.io/models/gguf/all-MiniLM-L6-v2-f16.gguf) from the model explorer on https://gpt4all.io/ fails with Unable to instantiate model: Unsupported model architecture: bert in the 2.6.0 version of the Python bindings.

Example Code

import gpt4all
gpt4all.Embed4All(model_name="all-MiniLM-L6-v2-f16.gguf", model_path=<download directory>)

This results in

  File "<stdin>", line 1, in <module>
  File "...\Lib\site-packages\gpt4all\gpt4all.py", line 58, in __init__
    self.gpt4all = GPT4All(model_name, n_threads=n_threads, device=device, **kwargs)
  File "...\Lib\site-packages\gpt4all\gpt4all.py", line 206, in __init__
    self.model = LLModel(self.config["path"], n_ctx, ngl)
  File "...\Lib\site-packages\gpt4all\_pyllmodel.py", line 218, in __init__
    raise RuntimeError(f"Unable to instantiate model: {'null' if s is None else s.decode()}")
RuntimeError: Unable to instantiate model: Unsupported model architecture: bert

Steps to Reproduce

1. Install gpt4all version 2.6.0 from pip: pip install gpt4all==2.6.0
2. Execute example code

Expected Behavior

The model available via the model explorer on the website also works with the Python bindings.

Your Environment

• Bindings version (e.g. "Version" from pip show gpt4all): Python 2.6.0
• Operating System: Windows (also reproduced on Ubuntu)
• Chat model used (if applicable): all-MiniLM-L6-v2-f16.gguf