🐛 Bug Report: HuggingFace Mistral models not working

[Victorivus]

📜 Description

Using a HuggingFace model from the Mistral family.

👟 Reproduction steps

with the parameters from the documentation:
LLM_NAME=huggingface
EMBEDDINGS_NAME=sentence-transformers/all-mpnet-base-v2

and setting the llm_name to 'mistralai/Mistral-7B-Instruct-v0.1' instead of 'Arc53/DocsGPT-7B' in the DocsGPT/application/llm/huggingface.py file. You launch the app and ask a question.

In the DocsGPT/application/api/answer/routes.py completion = llm.gen_stream( call must be also replaced by completion = llm.gen( to avoid NotImplementedErrorand get a reponse

👍 Expected behavior

It should work and answer

👎 Actual Behavior with Screenshots

It crashes with KeyError: 'mistral'

💻 Operating system

Linux

What browsers are you seeing the problem on?

No response

🤖 What development environment are you experiencing this bug on?

Docker

🔒 Did you set the correct environment variables in the right path? List the environment variable names (not values please!)

LLM_NAME
EMBEDDINGS_NAME
FLASK_APP
FLASK_DEBUG
CELERY_BROKER_URL
CELERY_RESULT_BACKEND

📃 Provide any additional context for the Bug.

The bug is corrected updating the transformers library pip install -U transformers

📖 Relevant log output

Downloading tokenizer_config.json: 100%|█████████| 1.47k/1.47k [00:00<00:00, 4.23MB/s]
Downloading tokenizer.model: 100%|█████████████████| 493k/493k [00:00<00:00, 11.9MB/s]
Downloading tokenizer.json: 100%|████████████████| 1.80M/1.80M [00:00<00:00, 10.1MB/s]
Downloading (…)cial_tokens_map.json: 100%|██████████| 72.0/72.0 [00:00<00:00, 757kB/s]
Downloading config.json: 100%|███████████████████████| 571/571 [00:00<00:00, 7.32MB/s]
Debugging middleware caught exception in streamed response at a point where response headers were already sent.
Traceback (most recent call last):
  File "/VSCode/DocsGPT/venv/lib/python3.10/site-packages/werkzeug/wsgi.py", line 256, in __next__
    return self._next()
  File "/VSCode/DocsGPT/venv/lib/python3.10/site-packages/werkzeug/wrappers/response.py", line 32, in _iter_encoded
    for item in iterable:
  File "/VSCode/DocsGPT/application/api/answer/routes.py", line 99, in complete_stream
    llm = LLMCreator.create_llm(settings.LLM_NAME, api_key=api_key)
  File "/VSCode/DocsGPT/application/llm/llm_creator.py", line 24, in create_llm
    return llm_class(*args, **kwargs)
  File "/VSCode/DocsGPT/application/llm/huggingface.py", line 23, in __init__
    model = AutoModelForCausalLM.from_pretrained(llm_name)
  File "/VSCode/DocsGPT/venv/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py", line 456, in from_pretrained
    config, kwargs = AutoConfig.from_pretrained(
  File "/VSCode/DocsGPT/venv/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py", line 957, in from_pretrained
    config_class = CONFIG_MAPPING[config_dict["model_type"]]
  File "/VSCode/DocsGPT/venv/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py", line 671, in __getitem__
    raise KeyError(key)
KeyError: 'mistral'
127.0.0.1 - - [12/Dec/2023 15:08:27] "POST /stream HTTP/1.1" 200 -
^C
worker: Hitting Ctrl+C again will terminate all running tasks!

worker: Warm shutdown (MainProcess)

👀 Have you spent some time to check if this bug has been raised before?

• I checked and didn't find similar issue

🔗 Are you willing to submit PR?

Yes, I am willing to submit a PR!

🧑‍⚖️ Code of Conduct

• I agree to follow this project's Code of Conduct

[Victorivus]

This bug gets fixed with a pip install -U transformers.

The modified libs are here: #801