gpt4all Issue: gpt4all==0.2.3 for AARCH64/ARM64

Issue you'd like to raise.

When installing gpt4all using pip the highest version supported is 0.1.7.

Please add support for pip installation of the latest version (0.2.3) of gpt4all on aarch64 machines.

Or alternatively please let me know how I can install the latest gpt4all on ubuntu 22.04 / aarch64 machine.

Thank you.

Suggestion:

No response

May 28 '23 14:05 mateenmalik

+1

May 31 '23 13:05 YoeJates

Have the same issue: ERROR: Could not find a version that satisfies the requirement gpt4all==0.2.3 (from versions: 0.1.5, 0.1.6, 0.1.7) ERROR: No matching distribution found for gpt4all==0.2.3

Please could you help?

May 31 '23 18:05 AndrewWebDev

+1 got same problem

Jun 02 '23 09:06 pakcheera

+1

Jun 04 '23 07:06 kemalony

+1

Jun 04 '23 12:06 rocky2018

+1
mac m1pro

Jun 05 '23 07:06 just-wkj

I managed to install it using pip for python 3.11.2 (installed via pyenv) on my Mac M2. However, this failed with docker build . e.g. for an image FROM python:3.11.2-slim. I could however build and install gpt4all locally which appears to satisfy 0.2.3.

EDIT:

Local installation appears to fail due to validation issues. Output:

    llm = GPT4All(model=model_path, n_ctx=model_n_ctx, backend='gptj', callbacks=callbacks, verbose=False)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/pydantic/main.py", line 341, in __init__
    raise validation_error
pydantic.error_wrappers.ValidationError: 1 validation error for GPT4All
__root__
  Failed to retrieve model (type=value_error)

Jun 05 '23 08:06 Muxelmann

+1

Jun 07 '23 18:06 Tobiaske1234

+1

Jun 08 '23 06:06 iyuyoung

+1

Jun 09 '23 13:06 augelloantonio

+1

Jun 10 '23 22:06 franciscogar94

@mateenmalik The better title might be "Build wheels for Linux aarch64/arm64". You can still build gpt4all locally as described here: https://github.com/nomic-ai/gpt4all/blob/main/gpt4all-bindings/python/README.md. @Muxelmann you might need to download a model file from the bottom of https://gpt4all.io/index.html. This causes https://github.com/imartinez/privateGPT/issues/729.

Jun 17 '23 12:06 Pipeliner

I have this issue with gpt4all==0.3.3 as well, on a docker build under MacOS with M2. On the MacOS platform itself it works, though. The problem is with a Dockerfile build, with "FROM arm64v8/python:3.9" or even "FROM python:3.10.12". With this one it pip3/installs: "FROM tiangolo/uvicorn-gunicorn-fastapi:python3.9"; unfortunately it fails to load the ggml-gpt4all-j-v1.3-groovy.bin model.