add dockerfile

[stone-wlg]

Quickly Demo

$ docker build -t nomic-ai/gpt4all:1.0.0 . $ docker run -it --rm -v ./gpt4all-lora-quantized.bin:/opt/gpt4all nomic-ai/gpt4all:1.0.0

Multi Arch

$ docker buildx build --platform linux/amd64,linux/arm64 --push -t nomic-ai/gpt4all:1.0.0 .

Mac Intel

$ docker buildx build --platform linux/amd64 --push -t nomic-ai/gpt4all-amd64:1.0.0 . $ docker run -it --rm -v ./gpt4all-lora-quantized.bin:/opt/gpt4all nomic-ai/gpt4all-amd64:1.0.0 /opt/gpt4all/gpt4all-lora-quantized-OSX-intel -m /opt/gpt4all/gpt4all-lora-quantized.bin

Mac m1

$ docker buildx build --platform linux/arm64 --push -t nomic-ai/gpt4all-arm64:1.0.0 . $ docker run -it --rm -v ./gpt4all-lora-quantized.bin:/opt/gpt4all nomic-ai/gpt4all-arm64:1.0.0 /opt/gpt4all/gpt4all-lora-quantized-OSX-m1 -m /opt/gpt4all/gpt4all-lora-quantized.bin

[stone-wlg]

If your going to include the binary, why not include the model as well?

And this would only work for a linux x86 image. Would be good to make one for arm as well.

maybe a multi-arch?

1. model ? you mean gpt4all-lora-quantized.bin, it download by wget when build image
2. yes, working on it.

[mkellerman]

All I'm saying is that each time you build the image, it will download a huge file. Might as well either: download the binary and the model, or copy both of them from the local workspace.

But nice work on the contribution! 💪

[stone-wlg]

If your going to include the binary, why not include the model as well?

And this would only work for a linux x86 image. Would be good to make one for arm as well.

maybe a multi-arch?

1. multi-arch, i use docker buildx to build image of amd64 and arm64, both are fine. but i can not test arm64, could you test it? if it is ok, maybe add these instructions to README.md
2. local model file, i change some in Dockerfile.

[AndriyMulyar]

Out of date