dockerfile lable error image to dangling easy fix for you if you want :D

[sycomix]

Duplicates

• I have searched the existing issues

Steps to reproduce 🕹

=> => exporting layers 8.5s
=> => exporting manifest sha256:81b29524e6ca86716c44c2fa16b8dc312af04dd88c3e3c03af98f087b650c8f4 0.0s
=> => exporting config sha256:00a84a43fa0487344f94620038375e6d5607fc5f482c144bf8e938ceb7c76803 0.0s
=> => naming to dangling@sha256:81b29524e6ca86716c44c2fa16b8dc312af04dd88c3e3c03af98f087b650c8f4 0.0s
=> => unpacking to dangling@sha256:81b29524e6ca86716c44c2fa16b8dc312af04dd88c3e3c03af98f087b650c8f4

but something like perks it right up, the source image can be swapped for rocm, intel, arm etc. i habve a cuda gpu so i played to my strong suite.

Use an official CUDA runtime as a parent image

FROM nvidia/cuda:11.5.0-runtime-ubuntu20.04

Install Python and any necessary dependencies

RUN apt-get update && apt-get install python3.11 python3-pip -y

Set the working directory to /app

WORKDIR /app

Copy the scripts and requirements.txt files into the container at /app

COPY scripts/ /app/scripts/
COPY requirements.txt /app/

Install any necessary Python packages

RUN ls //requirements.txt|xargs -n 1 -P 3 pip install -r

Set any necessary environment variables

ENV CUDA_VISIBLE_DEVICES=all

Set the command to run when the container starts

CMD ["python3.11", "/bin/bash"]

Current behavior 😯

failure to store and run image

Expected behavior 🤔

store and run image

Your prompt 📝

# Paste your prompt here
``` doesnt get that far

[rhohndorf]

Since llama.cpp is CPU only we should not need CUDA right now. I'm looking into running quantized models ion GPU though.
The Dockerfile does not copy and models and the code doesn't download any on initialization so yeah, currently dockerizing should not work.