update readme

Dockerfile

@@ -1,7 +1,7 @@
-# You can adapt the base image based on the CUDA version installed on the device
+# adapt cuda version based on your cuda-version, TERMINAL: `nvidia-smi`
 FROM nvidia/cuda:11.4.3-cudnn8-runtime-ubuntu20.04
 
-#Set up environment
+# Set up environment
 RUN apt-get update && apt-get install --no-install-recommends --no-install-suggests -y curl
 RUN apt-get install -y git
 RUN apt-get install unzip
@@ -16,14 +16,13 @@ COPY . .
 
 # Create cache directory and set permissions
 RUN mkdir -p /restful-llama-3/cache && chmod -R 777 /restful-llama-3/cache
+# Make start_app.sh executable
+RUN chmod +x /restful-llama-3/start_app.sh
 
 ENV LC_ALL=C.UTF-8
 ENV LANG=C.UTF-8
 ENV HF_HOME=/restful-llama-3/cache
 ENV HF_DATASETS_CACHE=/restful-llama-3/cache
 
-# Make start_app.sh executable
-RUN chmod +x /restful-llama-3/start_app.sh
-
 EXPOSE 7860
 CMD ["./start_app.sh"]

README.md

@@ -14,13 +14,16 @@ A simple RESTful service for the Meta-Llama-3.1-8B-Instruct language model.
 
 ## Pre-requisites
 
-1. A CUDA enabled GPU machine, runs optimal with 24GB vRAM
+1. A CUDA enabled GPU Space, runs optimal with 24GB vRAM
 2. Access to [LLaMa-3.1 weights](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) from Huggingface
-3. Personal Access Token copied to Secret Access Token(space) named HUGGING_FACE_HUB_TOKEN
+3. New Public Hugging Face Space <https://huggingface.co/docs/hub/spaces-overview>, Blank Docker Container
+4. Personal Access Token (Read) <https://huggingface.co/docs/hub/security-tokens>, save it somewhere safe
+5. Secret Access Token in your space. Name: HUGGING_FACE_HUB_TOKEN, Value: Personal Access Token
+6. (Optional) Local .env file to store Personal Access Token
 
 ## Getting Started
 
-1. Copy/Push this repo via ssh to your huggingface-space, blank docker-container setup
+1. Push this repo via SSH to your Hugging Face space
 
 ## How to use
 

requirements.txt

@@ -1,6 +1,5 @@
 accelerate==0.33.0
 torch==2.4.0
 fastapi==0.112.0
-gunicorn==22.0.0
 transformers==4.43.3
 uvicorn==0.30.5