chore: cache long-running stages and use add vs copy

Dockerfile

@@ -1,11 +1,13 @@
 FROM continuumio/miniconda3
 
 WORKDIR /usr/src/app
-COPY . .
-# install dependcies
-RUN conda install -y pandas numpy scikit-learn
+# install dependencies
+ADD requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
+
 EXPOSE 7860
 ENV GRADIO_SERVER_NAME="0.0.0.0"
 
+ADD . .
+
 CMD ["python", "inference_app.py"]