Automating garment classification is the aim of this repo, pytorch's CNN model was implemented for image classification to automatically categorize new clothing listings, making it easier for a person, customer to find what they're looking for. This will assist in inventory management by quickly sorting items. The primary objective is to develop a machine learning model capable of accurately categorizing images of clothing items into distinct garment types such as shirts, trousers, shoes, etc.
This is the link to the common FashionMNIST dataset: https://github.com/zalandoresearch/fashion-mnist
Sample of the fabric items with label from the data
Above is the image network on the model, viewed on netron.app
Get the implementation of the webservice with fastapi in the serve.py script file
a quick results on the get and post services:
#Dockerfile development
--> the python 3.11-slim as the base image
FROM python:3.11-slim
--> create a working directory, '/app'
WORKDIR /app
--> copy all the files from local maachine directory to the app directory previously created on 2.
COPY . /app
--> install the libraries in the .txt file in desired environment, pip in this case
RUN pip install -r requirements.txt
--> to host locally on the docker image and expose through the port 8000 on the same image, using uvicorn + fastapi
CMD ["uvicorn", "serve:app", "--host", "0.0.0.0", "--port", "8000"]
https://github.com/MLHermit/mlzoomcamp/blob/main/README.md A link to the final certificate
https://github.com/DataTalksClub/machine-learning-zoomcamp a link to the parent dtc ml zoomcamp