move repositories:
in machine learning, a model checkpoint is a file containing the weights (i.e., the parameters) of a trained model at a particular point during training.
a model card is a document that provides information about a machine learning model. I.e.:
- https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md
- https://github.com/tatsu-lab/stanford_alpaca/blob/main/model_card.md
a ML framework. See example pytorch/
- for deploying the model i use flask
model-webapp/helloflask.py
is sample hello world flask app.
- run it with:
export FLASK_APP=helloflask.py
python -m flask run
- open browser (or curl -i) http://127.0.0.1:5000/
- debug mode can be activated by
export FLASK_ENV=development
- call app.run function in main
if __name__ == '__main__':
app.run(debug=True)
famous ml dataset: fisher's iris data set (iris.data) https://archive.ics.uci.edu/ml/datasets/iris