Cn autograd debugging - Githubissues

tufts-ml / covid19-forecasting

Simulation software for forecasting demand at various stages of hospitalization

MIT License

3 stars 0 forks source link

Cn autograd debugging #36

Open cuong450 opened 3 years ago

cuong450 commented 3 years ago

Here is Python notebook showing inefficient gradient descent with the use of PopulationData_ag.py. I think you would be most interested in reviewing the relevant functions within PopulationData_ag.py:

get_forecasted_data
get_loss_given_truthdict_and_params
get_grad_of_loss

cuong450 commented 3 years ago

Hi Mike, can you take a look at autograd_for_poisson.py and tell me what I am doing wrong in trying to get the logpmf gradient?