Skip to content

Conversation

@robertsjames
Copy link
Collaborator

No description provided.

…/stepped spectrum, when using source groups.
…atch, execute the tf.map_fn over these. Now, need to access them in the function, and similarly read in the corresponding electron/photon domains.
… to the saved quanta tensors for the energies in each batch.
josh0-jrg and others added 30 commits November 7, 2024 06:34
TODO: make FV depend on time-stamp
Wei is working on Xe124 DEC model
Had major issues with input cleanliness, limited by a fudge factor of no x5 variation
Testing it works in simulate and diff rate land
This is not a good fix
Works for mono-energetic sources but fails for flat sources!
Alt method of matching shape... didn't work
Still fails, potential solution is an Exnevts tensor or, worse, an E x yields tensor but the latter may break gradients
Ends a long and painful saga, need to find a better implementation of bound calculation now
Kepy using default feild for ions dimension assuming low variability
Leads to a factor ~2-5x speed up, depend on sample size and how subsequent blocks scale, works in simulate and therfore estimate_mu. Be careful to check this doesn't bias your result
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants