Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

New inference 🎨 #203

Open
wants to merge 159 commits into
base: main
Choose a base branch
from
Open

New inference 🎨 #203

wants to merge 159 commits into from

Conversation

jcblemai
Copy link
Collaborator

@jcblemai jcblemai commented Apr 12, 2024 •
edited

Functionalities to support several inference algorithms, make flepimop install a one-line thing, overhaul file saving and analysis

  • convert the internal representation of epi states from data frame to xarray
  • save epi state only on command
  • update documentation for new packages
  • implements class for statistics
  • implement likelihood computation
  • Finish and test statistic class FOR ROUND
  • Make so we can resume FOR ROUND while not fitting all parameters but some of them (the other ? drawn ??)
  • make the folder draw thing
  • workflow config
  • make an inference benchmark
  • so many things...
  • implement emcee as a script
  • rewrite resume and continuation batch
  • rewrite current inference algorithm
  • make it possible to batch with and without MPI, or on single node.

I made a mistake and did here the changes to remove time and to switch the R code to date, sorry about that. The changed R files are because of this.

jcblemai and others added 30 commits April 9, 2024 14:47
@jcblemai
starting to gempyorize the things
c438cff
@jcblemai
seir simulation is a xarray now
f3832ba
@jcblemai
interface for inference
63e9359
@jcblemai
R uses that to write on accept
009a183
@jcblemai
updated the name of the cached storage within interface
56a611a
@jcblemai
progressing on stat
82a947a
@jcblemai
compute log of loss
bee1635
@jcblemai
logloss
0f2e5f7
@jcblemai
starting to get an inference structure
53d03a2
@jcblemai
simpler inferenceParameter structure
483abe6
@jcblemai
class seems to be quite done
bf96096
@jcblemai
better
c8d60b2
@jcblemai
atomic simulation, duplicate
55dadb7
@jcblemai
entry point for calibration, TODO
bb5e677
@jcblemai
Re-adding the parameter manager, after changing name
9bf3f9b
@jcblemai
vectorized inference param
758046d
@jcblemai
example and bbenchmark
619444c
@jcblemai
a much clearner interface and preparing the config
bb1a114
@jcblemai
here you go @saraloo
797294b
@jcblemai
new syntax
f949901
@jcblemai
example generated data
f1cc4cf
@jcblemai
benchmark
0178f87
@jcblemai
typo
0513a11
@jcblemai
get rid of time in column (only date)
bcd0c4f
@jcblemai
trying to remove time from the R code
08f5f0e
@saraloo
add other distributions to logloss
34b2314
@jcblemai
Merge branch 'gempyor-plugin' into new_inference
9f87b18
@jcblemai
progress
45d885b
@jcblemai
merge
84f18b8
@jcblemai
several regularization
90bd280
jcblemai and others added 30 commits April 30, 2024 19:00
@jcblemai
slow progression
cc4135d
@jcblemai
load the new object as part of the modelInfo
a159a74
@jcblemai
complete reorg
334bd18
@jcblemai
adding file changed name
616c089
@jcblemai
it now work
a20f3ff
@jcblemai
forgot seed
cd76179
@jcblemai
rewrote save chain function, the feeling of doing everything 3 times
dec6839
@jcblemai
another forgotten file
3b7123c
@jcblemai
it works, the script
f418488
@jcblemai
black
f1aa942
@saraloo
fix maine data
a6145ff
@saraloo
add rolling mean to flepiconfig
d8f4d0b
@saraloo
typo
6c7d308
@jcblemai
trying to optimize
edd0020
@jcblemai
Merge branch 'new_inference' of https://github.com/HopkinsIDD/flepiMoP …
e53dd26 
…into new_inference
@jcblemai
error in R call
dfa8701
@jcblemai
benchmark
7a6f1a7
@jcblemai
replace all the cat by print for error
d45620d
@jcblemai
fix for inference
abdbec8
@jcblemai
regularisation in projection period
735e508
@saraloo
fix init file saving
1e0c56d
@jcblemai
more reg
191bb07
@jcblemai
make it work also when it does not save seir
d91fcd7
@jcblemai
progress on benchmark
ad50c8f
@jcblemai
fix inference making a quasi-random walk
62c9957
@jcblemai
Merge branch 'new_inference' of https://github.com/HopkinsIDD/flepiMoP …
de758e3 
…into new_inference
@jcblemai
adding a new plot quickly
0227987
@jcblemai
fix
bcb5d99
@jcblemai
remove the hack
829356e
@jcblemai
pruning for R18
566285f
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees

@jcblemai jcblemai

Labels
inference Concerns the parameter inference framework
Projects
None yet
Milestone
Inference Rebuild
Development

Successfully merging this pull request may close these issues.

Inference file writing
2 participants
@jcblemai @saraloo