Feat/dsep nichr/less based sampler#10

Open

PaxPerVeritatem wants to merge 68 commits into

mainfrom

feat/dsep-nichr/LESSBasedSampler

Commits on Mar 21, 2025

transferred code from fork to new feature branch
nikch
committed
added job file and experiment script
nikch
committed
Ready to run full finetune experiment
nikch
committed
fixed selectivesampler enumerate "typo"
nikch
committed
removed hpc files
nikch
committed
singlesampler added
nikch
committed

Commits on Mar 25, 2025

initial LESS branch commit
Daniel
committed
LESSBasedSampler first implementation
nikch
committed

Commits on Mar 26, 2025

Commits on Mar 28, 2025

begun inital debugging for running LESSBasedSampler with subsampler
Daniel
committed
Think warmup works in subprocess now
nikch
committed
LESS step 1) LESS_sampler mask + model signiture bug fix begun
Daniel
committed
RadT only logs to default
Daniel
committed

Commits on Mar 31, 2025

ready to do experiments with RadT on HPC
Daniel
committed
experiments will be done without radT
Daniel
committed
minor changes
nikch
committed
Merged with singleSampler
nikch
committed
Merge branch 'feat/nichr/RSRS_sampler' into feat/dsep-nichr/LESSBasedSampler
nikch
committed

Commits on Apr 1, 2025

LESS recipes, configs and LESS torch-tune implementation encapsulated into seperate folders
Daniel
committed
ready to attempt first merge into LESSExperiments branch
Daniel
committed
test implementation for merge
Daniel
committed
removed test implementation
Daniel
committed

Commits on Apr 2, 2025

Commits on Apr 7, 2025

changed datasets to only use cleaned versions to enforce sanitation
Daniel
committed

Commits on Apr 8, 2025

Commits on Apr 9, 2025

Commits on Apr 14, 2025

added mkdir() call in checkpointer.py to ML-flow model dir creation and error handling. Added additional comment to eval_finetune.yaml configs to ease use of ML-flow model loading
Daniel
committed

Commits on Apr 15, 2025

LESS step2 starts now
nikch
committed

Commits on Apr 16, 2025

Commits on Apr 17, 2025

Commits on Apr 18, 2025

LESS step 3.1 implemented, calculating validation gradients works
Daniel
committed

Commits on Apr 21, 2025

Step1 now writes average learning rate to a csv file, step2 now takes multiple checkpoints and creates directories for each. Also trying to implement step 3.2
nikch
committed

Commits on Apr 22, 2025

Commits on Apr 23, 2025

Commits on Apr 28, 2025

Cleaned up step 3.1 (removed commented code and changed function name).
Daniel
committed
continueing work on step3.2
Daniel
committed

Commits on Apr 29, 2025

Setup function in step 3.2 done, continueing with testing
Daniel
committed
totally lost, initiallizeing Extreme Programming
Daniel
committed

Commits on Apr 30, 2025

.
niklas-chr
committed

Commits on May 6, 2025

Commits on May 12, 2025

Commits on May 29, 2025

done
niklas-chr
committed