Branch update#13

Merged

gperdrizet merged 20 commits intotelegramfrom main

Jul 21, 2024

+2,400-714

Commits on Jul 11, 2024

Added hf_model_string to results of model loading benchmark, restarting runs was not working correctly without it.
gperdrizet
committed
Merge branch 'main' into benchmarking to get updated requierments.txt on branch.
gperdrizet
committed
Added fence to avoid asking for first element of run dictionaries list if it's empty because all runs are complete.
gperdrizet
committed
Added note about HuggingFace login for gated models.
gperdrizet
committed
Finished dealing with the edge case where we are asked to resume an experiment which is already complete.
gperdrizet
committed
Moved CPU core count assignment to batch runner function, i.e. before the LLM(s) are loaded after getting strange results from model loading benchmark.
gperdrizet
committed
Fixed bug where CPU thread count was not properly being set in LLM class during benchmarking runs.
gperdrizet
committed

Commits on Jul 16, 2024

Updated plots with new data, refactored factor level exclusion logic out of plotting functions. Added some other quality of life improvements.
gperdrizet
committed

Commits on Jul 18, 2024

Started work on cleaning up and updating notebooks with new benchmarking data, added master outline of notebooks.
gperdrizet
committed

Commits on Jul 19, 2024

Commits on Jul 20, 2024