sbrl-github/research-rainfallradar
1
0
Fork 0
mirror of https://github.com/sbrl/research-rainfallradar synced 2024-10-22 11:43:00 +00:00
Code Issues Projects Releases Packages Wiki Activity

README: Document SPLIT_VALIDATE, SPLIT_TEST

Browse source
This commit is contained in:
Starbeamrainbowlabs 2024-08-30 18:49:38 +01:00
parent 0761651ccf
commit f46a86c9a4
Signed by: sbrl
GPG key ID: 1BE5172E637709C2
1 changed files with 2 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
README.md
View file

@ -158,6 +158,8 @@ LEARNING_RATE=0.001 | The learning rate to use. Default: 0.001.
UPSAMPLE=2 | How much to upsample by at the beginning of the model. A value of disables upscaling. Default: 2. UPSAMPLE=2 | How much to upsample by at the beginning of the model. A value of disables upscaling. Default: 2.
STEPS_PER_EXECUTION=1 | How many steps to perform before surfacing from the GPU to e.g. do callbacks. Default: 1. STEPS_PER_EXECUTION=1 | How many steps to perform before surfacing from the GPU to e.g. do callbacks. Default: 1.
RANDSEED | The random seed to use when shuffling filepaths. Default: unset, which means use a random value. RANDSEED | The random seed to use when shuffling filepaths. Default: unset, which means use a random value.
SPLIT_VALIDATE | Percentage of the available files in the dataset to be allocated to the validation split. Default: 0.2
SPLIT_TEST | Percentage of the available files in the dataset to be allocated to the test split. Default: 0.2
JIT_COMPILE | Set to any value to compile the model with XLA. Defaults to unset; set to any value to enable. JIT_COMPILE | Set to any value to compile the model with XLA. Defaults to unset; set to any value to enable.
PREDICT_COUNT=25 | The number of items from the (SCRAMBLED) dataset to make a prediction for. PREDICT_COUNT=25 | The number of items from the (SCRAMBLED) dataset to make a prediction for.
PREDICT_AS_ONE | [prediction only] Set to any value to avoid splitting the input dataset into training/validation and instead treat it as a single dataset. Default: False (treat it as training/validation) PREDICT_AS_ONE | [prediction only] Set to any value to avoid splitting the input dataset into training/validation and instead treat it as a single dataset. Default: False (treat it as training/validation)