Label by rule: infected (2) vs uninfected (1); hand annotations removed

settings/classify_settings.csv

@@ -1,52 +1,52 @@
-Key,Value
-amsgrad,True
-annotation_column,annotate
-augment,False
-batch_size,32
-class_balance,none
-class_folder_names,"['nc', 'pc']"
-classes,"['1', '2']"
-classifier_evaluation,True
-crop_source,auto
-cross_validation_enabled,False
-cross_validation_folds,0
-custom_model_path,
-cv_group_by,well
-dropout_rate,0.1
-epochs,10
-evaluation_bins,10
-evaluation_calibration,temperature
-evaluation_fail_on_leakage,True
-gradient_accumulation,True
-gradient_accumulation_steps,4
-holdout_plate,
-image_size,224
-init_weights,True
-intermedeate_save,True
-leakage_audit_train_test,True
-leakage_hash_content,True
-leakage_require_identity,True
-learning_rate,0.001
-loss_type,focal_loss
-max_failure_rate,
-mixed_precision,False
-model_type,maxvit_t
-n_jobs,30
-nested_cv_inner_folds,0
-normalize,True
-optimizer_type,adamw
-pin_memory,False
-plot,True
-resume_checkpoint,
-schedule,cosine
-src,<dataset>
-strict_errors,
-tensorboard,True
-test,False
-train,True
-train_channels,"['r', 'g', 'b']"
-train_mode,erm
-use_checkpoint,True
-val_split,0.1
-verbose,False
-weight_decay,1e-05
+Key,Value
+amsgrad,True
+annotation_column,infected
+augment,False
+batch_size,32
+class_balance,none
+class_folder_names,"['nc', 'pc']"
+classes,"['1', '2']"
+classifier_evaluation,True
+crop_source,auto
+cross_validation_enabled,False
+cross_validation_folds,0
+custom_model_path,
+cv_group_by,well
+dropout_rate,0.1
+epochs,10
+evaluation_bins,10
+evaluation_calibration,temperature
+evaluation_fail_on_leakage,True
+gradient_accumulation,True
+gradient_accumulation_steps,4
+holdout_plate,
+image_size,224
+init_weights,True
+intermedeate_save,True
+leakage_audit_train_test,True
+leakage_hash_content,True
+leakage_require_identity,True
+learning_rate,0.001
+loss_type,focal_loss
+max_failure_rate,
+mixed_precision,False
+model_type,maxvit_t
+n_jobs,30
+nested_cv_inner_folds,0
+normalize,True
+optimizer_type,adamw
+pin_memory,False
+plot,True
+resume_checkpoint,
+schedule,cosine
+src,<dataset>
+strict_errors,
+tensorboard,True
+test,False
+train,True
+train_channels,"['r', 'g', 'b']"
+train_mode,erm
+use_checkpoint,True
+val_split,0.1
+verbose,False
+weight_decay,1e-05