add missing changes from original repository

configs/callbacks/default.yaml

@@ -1,7 +1,7 @@
 model_checkpoint:
  _target_: pytorch_lightning.callbacks.ModelCheckpoint
- monitor: "val/acc" # name of the logged metric which determines when model is improving
- mode: "max" # "max" means higher metric value is better, can be also "min"
+ monitor: "val/mae" # name of the logged metric which determines when model is improving
+ mode: "min" # "max" means higher metric value is better, can be also "min"
  save_top_k: 1 # save k best models (determined by above metric)
  save_last: True # additionaly always save model from last epoch
  verbose: False
@@ -11,8 +11,8 @@ model_checkpoint:
 
 early_stopping:
  _target_: pytorch_lightning.callbacks.EarlyStopping
- monitor: "val/acc" # name of the logged metric which determines when model is improving
- mode: "max" # "max" means higher metric value is better, can be also "min"
+ monitor: "val/mae" # name of the logged metric which determines when model is improving
+ mode: "min" # "max" means higher metric value is better, can be also "min"
  patience: 100 # how many validation epochs of not improving until training stops
  min_delta: 0 # minimum change in the monitored metric needed to qualify as an improvement
 

configs/datamodule/focus150.yaml

@@ -0,0 +1,8 @@
+_target_: src.datamodules.focus_datamodule.FocusDataModule
+
+data_dir: ${data_dir}/focus150 # data_dir is specified in config.yaml
+csv_file: ${data_dir}/focus150/metadata.csv
+batch_size: 64
+train_val_test_split_percentage: [0.7, 0.15, 0.15]
+num_workers: 0
+pin_memory: False

configs/experiment/focus_example.yaml

@@ -0,0 +1,31 @@
+# @package _global_
+
+# to execute this experiment run:
+# python train.py experiment=example
+
+defaults:
+ - override /datamodule: focus.yaml
+ - override /model: focus.yaml
+ - override /callbacks: default.yaml
+ - override /logger: tensorboard
+ - override /trainer: default.yaml
+
+# all parameters below will be merged with parameters from default configurations set above
+# this allows you to overwrite only specified parameters
+
+# name of the run determines folder name in logs
+name: "focus_example"
+seed: 12345
+
+trainer:
+ min_epochs: 1
+ max_epochs: 100
+
+model:
+ lin1_size: 128
+ lin2_size: 256
+ lin3_size: 64
+
+datamodule:
+ batch_size: 64
+ 

configs/hparams_search/focus150_optuna.yaml

@@ -0,0 +1,64 @@
+# @package _global_
+
+# example hyperparameter optimization of some experiment with Optuna:
+# python train.py -m hparams_search=mnist_optuna experiment=example
+
+defaults:
+ - override /datamodule: focus150.yaml
+ - override /model: focus150.yaml
+ - override /hydra/sweeper: optuna
+
+# choose metric which will be optimized by Optuna
+# make sure this is the correct name of some metric logged in lightning module!
+optimized_metric: "val/mae_best"
+
+name: "focus150_optuna"
+
+# here we define Optuna hyperparameter search
+# it optimizes for value returned from function with @hydra.main decorator
+# docs: https://hydra.cc/docs/next/plugins/optuna_sweeper
+hydra:
+ sweeper:
+ _target_: hydra_plugins.hydra_optuna_sweeper.optuna_sweeper.OptunaSweeper
+
+ # storage URL to persist optimization results
+ # for example, you can use SQLite if you set 'sqlite:///example.db'
+ storage: null
+
+ # name of the study to persist optimization results
+ study_name: focus150_hyperparameter_search
+
+ # number of parallel workers
+ n_jobs: 1
+
+ # 'minimize' or 'maximize' the objective
+ direction: minimize
+
+ # total number of runs that will be executed
+ n_trials: 25
+
+ # choose Optuna hyperparameter sampler
+ # docs: https://optuna.readthedocs.io/en/stable/reference/samplers.html
+ sampler:
+ _target_: optuna.samplers.TPESampler
+ seed: 12345
+ n_startup_trials: 10 # number of random sampling runs before optimization starts
+
+ # define range of hyperparameters
+ search_space:
+ datamodule.batch_size:
+ type: categorical
+ choices: [32, 64, 128]
+ model.lr:
+ type: float
+ low: 0.0001
+ high: 0.2
+ model.lin1_size:
+ type: categorical
+ choices: [32, 64, 128, 256, 512, 1024, 2048]
+ model.lin2_size:
+ type: categorical
+ choices: [32, 64, 128, 256, 512, 1024, 2048]
+ model.lin3_size:
+ type: categorical
+ choices: [32, 64, 128, 256, 512, 1024, 2048]

configs/hparams_search/focus_optuna.yaml

@@ -0,0 +1,62 @@
+# @package _global_
+
+# example hyperparameter optimization of some experiment with Optuna:
+# python train.py -m hparams_search=mnist_optuna experiment=example
+
+defaults:
+ - override /hydra/sweeper: optuna
+
+# choose metric which will be optimized by Optuna
+# make sure this is the correct name of some metric logged in lightning module!
+optimized_metric: "val/mae_best"
+
+name: "focus_optuna"
+
+# here we define Optuna hyperparameter search
+# it optimizes for value returned from function with @hydra.main decorator
+# docs: https://hydra.cc/docs/next/plugins/optuna_sweeper
+hydra:
+ sweeper:
+ _target_: hydra_plugins.hydra_optuna_sweeper.optuna_sweeper.OptunaSweeper
+
+ # storage URL to persist optimization results
+ # for example, you can use SQLite if you set 'sqlite:///example.db'
+ storage: null
+
+ # name of the study to persist optimization results
+ study_name: focus_hyperparameter_search
+
+ # number of parallel workers
+ n_jobs: 1
+
+ # 'minimize' or 'maximize' the objective
+ direction: minimize
+
+ # total number of runs that will be executed
+ n_trials: 25
+
+ # choose Optuna hyperparameter sampler
+ # docs: https://optuna.readthedocs.io/en/stable/reference/samplers.html
+ sampler:
+ _target_: optuna.samplers.TPESampler
+ seed: 12345
+ n_startup_trials: 10 # number of random sampling runs before optimization starts
+
+ # define range of hyperparameters
+ search_space:
+ datamodule.batch_size:
+ type: categorical
+ choices: [32, 64, 128]
+ model.lr:
+ type: float
+ low: 0.0001
+ high: 0.2
+ model.lin1_size:
+ type: categorical
+ choices: [32, 64, 128, 256, 512, 1024, 2048]
+ model.lin2_size:
+ type: categorical
+ choices: [32, 64, 128, 256, 512, 1024, 2048]
+ model.lin3_size:
+ type: categorical
+ choices: [32, 64, 128, 256, 512, 1024, 2048]

configs/model/focus150.yaml

@@ -0,0 +1,9 @@
+_target_: src.models.focus_module.FocusLitModule
+
+input_size: 67500
+lin1_size: 256
+lin2_size: 256
+lin3_size: 256
+output_size: 1
+lr: 0.001
+weight_decay: 0.0005

configs/train.yaml

@@ -3,10 +3,10 @@
 # specify here default training configuration
 defaults:
  - _self_
- - datamodule: focus.yaml
- - model: focus.yaml
+ - datamodule: focus150.yaml
+ - model: focus150.yaml
  - callbacks: default.yaml
- - logger: null # set logger here or use command line (e.g. `python train.py logger=tensorboard`)
+ - logger: tensorboard # set logger here or use command line (e.g. `python train.py logger=tensorboard`)
  - trainer: long.yaml
  - log_dir: default.yaml
 

configs/trainer/default.yaml

@@ -1,6 +1,6 @@
 _target_: pytorch_lightning.Trainer
 
-gpus: 0
+gpus: 1
 
 min_epochs: 1
 max_epochs: 10

configs/trainer/long.yaml

@@ -1,6 +1,6 @@
 _target_: pytorch_lightning.Trainer
 
-gpus: 0
+gpus: 1
 
 min_epochs: 1
 max_epochs: 100

src/datamodules/focus_datamodule.py

@@ -22,6 +22,8 @@ class FocusDataSet(Dataset):
  on a sample.
  """
  self.metadata = pd.read_csv(csv_file)
+ self.col_index_path = self.metadata.columns.get_loc("image_path")
+ self.col_index_focus = self.metadata.columns.get_loc("focus_value")
  self.root_dir = root_dir
  self.transform = transform
 
@@ -45,9 +47,11 @@ class FocusDataSet(Dataset):
  if torch.is_tensor(idx):
  idx = idx.tolist()
 
- img_name = os.path.join(self.root_dir, self.metadata.iloc[idx, 1])
+ img_name = os.path.join(
+ self.root_dir, self.metadata.iloc[idx, self.col_index_path]
+ )
  image = io.imread(img_name)
- focus_value = self.metadata.iloc[idx, 5]
+ focus_value = self.metadata.iloc[idx, self.col_index_focus]
  sample = {"image": image, "focus_value": focus_value}
 
  if self.transform:

src/models/focus_module.py

@@ -134,6 +134,7 @@ class FocusLitModule(LightningModule):
  return {"loss": loss, "preds": preds, "targets": targets}
 
  def test_epoch_end(self, outputs: List[Any]):
+ print(outputs)
  pass
 
  def on_epoch_end(self):