[Conditionning] Min max scaling + [Visualization] Changing the run names

config.yml

@@ -21,6 +21,14 @@ ApproximativeMean: "torch.tensor([ 4.2457e+01,  7.4651e+00,  1.6738e+02,  1.3576
 ApproximativeSTD: "torch.tensor([5.8939e-01, 8.1625e-01, 1.4535e+02, 5.4952e+07, 1.7543e-02, 1.3846e+02,
         2.1302e-01, 1.9558e+00, 4.1455e+00, 1.2408e+01, 2.2938e-02, 9.9070e-02,
         1.9490e-01, 9.2847e-03, 2.2575e+00, 8.5310e-02, 7.8280e-02, 8.6237e-02])"
+ApproximativeMaxi: "torch.tensor([ 4.3479e+01,  9.0000e+00,  4.9267e+02,  1.4528e+09,  2.4088e+00,
+         2.7824e+03,  1.5576e+00,  6.2457e+00,  2.5120e+02,  2.7188e+02,
+         8.1683e+00,  3.2447e-01,  3.9041e+01,  2.7162e+00,  2.9419e+01,
+         8.6284e-01,  7.6471e-01, -7.7745e-02])"
+ApproximativeMini: "torch.tensor([ 4.1479e+01,  6.0000e+00,  1.0182e+00,  1.2623e+09,  2.2433e+00,
+         1.0910e+01,  1.0000e-11,  1.0000e-11, -1.1467e+01,  1.9718e+02,
+         7.9218e+00,  1.0000e-11,  3.7171e+01,  2.5584e+00,  1.2075e+01,
+        -1.2436e+00, -9.9256e-01, -8.8131e-01])"
 
 #Optimizer selection
 Optimizer: Adam # in {Adam}

dataloader.py

@@ -201,6 +201,10 @@ def get_stats_train_dataset(
     mean = 0.
     std = 0.
     nb_samples = 0.
+    it = iter(train_loader)
+    X, Y = next(it)
+    maxi = X.max(dim=0).values.max(dim=0).values
+    mini = X.min(dim=0).values.min(dim=0).values
 
     print("Computing the statistics of the train dataset")
 
@@ -209,6 +213,9 @@ def get_stats_train_dataset(
         # Update the total sum
         mean += data[0].mean(dim=0).mean(dim=0) * data[0].size(0)
         std += data[0].std(dim=0).mean(dim=0) * data[0].size(0)
+        #Update mini and maxi
+        maxi = torch.stack((data[0].max(dim=0).values.max(dim=0).values, maxi)).max(dim=0).values
+        mini = torch.stack((data[0].min(dim=0).values.min(dim=0).values, mini)).min(dim=0).values
         # Update the number of samples
         nb_samples += data[0].size(0)
 
@@ -216,7 +223,7 @@ def get_stats_train_dataset(
     mean /= nb_samples
     std /= nb_samples
 
-    return mean, std
+    return mean, std, maxi, mini
 
 def get_test_dataloader(
     filepath,
@@ -259,6 +266,9 @@ def transform_remove_space_time():
 def transform_normalize_with_train_statistics(MEAN, STD):
     return lambda attributes : torch.div(torch.sub(attributes, MEAN), STD)
 
+def transform_min_max_scaling(MIN, MAX):
+    return lambda attributes : torch.div(torch.sub(attributes, MIN), torch.sub(MAX,MIN))
+    
 def composite_transform(f,g):
     return lambda attributes : f(g(attributes))
 
@@ -292,8 +302,8 @@ if __name__ == "__main__":
         train_transform =composite_transform(transform_remove_space_time(), transform_normalize_with_train_statistics(MEAN, STD))
     )
 
-    """
-    #mean, std = get_stats_train_dataset(        
+    
+    mean, std, maxi, mini = get_stats_train_dataset(        
         filepath = trainpath,
         num_days = num_days,
         batch_size = batch_size,
@@ -303,9 +313,12 @@ if __name__ == "__main__":
         overwrite_index=True,
         max_num_samples=max_num_samples,
     )
-    """
-
+    
+    print("Mini = ")
+    print(mini)
 
+    print("Maxi =")
+    print(maxi)
 
     it = iter(train_loader)
     X, Y = next(it)

main.py

@@ -49,8 +49,10 @@ if __name__ == "__main__":
     if approx_stats:
         MEAN    = eval(cfg["ApproximativeMean"])
         STD     = eval(cfg["ApproximativeSTD"])
+        MAX    = eval(cfg["ApproximativeMaxi"])
+        MIN    = eval(cfg["ApproximativeMini"])
     else :
-        MEAN, STD = get_stats_train_dataset(trainpath,
+        MEAN, STD, MAX, MIN = get_stats_train_dataset(trainpath,
         num_days,
         batch_size,
         num_workers,
@@ -73,8 +75,8 @@ if __name__ == "__main__":
         max_num_samples=max_num_samples,
         #train_transform = dataloader.transform_remove_space_time(),
         #valid_transform = dataloader.transform_remove_space_time()
-        train_transform=dataloader.composite_transform(dataloader.transform_remove_space_time(), dataloader.transform_normalize_with_train_statistics(MEAN, STD)),
-        valid_transform=dataloader.composite_transform(dataloader.transform_remove_space_time(), dataloader.transform_normalize_with_train_statistics(MEAN, STD))
+        train_transform=dataloader.composite_transform(dataloader.transform_remove_space_time(), dataloader.transform_min_max_scaling(MIN, MAX)),
+        valid_transform=dataloader.composite_transform(dataloader.transform_remove_space_time(), dataloader.transform_min_max_scaling(MIN, MAX))
     )
 
     if use_cuda :
@@ -92,8 +94,8 @@ if __name__ == "__main__":
 
     optimizer = optimizer(cfg, network)
 
-    logdir = utils.create_unique_logpath(cfg["LogDir"], cfg["Model"]["Name"])
-    wandb.run.name = utils.generate_unique_logpath("", cfg["Model"]["Name"])
+    logdir, raw_run_name = utils.create_unique_logpath(cfg["LogDir"], cfg["Model"]["Name"])
+    wandb.run.name = raw_run_name
     network_checkpoint = model.ModelCheckpoint(logdir + "/best_model.pt", network)
 
 
@@ -114,7 +116,7 @@ if __name__ == "__main__":
         wandb.log({"val_loss": val_loss})
 
 
-    create_submission.create_submission(network, dataloader.composite_transform(dataloader.transform_remove_space_time(), dataloader.transform_normalize_with_train_statistics(MEAN, STD)), device)
+    create_submission.create_submission(network, dataloader.composite_transform(dataloader.transform_remove_space_time(), dataloader.transform_min_max_scaling(MIN, MAX)), device)
     """
     logdir = generate_unique_logpath(top_logdir, "linear")
     print("Logging to {}".format(logdir))

utils.py

@@ -14,4 +14,4 @@ def create_unique_logpath(top_logdir, raw_run_name):
         os.mkdir(top_logdir)
     logdir = generate_unique_logpath(top_logdir, raw_run_name)
     os.mkdir(logdir)
-    return logdir
+    return logdir, raw_run_name
\ No newline at end of file