a

.ipynb_checkpoints/app-checkpoint.py

@@ -23,6 +23,7 @@ class ImageClassificationCollator:
             encodings = self.feature_extractor([x[0] for x in batch], return_tensors='pt')
             encodings['labels'] = torch.tensor([x[1] for x in batch], dtype=torch.long)
             return encodings 
+        
 class Classifier(pl.LightningModule):
 
         def __init__(self, model, lr: float = 2e-5, **kwargs):
@@ -54,10 +55,7 @@ def video_identity(video,user_name,class_name,trainortest,ready):
     if ready=='yes':
 
         data_dir = Path(str(user_name)+'/train')
-        transform = transforms.Compose([
-            transforms.ToTensor(),
-            transforms.ConvertImageDtype(torch.float)
-            ]) 
+
         train_ds = ImageFolder(data_dir, transform=transform)
 
         

.ipynb_checkpoints/requirements-checkpoint.txt

@@ -2,7 +2,7 @@ opencv-python
 encoded-video
 torch==2.0.0
 numpy
-pytorch-lightning
-torchvision
-transformers
+pytorch-lightning==2.0.1
+torchvision==0.15.1
+transformers==4.29.0
 pathlib

app.py

@@ -23,6 +23,7 @@ class ImageClassificationCollator:
             encodings = self.feature_extractor([x[0] for x in batch], return_tensors='pt')
             encodings['labels'] = torch.tensor([x[1] for x in batch], dtype=torch.long)
             return encodings 
+        
 class Classifier(pl.LightningModule):
 
         def __init__(self, model, lr: float = 2e-5, **kwargs):
@@ -54,10 +55,7 @@ def video_identity(video,user_name,class_name,trainortest,ready):
     if ready=='yes':
 
         data_dir = Path(str(user_name)+'/train')
-        transform = transforms.Compose([
-            transforms.ToTensor(),
-            transforms.ConvertImageDtype(torch.float)
-            ]) 
+
         train_ds = ImageFolder(data_dir, transform=transform)
 
         

requirements.txt

@@ -2,7 +2,7 @@ opencv-python
 encoded-video
 torch==2.0.0
 numpy
-pytorch-lightning
-torchvision
-transformers
+pytorch-lightning==2.0.1
+torchvision==0.15.1
+transformers==4.29.0
 pathlib