feat: optuna optimization performed

CNN-model/dataset_loaders/LoremIpsumDataset.py

@@ -17,19 +17,18 @@ class LoremIpsumDataset(Dataset):
         path = join(curdir, "data")
         self._root = path
         # Convert text to bytes (UTF-8 encoded)
-        self.dataset = torch.tensor([ord(c) for c in list(_text)], dtype=torch.long)
-
-        sequence_count = self.dataset.shape[0] // 128 # how many vectors of 128 elements can we make
-        self.dataset = self.dataset[:sequence_count * 128]
-        self.dataset = self.dataset.view(-1, 128)
-
-        print(self.dataset.shape)
+        self.dataset = torch.tensor([ord(c) % 256 for c in list(_text)], dtype=torch.long)
+        self.context_length = 128
 
     def __len__(self):
         # Number of possible sequences of length sequence_length
-        return self.dataset.size(0)
+        return self.dataset.size(0) - self.context_length
 
     def __getitem__(self, idx):
+        x = self.dataset[idx: idx + self.context_length]
+        y = self.dataset[idx + self.context_length]
+
         if self.transform is not None:
-            return self.transform(self.dataset[idx])
-        return self.dataset[idx]
+            x = self.transform(x)
+
+        return x, y