yume before training on gpu

.github/workflows/hugging-face.yaml

@@ -0,0 +1,19 @@
+name: Unit Tests
+
+on:
+  push:
+    branches:
+      - "main"
+# TODO setup for huggingface space
+jobs:
+  test:
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v2
+
+      - name: Setup Python
+        uses: actions/setup-python@v2
+        with:
+          python-version: 3.x

interface.py

@@ -3,10 +3,11 @@ import torch
 from torchvision.utils import make_grid
 from torchvision.transforms import ToPILImage
 
+
 def main():
     st.title("Image Generation")
     st.write("Made with GANS from scratch")
-    
-    
-if __name__ == '__main__':
+
+
+if __name__ == "__main__":
     main()

sampler.py

@@ -1,7 +1,7 @@
-from yume import Yume
+from .yume import Yume
 
 yume = Yume()
 
 yume.load_pretrained()
 
-yume.generate()
+yume.generate()

tests/test_datasets.py

@@ -1,7 +1,13 @@
 import unittest
 
+
 class TestDatasets(unittest.TestCase):
-    pass
+    def test_download():
+        pass
+
+    def test_encode():
+        pass
+
 
-if __name__ == '__main__':
-    unittest.main()
+if __name__ == "__main__":
+    unittest.main()

tests/test_demo.py

@@ -0,0 +1,19 @@
+# test_demo.py
+import unittest
+
+
+class TestDemoFunctions(unittest.TestCase):
+
+    def test_add_numbers(self):
+        self.assertEqual(add_numbers(2, 3), 5)
+        self.assertEqual(add_numbers(-1, 1), 0)
+        # Add more test cases as needed
+
+    def test_multiply_numbers(self):
+        self.assertEqual(multiply_numbers(2, 3), 6)
+        self.assertEqual(multiply_numbers(-1, 1), -1)
+        # Add more test cases as needed
+
+
+if __name__ == "__main__":
+    unittest.main()

tests/test_pretrained.py

@@ -1,11 +1,13 @@
 import unittest
 
+
 class TestPretrained(unittest.TestCase):
     def test_download():
         pass
-    
+
     def test_generation():
         pass
 
-if __name__ == '__main__':
-    unittest.main()
+
+if __name__ == "__main__":
+    unittest.main()

tests/test_tokenizer.py

@@ -1,15 +1,16 @@
 import unittest
 
+
 class TestTokenizer(unittest.TestCase):
     def test_encode():
         pass
-    
+
     def test_decode():
         pass
-    
+
     def test_equal_result():
         pass
-    
-    
-if __name__ == '__main__':
-    unittest.main()
+
+
+if __name__ == "__main__":
+    unittest.main()

tests/test_transformer.py

@@ -1,12 +1,13 @@
 import unittest
 
+
 class TestTransformer(unittest.TestCase):
     def test_loading_model():
         pass
-    
+
     def test_parameters():
         pass
-    
 
-if __name__ == '__main__':
-    unittest.main()
+
+if __name__ == "__main__":
+    unittest.main()

training.py

@@ -1,5 +1,5 @@
-from yume import Yume
+from .yume import Yume
 
 yume = Yume()
 
-yume.train()
+yume.train()

version.py

@@ -1 +1 @@
-__version__ = "20231117"
+__version__ = "20231117"

yume/config.py

@@ -1,23 +1,20 @@
-
 class Config:
-    def __init__(self,num_epoch:int,
-                     block_sized = 1024,
-    vocab_sized = 50304,
-    n_layer = 12,
-    n_head = 12,
-    n_embd = 768,
-    dropout = 0.0,
-    bias = True,
-                 
-                 ) -> None:
+    def __init__(
+        self,
+        num_epoch: int,
+        block_sized=1024,
+        vocab_sized=50304,
+        n_layer=12,
+        n_head=12,
+        n_embd=768,
+        dropout=0.0,
+        bias=True,
+    ) -> None:
         self.num_epoch = num_epoch
         self.block_sized = 1024
-        self.vocab_sized = 50304 
+        self.vocab_sized = 50304
         self.n_layerd = 12
         self.n_headd = 12
         self.n_embdd = 768
         self.dropout = 0.0
         self.bias = True
-        
-
-config = Config()

yume/dataset.py

@@ -1,5 +1,11 @@
 from datasets import load_dataset
 
+
+# TODO setup dataset
 class Trainset:
     def __init__(self) -> None:
         pass
+
+
+def use_dataset(name=""):
+    pass

yume/models.py

@@ -1,34 +1,49 @@
 import torch
 from torch import nn
 import torch.nn.functional as F
+from .config import Config
+from .utils import encode, decode
 
+
+# TODO setup models
 class SelfAttention(nn.Module):
-    def __init__(self, ) -> None:
+    def __init__(self, config: Config) -> None:
         super().__init__()
-        
-    def forward(self):
+
+    def forward(self, x):
         pass
 
+
 class MLP(nn.Module):
-    def __init__(self, ) -> None:
+    def __init__(self, config: Config) -> None:
         super().__init__()
-        
-    def forward(self):
+
+    def forward(self, x):
         pass
 
+
 class Block(nn.Module):
-    def __init__(self, ) -> None:
+    def __init__(self, config: Config) -> None:
         super().__init__()
-        
-    def forward(self):
+
+    def forward(self, x):
         pass
-    
+
 
 class GPT(nn.Module):
-    def __init__(self, config):
+    def __init__(self, config: Config):
         super().__init__()
 
-        self.device = 'cuda' if torch.cuda.is_available() else 'cpu'
-    
-    def forward(x):
-        pass
+        self.device = "cuda" if torch.cuda.is_available() else "cpu"
+
+    def forward(self, x):
+        pass
+
+    def _init_weights(self):
+        pass
+
+    def configure_optimizer(self):
+        pass
+
+    def generate(self):
+        pass

yume/tokenizer.py

@@ -1,5 +1,15 @@
+# TODO setup tokenizer
 class Tokenizer:
     def __init__(self):
         pass
-    
-    
+
+    def train(self):
+        pass
+
+    def encode(self, texts):
+        result = ""
+        return []
+
+    def decode(self, tensors):
+        result = ""
+        return tensors

yume/utils.py

@@ -1,8 +1,18 @@
+from .tokenizer import Tokenizer
+
+
+# TODO setup utils
 def encode(text):
-    pass
+    tokenizer = Tokenizer()
+    result = tokenizer.encode(text)
+    return result
+
+
+def decode(tensor):
+    tokenizer = Tokenizer()
+    result = tokenizer.decode(tensor)
+    return result
 
-def decode(text):
-    pass
 
-def transform_data(data):
+def load_data(data):
     pass

yume/yume.py

@@ -6,6 +6,7 @@ from .models import GPT
 
 config = Config()
 
+
 class Yume:
     def __init__(self, config):
         super().__init__()
@@ -18,4 +19,4 @@ class Yume:
         pass
 
     def generate(self):
-        pass
+        pass