Upload folder using huggingface_hub

MultiplicationNet.py

@@ -3,7 +3,7 @@ import torch.nn as nn
 class MultiplicationNet(nn.Module):
     def __init__(self):
         super(MultiplicationNet, self).__init__()
-        layer_sizes = [2, 256, 512, 1024, 1]
+        layer_sizes = [2, 512, 1024, 2048, 1]
         layers = []
         for i in range(len(layer_sizes) - 1):
             layers.append(nn.Linear(layer_sizes[i], layer_sizes[i + 1]))

device.py

@@ -5,3 +5,5 @@ device: str
 if torch.cuda.is_available(): device = 'cuda'
 elif torch.mps.is_available(): device = 'mps'
 else: device = 'cpu'
+
+print(f"Device set to {torch.Tensor().to(device).device}")

model.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cefcd52977cd055794c10dba186f321ad2b5aacee73975f99371d3843c0a524c
-size 2639793
+oid sha256:1e68ce613b4003c2d7fee3ed8fafc38b6b53fb744e7a50f237710d63fad450bc
+size 10517425

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ef5ab55b5daa68066891c2991cfe1628f76581bf0106deedb4b2fa907ae4dead
-size 2635388
+oid sha256:99a3bb67679f9a5ee8ea27df6fed196e33d35fb20bf73ca7192a03f0ff16017c
+size 10513028

test.py

@@ -0,0 +1,20 @@
+from run import run
+from tqdm import trange
+import random
+
+lacc = []
+
+def test(round):
+    pbar = trange(round)
+    for r in pbar:
+        x = random.randint(1, 100)
+        y = random.randint(1, 100)
+        pans = run([x, y])
+        tans = x * y
+        acc = 1 - abs(tans - pans) / tans
+        lacc.append(acc)
+        pbar.set_description_str(f"Round {r}, Average Accuracy = {sum(lacc) / len(lacc) * 100}%")
+    print(f"Final Accuracy: {sum(lacc) / len(lacc) * 100}%")
+
+if __name__ == '__main__':
+    test(100000)

train.py

@@ -12,8 +12,8 @@ def generate_data(num_samples, min_val=0, max_val=100):
     return np.hstack([x1, x2]), y
 
 def train():
-    num_samples = 20000
-    num_epochs = 20000
+    num_samples = 10000
+    num_epochs = 30000
     learning_rate = 0.01
 
     x, y = generate_data(num_samples)
@@ -23,7 +23,7 @@ def train():
     model = MultiplicationNet().to(device)
     criterion = nn.MSELoss().to(device)
     optimizer = optim.Adam(model.parameters(), lr=learning_rate)
-    scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=100, gamma=0.93)
+    scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=100, gamma=0.95)
 
     for epoch in range(num_epochs):
         outputs = model(x_train)