add profiling for mnist net

test/test_conv_speed.py

@@ -14,6 +14,7 @@ import time
 import cProfile
 import pstats
 import unittest
+import numpy as np
 from tinygrad.tensor import Tensor
 
 def profile_conv(bs, chans, conv, cnt=10):
@@ -30,27 +31,55 @@ def profile_conv(bs, chans, conv, cnt=10):
     bpt += (et2-et1)
   return fpt/cnt, bpt/cnt
 
-class TestConvSpeed(unittest.TestCase):
+def start_profile():
-  def test_forward_backward_3x3(self):
+  import time
-    # warmup
-    profile_conv(128, 16, 3, cnt=1)
-
-    # profile
   pr = cProfile.Profile(timer=lambda: int(time.time()*1e9), timeunit=1e-6)
   pr.enable()
-    fpt, bpt = profile_conv(128, 16, 3)
+  return pr
+
+def stop_profile(pr, sort='cumtime'):
   pr.disable()
   ps = pstats.Stats(pr)
   ps.strip_dirs()
-    ps.sort_stats('cumtime')
+  ps.sort_stats(sort)
   ps.print_stats(0.3)
 
   if prof is not None:
     prof.print_stats()
 
+class TestConvSpeed(unittest.TestCase):
+  def test_forward_backward_3x3(self):
+    # warmup
+    profile_conv(128, 16, 3, cnt=1)
+
+    pr = start_profile()
+    fpt, bpt = profile_conv(128, 16, 3)
+    stop_profile(pr)
+
     print("forward pass:  %.3f ms" % (fpt*1000))
     print("backward pass: %.3f ms" % (bpt*1000))
 
+  def test_mnist(self):
+    # https://keras.io/examples/vision/mnist_convnet/
+    conv = 3
+    inter_chan, out_chan = 32, 64
+    c1 = Tensor.randn(inter_chan,1,conv,conv)
+    c2 = Tensor.randn(out_chan,inter_chan,conv,conv)
+    l1 = Tensor.randn(out_chan*5*5, 10)
+
+    for i in range(6):
+      x = Tensor.randn(128, 1, 28, 28)
+      x = x.conv2d(c1).relu().maxpool2x2()
+      x = x.conv2d(c2).relu().maxpool2x2()
+      x = x.reshape(Tensor(np.array((x.shape[0], -1))))
+      out = x.dot(l1).logsoftmax()
+      out.mean().backward()
+      if i == 0:
+        pr = start_profile()
+
+    stop_profile(pr, sort='time')
+
+
 if __name__ == '__main__':
   unittest.main()
 

tinygrad/utils.py

@@ -68,6 +68,9 @@ def im2col(x, H, W):
       tx[:, Y, X] = x[:, :, Y:Y+H, X:X+W].reshape(bs, -1)
   """
 
+  # all the time is spent here
+  tx = tx.ravel()
+
   return tx.reshape(-1, cin*W*H)
 
 def col2im(tx, H, W, OY, OX):