Merge pull request #459 from Qsingle:fix_overflow_of_flops_calculate

GitOrigin-RevId: c1333e2089c5aa9379d64a3e37194e750704b02e

imperative/python/megengine/utils/module_stats.py

@@ -90,7 +90,7 @@ def flops_convNd(module: M.Conv2d, inputs, outputs):
     bias = 1 if module.bias is not None else 0
     # N x Cout x H x W x  (Cin x Kw x Kh + bias)
     return np.prod(outputs[0].shape) * (
-        module.in_channels // module.groups * np.prod(module.kernel_size) + bias
+        float(module.in_channels // module.groups) * np.prod(module.kernel_size) + bias
     )
 
 

imperative/python/megengine/utils/network_node.py

@@ -487,7 +487,7 @@ def flops_conv(opnode: ConvolutionForward, inputs, outputs):
     NCHW = np.prod(outputs[0].shape)
     bias = 1 if isinstance(opnode, ConvBiasForward) else 0
     # N x Cout x H x W x  (Cin x Kw x Kh)
-    return NCHW * (num_input * kw * kh + bias)
+    return NCHW * (float(num_input * kw * kh) + bias)
 
 
 @register_receptive_field(ConvolutionForward, ConvBiasForward)