fix(util): ensure integer receptive field size (#1645)

pyannote/audio/utils/receptive_field.py

@@ -20,8 +20,6 @@
 # OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 # SOFTWARE.
 
-
-import math
 from typing import List
 
 
@@ -52,9 +50,7 @@ def conv1d_num_frames(
     ------
     https://pytorch.org/docs/stable/generated/torch.nn.Conv1d.html#torch.nn.Conv1d
     """
-    return math.floor(
-        1 + (num_samples + 2 * padding - dilation * (kernel_size - 1) - 1) / stride
-    )
+    return 1 + (num_samples + 2 * padding - dilation * (kernel_size - 1) - 1) // stride
 
 
 def multi_conv_num_frames(
@@ -105,6 +101,7 @@ def multi_conv_receptive_field_size(
     dilation: List[int] = None,
 ) -> int:
     receptive_field_size = num_frames
+
     for k, s, d in reversed(list(zip(kernel_size, stride, dilation))):
         receptive_field_size = conv1d_receptive_field_size(
             num_frames=receptive_field_size,
@@ -140,7 +137,7 @@ def conv1d_receptive_field_center(
     """
 
     effective_kernel_size = 1 + (kernel_size - 1) * dilation
-    return frame * stride + (effective_kernel_size - 1) / 2 - padding
+    return frame * stride + (effective_kernel_size - 1) // 2 - padding
 
 
 def multi_conv_receptive_field_center(