Add TFDWConv() depth_multiplier (ultralytics#7858)

Enabled grouped non c1 == c2 convolutions in TF YOLOv5 models.

Author: glenn-jocher

models/tf.py

@@ -91,9 +91,10 @@ class TFDWConv(keras.layers.Layer):
     def __init__(self, c1, c2, k=1, s=1, p=None, act=True, w=None):
         # ch_in, ch_out, weights, kernel, stride, padding, groups
         super().__init__()
-        assert c1 == c2, f'TFDWConv() input={c1} must equal output={c2} channels'
+        assert c2 % c1 == 0, f'TFDWConv() output={c2} must be a multiple of input={c1} channels'
         conv = keras.layers.DepthwiseConv2D(
             kernel_size=k,
+            depth_multiplier=c2 // c1,
             strides=s,
             padding='SAME' if s == 1 else 'VALID',
             use_bias=not hasattr(w, 'bn'),