mobilenet_v3. Update pb_converted

Author: dmitrygolovkin

caffe_models/mobilenet/caffe_model/mobilenet_v3/README.md

@@ -5,12 +5,12 @@ MobileNet v3 model could be converted from frozen pb or tflite to caffe model.
 
 Conversion command:
 1. Model converted from TF: pb_converted/v3-small_224_0.75_float_pb_convert.prototxt
-> evconvert tf2ev -g v3-small_224_0.75_float.pb -p input 1 224 224 3
+> evconvert tf2ev -g v3-small_224_0.75_float.pb -p input 1 224 224 3 
 
 2. (Old) model converted from tflite: v3-small_224_0.75_float_convert.prototxt
-> evconvert tflite2ev -m v3-small_224_0.75_float.tflite
+> evconvert tflite2ev -m v3-small_224_0.75_float.tflite 
 This model contains Permute layers that should have been optimized away during the conversion.
-The model converted from TF does not have these layers
+The model converted from TF does not have these layers.
 
 The *_nohardsiwsh.prototxt models are generated manually by changing all HardSwish layers into ReLU layers
 

caffe_models/mobilenet/caffe_model/mobilenet_v3/pb_converted/v3-small_224_0.75_float_pb_convert.caffemodel

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5886403309fe4474752368d01824500f5bf5184f50a3288397f17470b1d3d64d
-size 8313163
+oid sha256:9b86d4f5583a5a90e318e72fb8813a1a145091d34590bf4673e005f6c7def475
+size 8229168