Image classification model development: updated pretrained models to larger newly-trained model, fixed typo in model_converter.py which sliced characters from the model name that would fail convert_to_tflite.sh, removed old MD5 models

Author: obaileyw-uncc

benchmark/training/image_classification/model_converter.py

@@ -15,7 +15,7 @@
 tfmodel_path = 'trained_models/' + model_name
 tfmodel = tf.keras.models.load_model(tfmodel_path)
 cifar_10_dir = 'cifar-10-batches-py'
-model_name = model_name[:-3]
+# model_name = model_name[:-3]
 
 def representative_dataset_generator():
     train_data, train_filenames, train_labels, test_data, test_filenames, test_labels, label_names = \