3-3-4-TFLiteConverter.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"3_3_4_Colab.ipynb","provenance":[]},"kernelspec":{"name":"python3","display_name":"Python 3"}},"cells":[{"cell_type":"code","metadata":{"id":"D1J15Vh_1Jih","cellView":"both"},"source":["try:\n","  # %tensorflow_version only exists in Colab.\n","  %tensorflow_version 2.x\n","except Exception:\n","  pass"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"O6k2Pg0gTYB8"},"source":["import tensorflow as tf\n","import numpy as np\n","from tensorflow.keras import Sequential\n","from tensorflow.keras.layers import Dense"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"YAwdHf6ySQGt"},"source":["l0 = Dense(units=1, input_shape=[1])\n","model = Sequential([l0])\n","model.compile(optimizer='sgd', loss='mean_squared_error')\n","\n","xs = np.array([-1.0, 0.0, 1.0, 2.0, 3.0, 4.0], dtype=float)\n","ys = np.array([-3.0, -1.0, 1.0, 3.0, 5.0, 7.0], dtype=float)\n","\n","model.fit(xs, ys, epochs=500)\n","\n","print(model.predict([10.0]))\n","print(\"Here is what I learned: {}\".format(l0.get_weights()))"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"SOs_IDM6ToaM"},"source":["export_dir = 'saved_model/1'\n","tf.saved_model.save(model, export_dir)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"lWSlkprhTsWE"},"source":["# Convert the model.\n","converter = tf.lite.TFLiteConverter.from_saved_model(export_dir)\n","tflite_model = converter.convert()"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"lsaEjJfrTujk"},"source":["import pathlib\n","tflite_model_file = pathlib.Path('model.tflite')\n","tflite_model_file.write_bytes(tflite_model)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"fseX4pkhTzS0"},"source":["# Load TFLite model and allocate tensors.\n","interpreter = tf.lite.Interpreter(model_content=tflite_model)\n","interpreter.allocate_tensors()\n","\n","# Get input and output tensors.\n","input_details = interpreter.get_input_details()\n","output_details = interpreter.get_output_details()\n","print(input_details)\n","print(output_details)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"k_Ij8_BvU0KV"},"source":["to_predict = np.array([[10.0]], dtype=np.float32)\n","print(to_predict)\n","interpreter.set_tensor(input_details[0]['index'], to_predict)\n","interpreter.invoke()\n","tflite_results = interpreter.get_tensor(output_details[0]['index'])\n","print(tflite_results)"],"execution_count":null,"outputs":[]}]}