{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "c8Cx-rUMVX25"
},
"source": [
"##### Copyright 2019 The TensorFlow Authors."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"cellView": "form",
"execution": {
"iopub.execute_input": "2022-12-15T00:55:58.585786Z",
"iopub.status.busy": "2022-12-15T00:55:58.585221Z",
"iopub.status.idle": "2022-12-15T00:55:58.589494Z",
"shell.execute_reply": "2022-12-15T00:55:58.588932Z"
},
"id": "I9sUhVL_VZNO"
},
"outputs": [],
"source": [
"#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n",
"# you may not use this file except in compliance with the License.\n",
"# You may obtain a copy of the License at\n",
"#\n",
"# https://www.apache.org/licenses/LICENSE-2.0\n",
"#\n",
"# Unless required by applicable law or agreed to in writing, software\n",
"# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
"# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
"# See the License for the specific language governing permissions and\n",
"# limitations under the License."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "6Y8E0lw5eYWm"
},
"source": [
"# 훈련 후 float16 양자화"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "CGuqeuPSVNo-"
},
"source": [
""
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "BTC1rDAuei_1"
},
"source": [
"## 개요\n",
"\n",
"[TensorFlow Lite는](https://www.tensorflow.org/lite/) 이제 TensorFlow에서 TensorFlow Lite의 flat buffer 형식으로 모델을 변환하는 동안 가중치를 16bit 부동 소수점 값으로 변환하는 것을 지원합니다. 그 결과 모델 크기가 2배 감소합니다. GPU와 같은 일부 하드웨어는 감소한 정밀도 산술로 기본적으로 계산할 수 있으므로 기존 부동 소수점 실행보다 속도가 향상됩니다. Tensorflow Lite GPU 대리자는 이러한 방식으로 실행되도록 구성될 수 있습니다. 그러나 float16 가중치로 변환된 모델은 추가 수정없이도 CPU에서 계속 실행될 수 있습니다. float16 가중치는 첫 번째 추론 이전에 float32로 업 샘플링됩니다. 이를 통해 지연 시간과 정확성에 미치는 영향을 최소화하는 대신 모델 크기를 크게 줄일 수 있습니다.\n",
"\n",
"이 가이드에서는 MNIST 모델을 처음부터 훈련하고 TensorFlow에서 정확성을 확인한 다음 모델을 float16 양자화를 사용하여 Tensorflow Lite flatbuffer로 변환합니다. 마지막으로 변환된 모델의 정확성을 확인하고 원래 float32 모델과 비교합니다."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "2XsEP17Zelz9"
},
"source": [
"## MNIST 모델 빌드하기"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "dDqqUIZjZjac"
},
"source": [
"### 설정"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:55:58.593264Z",
"iopub.status.busy": "2022-12-15T00:55:58.592744Z",
"iopub.status.idle": "2022-12-15T00:56:00.634708Z",
"shell.execute_reply": "2022-12-15T00:56:00.633921Z"
},
"id": "gyqAw1M9lyab"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"2022-12-15 00:55:59.580016: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory\n",
"2022-12-15 00:55:59.580120: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvinfer_plugin.so.7: cannot open shared object file: No such file or directory\n",
"2022-12-15 00:55:59.580130: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.\n"
]
}
],
"source": [
"import logging\n",
"logging.getLogger(\"tensorflow\").setLevel(logging.DEBUG)\n",
"\n",
"import tensorflow as tf\n",
"from tensorflow import keras\n",
"import numpy as np\n",
"import pathlib"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "eQ6Q0qqKZogR"
},
"source": [
"### 모델 훈련 및 내보내기"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:00.639774Z",
"iopub.status.busy": "2022-12-15T00:56:00.638819Z",
"iopub.status.idle": "2022-12-15T00:56:13.863371Z",
"shell.execute_reply": "2022-12-15T00:56:13.862615Z"
},
"id": "hWSAjQWagIHl"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r",
" 1/1875 [..............................] - ETA: 1:42:52 - loss: 2.2809 - accuracy: 0.0938"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 20/1875 [..............................] - ETA: 5s - loss: 2.1006 - accuracy: 0.3688 "
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 41/1875 [..............................] - ETA: 4s - loss: 1.8007 - accuracy: 0.5587"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 62/1875 [..............................] - ETA: 4s - loss: 1.5109 - accuracy: 0.6326"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 84/1875 [>.............................] - ETA: 4s - loss: 1.2884 - accuracy: 0.6849"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 105/1875 [>.............................] - ETA: 4s - loss: 1.1476 - accuracy: 0.7143"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 126/1875 [=>............................] - ETA: 4s - loss: 1.0405 - accuracy: 0.7383"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 147/1875 [=>............................] - ETA: 4s - loss: 0.9526 - accuracy: 0.7587"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 168/1875 [=>............................] - ETA: 4s - loss: 0.8860 - accuracy: 0.7736"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 189/1875 [==>...........................] - ETA: 4s - loss: 0.8277 - accuracy: 0.7865"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 209/1875 [==>...........................] - ETA: 4s - loss: 0.7895 - accuracy: 0.7962"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 229/1875 [==>...........................] - ETA: 4s - loss: 0.7567 - accuracy: 0.8043"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 250/1875 [===>..........................] - ETA: 4s - loss: 0.7170 - accuracy: 0.8141"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 271/1875 [===>..........................] - ETA: 3s - loss: 0.6864 - accuracy: 0.8208"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 292/1875 [===>..........................] - ETA: 3s - loss: 0.6602 - accuracy: 0.8269"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 313/1875 [====>.........................] - ETA: 3s - loss: 0.6404 - accuracy: 0.8314"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 334/1875 [====>.........................] - ETA: 3s - loss: 0.6227 - accuracy: 0.8353"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 356/1875 [====>.........................] - ETA: 3s - loss: 0.6053 - accuracy: 0.8388"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 377/1875 [=====>........................] - ETA: 3s - loss: 0.5911 - accuracy: 0.8421"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 399/1875 [=====>........................] - ETA: 3s - loss: 0.5744 - accuracy: 0.8459"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 420/1875 [=====>........................] - ETA: 3s - loss: 0.5590 - accuracy: 0.8492"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 441/1875 [======>.......................] - ETA: 3s - loss: 0.5477 - accuracy: 0.8525"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 462/1875 [======>.......................] - ETA: 3s - loss: 0.5358 - accuracy: 0.8553"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 483/1875 [======>.......................] - ETA: 3s - loss: 0.5246 - accuracy: 0.8584"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 504/1875 [=======>......................] - ETA: 3s - loss: 0.5157 - accuracy: 0.8609"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 525/1875 [=======>......................] - ETA: 3s - loss: 0.5071 - accuracy: 0.8631"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 546/1875 [=======>......................] - ETA: 3s - loss: 0.4968 - accuracy: 0.8653"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 567/1875 [========>.....................] - ETA: 3s - loss: 0.4907 - accuracy: 0.8668"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 589/1875 [========>.....................] - ETA: 3s - loss: 0.4811 - accuracy: 0.8695"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 611/1875 [========>.....................] - ETA: 3s - loss: 0.4756 - accuracy: 0.8710"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 632/1875 [=========>....................] - ETA: 3s - loss: 0.4673 - accuracy: 0.8732"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 654/1875 [=========>....................] - ETA: 2s - loss: 0.4606 - accuracy: 0.8747"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 676/1875 [=========>....................] - ETA: 2s - loss: 0.4559 - accuracy: 0.8760"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 696/1875 [==========>...................] - ETA: 2s - loss: 0.4510 - accuracy: 0.8771"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 716/1875 [==========>...................] - ETA: 2s - loss: 0.4454 - accuracy: 0.8784"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 737/1875 [==========>...................] - ETA: 2s - loss: 0.4400 - accuracy: 0.8798"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 758/1875 [===========>..................] - ETA: 2s - loss: 0.4339 - accuracy: 0.8814"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 779/1875 [===========>..................] - ETA: 2s - loss: 0.4284 - accuracy: 0.8833"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 800/1875 [===========>..................] - ETA: 2s - loss: 0.4237 - accuracy: 0.8843"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 820/1875 [============>.................] - ETA: 2s - loss: 0.4183 - accuracy: 0.8855"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 840/1875 [============>.................] - ETA: 2s - loss: 0.4156 - accuracy: 0.8864"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 860/1875 [============>.................] - ETA: 2s - loss: 0.4113 - accuracy: 0.8876"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 880/1875 [=============>................] - ETA: 2s - loss: 0.4066 - accuracy: 0.8889"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 900/1875 [=============>................] - ETA: 2s - loss: 0.4047 - accuracy: 0.8895"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 921/1875 [=============>................] - ETA: 2s - loss: 0.4003 - accuracy: 0.8907"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 942/1875 [==============>...............] - ETA: 2s - loss: 0.3958 - accuracy: 0.8918"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 963/1875 [==============>...............] - ETA: 2s - loss: 0.3912 - accuracy: 0.8930"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
" 984/1875 [==============>...............] - ETA: 2s - loss: 0.3873 - accuracy: 0.8941"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1005/1875 [===============>..............] - ETA: 2s - loss: 0.3838 - accuracy: 0.8950"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1025/1875 [===============>..............] - ETA: 2s - loss: 0.3799 - accuracy: 0.8962"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1046/1875 [===============>..............] - ETA: 2s - loss: 0.3774 - accuracy: 0.8969"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1067/1875 [================>.............] - ETA: 1s - loss: 0.3742 - accuracy: 0.8978"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1088/1875 [================>.............] - ETA: 1s - loss: 0.3706 - accuracy: 0.8987"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1109/1875 [================>.............] - ETA: 1s - loss: 0.3664 - accuracy: 0.8998"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1130/1875 [=================>............] - ETA: 1s - loss: 0.3628 - accuracy: 0.9006"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1151/1875 [=================>............] - ETA: 1s - loss: 0.3602 - accuracy: 0.9012"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1172/1875 [=================>............] - ETA: 1s - loss: 0.3571 - accuracy: 0.9020"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1193/1875 [==================>...........] - ETA: 1s - loss: 0.3545 - accuracy: 0.9028"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1214/1875 [==================>...........] - ETA: 1s - loss: 0.3520 - accuracy: 0.9035"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1235/1875 [==================>...........] - ETA: 1s - loss: 0.3484 - accuracy: 0.9044"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1256/1875 [===================>..........] - ETA: 1s - loss: 0.3468 - accuracy: 0.9050"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1276/1875 [===================>..........] - ETA: 1s - loss: 0.3441 - accuracy: 0.9057"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1296/1875 [===================>..........] - ETA: 1s - loss: 0.3427 - accuracy: 0.9061"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1317/1875 [====================>.........] - ETA: 1s - loss: 0.3397 - accuracy: 0.9068"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1338/1875 [====================>.........] - ETA: 1s - loss: 0.3372 - accuracy: 0.9076"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1359/1875 [====================>.........] - ETA: 1s - loss: 0.3347 - accuracy: 0.9083"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1379/1875 [=====================>........] - ETA: 1s - loss: 0.3325 - accuracy: 0.9087"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1400/1875 [=====================>........] - ETA: 1s - loss: 0.3304 - accuracy: 0.9093"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1421/1875 [=====================>........] - ETA: 1s - loss: 0.3283 - accuracy: 0.9097"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1443/1875 [======================>.......] - ETA: 1s - loss: 0.3259 - accuracy: 0.9101"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1464/1875 [======================>.......] - ETA: 1s - loss: 0.3244 - accuracy: 0.9105"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1485/1875 [======================>.......] - ETA: 0s - loss: 0.3222 - accuracy: 0.9110"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1507/1875 [=======================>......] - ETA: 0s - loss: 0.3202 - accuracy: 0.9117"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1529/1875 [=======================>......] - ETA: 0s - loss: 0.3178 - accuracy: 0.9123"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1551/1875 [=======================>......] - ETA: 0s - loss: 0.3157 - accuracy: 0.9128"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1573/1875 [========================>.....] - ETA: 0s - loss: 0.3139 - accuracy: 0.9133"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1595/1875 [========================>.....] - ETA: 0s - loss: 0.3120 - accuracy: 0.9138"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1617/1875 [========================>.....] - ETA: 0s - loss: 0.3100 - accuracy: 0.9143"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1638/1875 [=========================>....] - ETA: 0s - loss: 0.3082 - accuracy: 0.9148"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1659/1875 [=========================>....] - ETA: 0s - loss: 0.3062 - accuracy: 0.9153"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1681/1875 [=========================>....] - ETA: 0s - loss: 0.3045 - accuracy: 0.9159"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1702/1875 [==========================>...] - ETA: 0s - loss: 0.3034 - accuracy: 0.9161"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1723/1875 [==========================>...] - ETA: 0s - loss: 0.3015 - accuracy: 0.9166"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1744/1875 [==========================>...] - ETA: 0s - loss: 0.2996 - accuracy: 0.9171"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1765/1875 [===========================>..] - ETA: 0s - loss: 0.2979 - accuracy: 0.9175"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1785/1875 [===========================>..] - ETA: 0s - loss: 0.2961 - accuracy: 0.9179"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1806/1875 [===========================>..] - ETA: 0s - loss: 0.2944 - accuracy: 0.9185"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1827/1875 [============================>.] - ETA: 0s - loss: 0.2928 - accuracy: 0.9190"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1847/1875 [============================>.] - ETA: 0s - loss: 0.2918 - accuracy: 0.9192"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1867/1875 [============================>.] - ETA: 0s - loss: 0.2905 - accuracy: 0.9195"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
"1875/1875 [==============================] - 9s 3ms/step - loss: 0.2901 - accuracy: 0.9197 - val_loss: 0.1384 - val_accuracy: 0.9614\n"
]
},
{
"data": {
"text/plain": [
""
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Load MNIST dataset\n",
"mnist = keras.datasets.mnist\n",
"(train_images, train_labels), (test_images, test_labels) = mnist.load_data()\n",
"\n",
"# Normalize the input image so that each pixel value is between 0 to 1.\n",
"train_images = train_images / 255.0\n",
"test_images = test_images / 255.0\n",
"\n",
"# Define the model architecture\n",
"model = keras.Sequential([\n",
" keras.layers.InputLayer(input_shape=(28, 28)),\n",
" keras.layers.Reshape(target_shape=(28, 28, 1)),\n",
" keras.layers.Conv2D(filters=12, kernel_size=(3, 3), activation=tf.nn.relu),\n",
" keras.layers.MaxPooling2D(pool_size=(2, 2)),\n",
" keras.layers.Flatten(),\n",
" keras.layers.Dense(10)\n",
"])\n",
"\n",
"# Train the digit classification model\n",
"model.compile(optimizer='adam',\n",
" loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
" metrics=['accuracy'])\n",
"model.fit(\n",
" train_images,\n",
" train_labels,\n",
" epochs=1,\n",
" validation_data=(test_images, test_labels)\n",
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "5NMaNZQCkW9X"
},
"source": [
"예를 들어, 단일 epoch에 대해서만 모델을 훈련시켰으므로 ~96% 정확성으로만 훈련합니다."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "xl8_fzVAZwOh"
},
"source": [
"### TensorFlow Lite 모델로 변환하기\n",
"\n",
"TensorFlow Lite [Converter](https://www.tensorflow.org/lite/models/convert)를 사용하여 이제 훈련된 모델을 TensorFlow Lite 모델로 변환할 수 있습니다.\n",
"\n",
"`TFLiteConverter`를 사용하여 모델을 로드합니다."
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:13.867366Z",
"iopub.status.busy": "2022-12-15T00:56:13.866864Z",
"iopub.status.idle": "2022-12-15T00:56:14.972324Z",
"shell.execute_reply": "2022-12-15T00:56:14.971552Z"
},
"id": "_i8B2nDZmAgQ"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp6lyb31lw/assets\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp6lyb31lw/assets\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"2022-12-15 00:56:14.848472: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:362] Ignored output_format.\n",
"2022-12-15 00:56:14.848515: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:365] Ignored drop_control_dependency.\n"
]
}
],
"source": [
"converter = tf.lite.TFLiteConverter.from_keras_model(model)\n",
"tflite_model = converter.convert()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "F2o2ZfF0aiCx"
},
"source": [
"`.tflite` 파일에 작성합니다."
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:14.976574Z",
"iopub.status.busy": "2022-12-15T00:56:14.975910Z",
"iopub.status.idle": "2022-12-15T00:56:14.979696Z",
"shell.execute_reply": "2022-12-15T00:56:14.979041Z"
},
"id": "vptWZq2xnclo"
},
"outputs": [],
"source": [
"tflite_models_dir = pathlib.Path(\"/tmp/mnist_tflite_models/\")\n",
"tflite_models_dir.mkdir(exist_ok=True, parents=True)"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:14.982567Z",
"iopub.status.busy": "2022-12-15T00:56:14.982325Z",
"iopub.status.idle": "2022-12-15T00:56:14.986876Z",
"shell.execute_reply": "2022-12-15T00:56:14.986282Z"
},
"id": "Ie9pQaQrn5ue"
},
"outputs": [
{
"data": {
"text/plain": [
"84820"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tflite_model_file = tflite_models_dir/\"mnist_model.tflite\"\n",
"tflite_model_file.write_bytes(tflite_model)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "7BONhYtYocQY"
},
"source": [
"대신 모델을 내보낼 때 float16으로 양자화하려면 먼저 기본 최적화를 사용하도록 `optimizations` 플래그를 지정합니다. 그런 다음 float16이 대상 플랫폼에서 지원되는 유형임을 지정합니다."
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:14.990125Z",
"iopub.status.busy": "2022-12-15T00:56:14.989544Z",
"iopub.status.idle": "2022-12-15T00:56:14.992716Z",
"shell.execute_reply": "2022-12-15T00:56:14.992143Z"
},
"id": "HEZ6ET1AHAS3"
},
"outputs": [],
"source": [
"converter.optimizations = [tf.lite.Optimize.DEFAULT]\n",
"converter.target_spec.supported_types = [tf.float16]"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "xW84iMYjHd9t"
},
"source": [
"마지막으로 평소와 같이 모델을 변환합니다. 기본적으로 변환된 모델은 호출 편의를 위해 여전히 float 입력 및 출력을 사용합니다."
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:14.996064Z",
"iopub.status.busy": "2022-12-15T00:56:14.995542Z",
"iopub.status.idle": "2022-12-15T00:56:15.955497Z",
"shell.execute_reply": "2022-12-15T00:56:15.954794Z"
},
"id": "yuNfl3CoHNK3"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"INFO:tensorflow:Assets written to: /tmpfs/tmp/tmpb6hmdxdl/assets\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"INFO:tensorflow:Assets written to: /tmpfs/tmp/tmpb6hmdxdl/assets\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"2022-12-15 00:56:15.826625: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:362] Ignored output_format.\n",
"2022-12-15 00:56:15.826673: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:365] Ignored drop_control_dependency.\n"
]
},
{
"data": {
"text/plain": [
"44624"
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tflite_fp16_model = converter.convert()\n",
"tflite_model_fp16_file = tflite_models_dir/\"mnist_model_quant_f16.tflite\"\n",
"tflite_model_fp16_file.write_bytes(tflite_fp16_model)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "PhMmUTl4sbkz"
},
"source": [
"결과 파일이 약 `1/2` 크기인지 확인하세요."
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:15.959373Z",
"iopub.status.busy": "2022-12-15T00:56:15.958831Z",
"iopub.status.idle": "2022-12-15T00:56:16.160415Z",
"shell.execute_reply": "2022-12-15T00:56:16.159397Z"
},
"id": "JExfcfLDscu4"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"total 128K\r\n",
"-rw-rw-r-- 1 kbuilder kbuilder 83K Dec 15 00:56 mnist_model.tflite\r\n",
"-rw-rw-r-- 1 kbuilder kbuilder 44K Dec 15 00:56 mnist_model_quant_f16.tflite\r\n"
]
}
],
"source": [
"!ls -lh {tflite_models_dir}"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "L8lQHMp_asCq"
},
"source": [
"## TensorFlow Lite 모델 실행하기"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "-5l6-ciItvX6"
},
"source": [
"Python TensorFlow Lite 인터프리터를 사용하여 TensorFlow Lite 모델을 실행합니다."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "Ap_jE7QRvhPf"
},
"source": [
"### 인터프리터에 모델 로드하기"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:16.165472Z",
"iopub.status.busy": "2022-12-15T00:56:16.164739Z",
"iopub.status.idle": "2022-12-15T00:56:16.171392Z",
"shell.execute_reply": "2022-12-15T00:56:16.170726Z"
},
"id": "Jn16Rc23zTss"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"INFO: Created TensorFlow Lite XNNPACK delegate for CPU.\n"
]
}
],
"source": [
"interpreter = tf.lite.Interpreter(model_path=str(tflite_model_file))\n",
"interpreter.allocate_tensors()"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:16.174734Z",
"iopub.status.busy": "2022-12-15T00:56:16.174248Z",
"iopub.status.idle": "2022-12-15T00:56:16.178057Z",
"shell.execute_reply": "2022-12-15T00:56:16.177408Z"
},
"id": "J8Pztk1mvNVL"
},
"outputs": [],
"source": [
"interpreter_fp16 = tf.lite.Interpreter(model_path=str(tflite_model_fp16_file))\n",
"interpreter_fp16.allocate_tensors()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "2opUt_JTdyEu"
},
"source": [
"### 하나의 이미지에서 모델 테스트하기"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:16.181411Z",
"iopub.status.busy": "2022-12-15T00:56:16.180941Z",
"iopub.status.idle": "2022-12-15T00:56:16.185958Z",
"shell.execute_reply": "2022-12-15T00:56:16.185325Z"
},
"id": "AKslvo2kwWac"
},
"outputs": [],
"source": [
"test_image = np.expand_dims(test_images[0], axis=0).astype(np.float32)\n",
"\n",
"input_index = interpreter.get_input_details()[0][\"index\"]\n",
"output_index = interpreter.get_output_details()[0][\"index\"]\n",
"\n",
"interpreter.set_tensor(input_index, test_image)\n",
"interpreter.invoke()\n",
"predictions = interpreter.get_tensor(output_index)"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:16.189170Z",
"iopub.status.busy": "2022-12-15T00:56:16.188683Z",
"iopub.status.idle": "2022-12-15T00:56:16.939117Z",
"shell.execute_reply": "2022-12-15T00:56:16.938229Z"
},
"id": "XZClM2vo3_bm"
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"import matplotlib.pylab as plt\n",
"\n",
"plt.imshow(test_images[0])\n",
"template = \"True:{true}, predicted:{predict}\"\n",
"_ = plt.title(template.format(true= str(test_labels[0]),\n",
" predict=str(np.argmax(predictions[0]))))\n",
"plt.grid(False)"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:16.942964Z",
"iopub.status.busy": "2022-12-15T00:56:16.942300Z",
"iopub.status.idle": "2022-12-15T00:56:16.947667Z",
"shell.execute_reply": "2022-12-15T00:56:16.946931Z"
},
"id": "3gwhv4lKbYZ4"
},
"outputs": [],
"source": [
"test_image = np.expand_dims(test_images[0], axis=0).astype(np.float32)\n",
"\n",
"input_index = interpreter_fp16.get_input_details()[0][\"index\"]\n",
"output_index = interpreter_fp16.get_output_details()[0][\"index\"]\n",
"\n",
"interpreter_fp16.set_tensor(input_index, test_image)\n",
"interpreter_fp16.invoke()\n",
"predictions = interpreter_fp16.get_tensor(output_index)"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:16.951272Z",
"iopub.status.busy": "2022-12-15T00:56:16.950652Z",
"iopub.status.idle": "2022-12-15T00:56:17.102693Z",
"shell.execute_reply": "2022-12-15T00:56:17.101986Z"
},
"id": "CIH7G_MwbY2x"
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"plt.imshow(test_images[0])\n",
"template = \"True:{true}, predicted:{predict}\"\n",
"_ = plt.title(template.format(true= str(test_labels[0]),\n",
" predict=str(np.argmax(predictions[0]))))\n",
"plt.grid(False)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "LwN7uIdCd8Gw"
},
"source": [
"### 모델 평가하기"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:17.106326Z",
"iopub.status.busy": "2022-12-15T00:56:17.106055Z",
"iopub.status.idle": "2022-12-15T00:56:17.111775Z",
"shell.execute_reply": "2022-12-15T00:56:17.111174Z"
},
"id": "05aeAuWjvjPx"
},
"outputs": [],
"source": [
"# A helper function to evaluate the TF Lite model using \"test\" dataset.\n",
"def evaluate_model(interpreter):\n",
" input_index = interpreter.get_input_details()[0][\"index\"]\n",
" output_index = interpreter.get_output_details()[0][\"index\"]\n",
"\n",
" # Run predictions on every image in the \"test\" dataset.\n",
" prediction_digits = []\n",
" for test_image in test_images:\n",
" # Pre-processing: add batch dimension and convert to float32 to match with\n",
" # the model's input data format.\n",
" test_image = np.expand_dims(test_image, axis=0).astype(np.float32)\n",
" interpreter.set_tensor(input_index, test_image)\n",
"\n",
" # Run inference.\n",
" interpreter.invoke()\n",
"\n",
" # Post-processing: remove batch dimension and find the digit with highest\n",
" # probability.\n",
" output = interpreter.tensor(output_index)\n",
" digit = np.argmax(output()[0])\n",
" prediction_digits.append(digit)\n",
"\n",
" # Compare prediction results with ground truth labels to calculate accuracy.\n",
" accurate_count = 0\n",
" for index in range(len(prediction_digits)):\n",
" if prediction_digits[index] == test_labels[index]:\n",
" accurate_count += 1\n",
" accuracy = accurate_count * 1.0 / len(prediction_digits)\n",
"\n",
" return accuracy"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:17.115170Z",
"iopub.status.busy": "2022-12-15T00:56:17.114619Z",
"iopub.status.idle": "2022-12-15T00:56:18.409912Z",
"shell.execute_reply": "2022-12-15T00:56:18.409179Z"
},
"id": "T5mWkSbMcU5z"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"0.9614\n"
]
}
],
"source": [
"print(evaluate_model(interpreter))"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "Km3cY9ry8ZlG"
},
"source": [
"float16 양자화된 모델에 대한 평가를 반복하여 다음을 얻습니다."
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {
"execution": {
"iopub.execute_input": "2022-12-15T00:56:18.413565Z",
"iopub.status.busy": "2022-12-15T00:56:18.413178Z",
"iopub.status.idle": "2022-12-15T00:56:19.730797Z",
"shell.execute_reply": "2022-12-15T00:56:19.730089Z"
},
"id": "-9cnwiPp6EGm"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"0.9614\n"
]
}
],
"source": [
"# NOTE: Colab runs on server CPUs. At the time of writing this, TensorFlow Lite\n",
"# doesn't have super optimized server CPU kernels. For this reason this may be\n",
"# slower than the above float interpreter. But for mobile CPUs, considerable\n",
"# speedup can be observed.\n",
"print(evaluate_model(interpreter_fp16))"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "L7lfxkor8pgv"
},
"source": [
"이 예에서는 정확성 차이가 없이 모델을 float16으로 양자화했습니다.\n",
"\n",
"GPU에서 fp16 양자화 모델을 평가하는 것도 가능합니다. 감소된 정밀도 값으로 모든 산술을 수행하려면 다음과 같이 앱에서 `TfLiteGPUDelegateOptions` 구조체를 만들고 `precision_loss_allowed`를 `1`로 설정해야 합니다.\n",
"\n",
"```\n",
"//Prepare GPU delegate.\n",
"const TfLiteGpuDelegateOptions options = {\n",
" .metadata = NULL,\n",
" .compile_options = {\n",
" .precision_loss_allowed = 1, // FP16\n",
" .preferred_gl_object_type = TFLITE_GL_OBJECT_TYPE_FASTEST,\n",
" .dynamic_batch_enabled = 0, // Not fully functional yet\n",
" },\n",
"};\n",
"```\n",
"\n",
"TFLite GPU 대리자 및 애플리케이션에서 사용하는 방법에 대한 자세한 설명서는 [여기에서](https://www.tensorflow.org/lite/performance/gpu_advanced?source=post_page---------------------------) 찾을 수 있습니다"
]
}
],
"metadata": {
"colab": {
"collapsed_sections": [],
"name": "post_training_float16_quant.ipynb",
"toc_visible": true
},
"kernelspec": {
"display_name": "Python 3",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.9.16"
}
},
"nbformat": 4,
"nbformat_minor": 0
}
![](\"https://www.tensorflow.org/images/tf_logo_32px.png\"){kind=logo}
TensorFlow.org에서 보기![](\"https://www.tensorflow.org/images/colab_logo_32px.png\"){kind=logo}
Google Colab에서 실행![](\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\")
GitHub에서 소스보기![](\"https://www.tensorflow.org/images/download_logo_32px.png\"){kind=logo}
노트북 다운로드