Use tensorflow.contrib.slim to training a simple CNN classification model
代码将更新
1.1 git clone this project
1.2 create a directory ./datasets/images/
1.3 generate training images:
python3 generate_train_data.py
1.4 generate tfrecord file:
python3 generate_tfrecord.py \
--images_path ./datasets/images/ \
--output_path ./datasets/train.record
1.5 create a directory ./training/
1.6 train CNN model:
python3 train.py \
--record_path ./datasets/train.record \
--logdir ./training/
1.7 visulize the loss curves:
tensorboard --logdir /home/.../training/
1.8 export frozen inference graph:
python3 export_inference_graph.py \
--input_type image_tensor \
--trained_checkpoint_prefix ./training/model.ckpt-xxx(num_steps) \
--output_directory path/to/exported_model_directory
1.9 evaluate the trained model:
python3 evaluate.py \
--frozen_graph_path path/to/exported_model_directory/frozen_inference_graph.pb