TensorFlow Speech Recognition Challenge Kaggle Competition Page
- example_conv
- cnn_trad_fpool3 [http://www.isca-speech.org/archive/interspeech_2015/papers/i15_1478.pdf]
- esc_cnn [http://karol.piczak.com/papers/Piczak2015-ESC-ConvNet.pdf]
- deep_cnn
- deeep_cnn
- deeeep_cnn
python3 train.py \
--data_url= \
--model_architecture=deeeep_cnn \
--how_many_training_steps=30000,30000,180000 \
--learning_rate=0.001,0.0001,0.001 \
--batch_size=500 \
--start_checkpoint=speech_commands_train/deeeep_cnn.ckpt-200000
Training Dataset [https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/download/train.7z]
python3 freeze.py \
--model_architecture=deeeep_cnn \
--start_checkpoint=speech_commands_train/deeeep_cnn.ckpt-240000 \
--output_file=graph.pb
python3 label_wav.py \
--graph=graph.pb \
--labels=speech_commands_train/deeeep_cnn_labels.txt \
--wav_dir=test_dataset
Testing Dataset [https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/download/test.7z]
- example_conv (77%)
- cnn_trad_fpool3 (77%)
- esc_cnn (77%)
- deep_cnn (79%)
- deeep_cnn (82%)
- deeeep_cnn (79%)