Hello,i am use colab to train. >100h
1. !python3 /content/DeepSpeech/[DeepSpeech.py](http://deepspeech.py/) --n_hidden 1600 --max_to_keep 8 --early_stop True --es_epochs 25 --dev_batch_size 10 --train_batch_size 16 --feature_win_step 25 --summary_dir /content/summary --train_cudnn True --checkpoint_dir /content/1600 --epochs 25 --train_files /content/vivos/train2222.csv --dev_files /content/vivos/test22.csv --learning_rate 0.000001 --export_tflite --export_dir /content/DeepSpeech/output_models/ --automatic_mixed_precision True --dropout_rate 0.3 --alphabet_config_path /content/vivos/alphabet2.txt
I save the checkpoint after each time runs out and train again. after epochs I get the results:
Epoch 0 | Training | Elapsed Time: 0:15:47 | Steps: 4287 | Loss: 39.171857
Epoch 0 | Validation | Elapsed Time: 0:01:28 | Steps: 228 | Loss: 192.677394 | Dataset: /content/vivos/test22.csv
I Saved new best validating model with loss 192.677394 to: /content/1600/best_dev-316469
Epoch 1 | Training | Elapsed Time: 0:15:41 | Steps: 4287 | Loss: 39.697071
Epoch 1 | Validation | Elapsed Time: 0:01:28 | Steps: 228 | Loss: 193.325539 | Dataset: /content/vivos/test22.csv
Can someone show me some problem causing this problem. Thank you so much.
