Results on TIMIT dataset for fine tuning

Deep Speech
#1

I have fine tuned TIMIT dataset (around 4.5 hours), with pretrained DeepSpeech v0.5.1 models.
Command:
python3 DeepSpeech.py --n_hidden 2048 --checkpoint_dir /home/iiit_admin/DeepSpeech-0.5.1/DeepSpeech-0.5.1/checkpoint/deepspeech-0.5.1-checkpoint/ --epochs 100 --train_files /home/iiit_admin/Desktop/TIMIT/timit_train.csv --test_files /home/iiit_admin/Desktop/TIMIT/timit_test.csv --train_batch_size 80 – test_batch_size 20 learning_rate 0.0001

I am getting WER as follows
Test on /home/iiit_admin/Desktop/TIMIT/timit_test.csv - WER: 0.960758, CER: 0.929685, loss: 141.058945-100 epochs.

  1. How to improve on this?
  2. Has anyone done fine tuning on TIMIT dataset?If yes, what WER is got?