experiments/STT
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Mention use of CuDNN RNN in release checkpoints

Browse Source 
X-DeepSpeech: NOBUILD
This commit is contained in:
Reuben Morais 2020-01-18 18:07:55 +00:00 committed by GitHub
parent 3cea430f7e
commit 94882fb1c9
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
doc/TRAINING.rst
View File

@ -207,7 +207,11 @@ For example, if you want to fine tune the entire graph using your own data in ``
mkdir fine_tuning_checkpoints
python3 DeepSpeech.py --n_hidden 2048 --checkpoint_dir path/to/checkpoint/folder --epochs 3 --train_files my-train.csv --dev_files my-dev.csv --test_files my_dev.csv --learning_rate 0.0001
Note: the released models were trained with ``--n_hidden 2048``\ , so you need to use that same value when initializing from the release models.
Note: the released models were trained with ``--n_hidden 2048``\ , so you need to use that same value when initializing from the release models. Since v0.6.0, the release models are also trained with ``--use_cudnn_rnn``\ , so you'll need to specify that as well. If you don't have a CUDA compatible GPU, then you can workaround it by using the ``--cudnn_checkpoint`` flag. Use ``--helpfull`` to get more information on how the flags work. If you try to load a release model without following these steps, you'll get an error similar to this:
.. code-block::
Key cudnn_lstm/rnn/multi_rnn_cell/cell_0/cudnn_compatible_lstm_cell/bias/Adam not found in checkpoint
Training with augmentation
^^^^^^^^^^^^^^^^^^^^^^^^^^