That was for DS but with STT, we have access to --load_checkpoint_dir and --save_checkpoint_dir flags instead.
if [ -f "/transfer-checkpoint/checkpoint" -a ! -f "/mnt/models/output_graph.tflite" ]; then
echo "Using checkpoint from ${TRANSFER_CHECKPOINT}"
# use --load_checkpoint_dir for transfer learning
LOAD_CHECKPOINT_FROM="--load_checkpoint_dir /transfer-checkpoint --save_checkpoint_dir /mnt/checkpoints"
else
LOAD_CHECKPOINT_FROM="--checkpoint_dir /mnt/checkpoints/"
fi;
Produced checkpoints are using absolute path where they should really be using relative ones.
Checkpoints FR STT 0.9
Here is English checkpoint for reference.
Fix should probably go into
package.sh
I think.Related : https://github.com/coqui-ai/STT/issues/2338 ^ STT's
lm_optimizer
throws aNotFoundError
from TF, since the path points to nowhere in transfer-learning configuration, as checkpoint path is/transfer-checkpoint/
instead of/mnt/checkpoints/
. https://github.com/common-voice/commonvoice-fr/blob/5699e59244d14bb14d5b7603b91c934b761c9194/DeepSpeech/CONTRIBUTING.md?plain=1#L76-L78 https://github.com/common-voice/commonvoice-fr/blob/5699e59244d14bb14d5b7603b91c934b761c9194/DeepSpeech/train.sh#L14-L17That was for DS but with STT, we have access to
--load_checkpoint_dir
and--save_checkpoint_dir
flags instead.Example from my branch stt140-cv9