** Apologies to all ............. I tested this using JammyPup64CE **
I did do this myself and found the results incredibly promising.
This example used US English as the language model.
Once the application was running, I had my radio on in the background and it automatically began transcribing the dialogue.
Then I opened up Abiword and the text "as if by magic" appeared in the open document
You will need to load up the DevX.SFS first (tested only in FP95) as you will need to use Python and Git.
Python 3.10 is included in the DevX.
Open up a terminal and enter the following one line at a time.
Ensuring each command has been completed.
Code: Select all
pip install cffi
Code: Select all
pip3 install vosk
Code: Select all
git clone https://github.com/ideasman42/nerd-dictation.git
Code: Select all
cd nerd-dictation
Code: Select all
wget https://alphacephei.com/kaldi/models/vosk-model-small-en-us-0.15.zip
Code: Select all
unzip vosk-model-small-en-us-0.15.zip
Code: Select all
mv vosk-model-small-en-us-0.15 model
Once you have completed the above, you are ready to test it yourself.
So, have something to say, or play some music ........etc
Then in the same terminal window enter this to begin the dictation
Code: Select all
./nerd-dictation begin --vosk-model-dir=./model &
You can open up Abiword or a similar to see the results at the same time as the open terminal.
To end the test, put the following into a terminal window.
Code: Select all
./nerd-dictation end
Impressed??
If so, additional language models are available here: