Hello, I’ve been trying to train a model but unfortunately 1) there are no sample datasets small enough for me to test the workflow (at least none that I know of) 2) I can’t afford to train with more resources.
Really, I just want to know
- How do you pause training?
VM or laptop, I need them to shut down at the very least for updates. But I can’t find information anywhere how to safely halt the training process. I know there are checkpoints, but that doesn’t tell me how or if I can safely turn off my computer during training if there’s no way to pass a command to pause training. - I’m hoping when training completes, a model will be generated. Is this model a file I can use?
- Are there small datasets I can work with? Documentation points to something from UPenn, but the link seems to no longer work, and online search points me to individual files I don’t know how to setup to pass to DeepSpeech. I’m not looking to produce an accurate model, but just something I can play with to test the training process.
Ideally, it would be a much smaller version of the English speech recognition dataset I could finish training in a couple hours on a laptop. The plan is to replace the English text with IPA symbols to test a concept.
Thank you for your help!