Are you guys considering releasing DeepSpeech for M1 macs? They are really powerful and since I have one in my hand, I tried to install and use DeepSpeech for 3 days non-stop. I tried installing it with conda, pyenv, default Python libraries, changing the arch settings, installing a new Python version from the official website, and still had this issue. I could proceed further into it, so I thought it would be a good idea to ask you.
No, we don’t have plans because:
- DeepSpeech update, grant and playbook
- [CALL FOR HACKING] Use of GitHub Actions
- we don’t have hardware for working on it
- we don’t have hardware for CI on it (GitHub Actions does not support)
- I have no time to work on that
- I don’t think TensorFlow r2.3 builds on M1.
Now, since we have GitHub Actions in place with coverage for macOS amd64 and including python bindings, if you have hardware you can try and setup local github actions runner and add support for it.
Yes I have M1 MacBook Pro right now. I managed to install and use TF with lots of tweaking and installing on x86_64 base. I thought that I could install or compile DeepSpeech as well. And actually installed it, but couldn’t import it.
I will check out the GitHub Actions as well. Thank you for your reply.
Any updates on DeepSpeech build on Apple silicon? I am planning on trying the build and any information about the process will be helpful.
Thanks in advance