-
Notifications
You must be signed in to change notification settings - Fork 17
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fixed a few things for M1 arm64 release #80
Conversation
fixed dialog by moving to main process via ipc, fixed ffmpeg and ffprobe bin by adding arm64 arch folder and logic in that module, tested locally with npm start prod and dev, haven't tried packaging yet
testing github acitons releases
asar true
asar false
USE_HARD_LINKS false
USE_HARD_LINKS false
+updated readme
If I understand the templating language correctly here, this snippet restricts the display of DeepSpeech as an option to only x64 Mac ( digital-paper-edit-electron/src/stt-settings/index.html Lines 268 to 272 in 6d0a438
Would a condition like |
Sorry for the delay at @dgw Does deepspeech work on Windows 10? I wanted to remove the deepspeech option for mac M1 ( I wasn't sure if it worked on linux and windows(?) so thought I'd remove that too. You could change it to - {process.platform==="darwin" && process.arch !=='arm64' &&(
+ {process.platform==="darwin"
+ && process.platform==="win32"
+ && process.platform === "linux"
+ && process.arch !=='arm64' &&(
<option value="deepspeech">
Mozilla Deepspeech (offline){" "}
</option>)} To bring back deepspeech to windows and linux (provided deepspeech is compatible with those OS) I'd welcome a PR, if you got time to try it out (on a windows machine?) 😇 Node docs, links |
Hi @pietrop, I ultimately moved on from this tool for the current project as it has too many limitations—especially given that I need to collaborate with others. A cloud service seemed to be the best option for something that "just works" for my employer. I'm happy to toy with stuff like this on my own projects, but the project timeline from work didn't allow for it here. Another thing I tried was tying together a couple of tools built on deepspeech using the WSL2 command line, which certainly got me farther than the paper-edit GUI. In that way, deepspeech "works on Windows 10", but it won't help with paper-edit until the promised graphical interface support rolls out. At that point it would become possible to install and run paper-edit inside WSL and use any STT library that works on a normal Linux desktop (but there's obviously no Windows Subsystem for macOS 😅). As far as testing deepspeech compatibility within the app, natively on Windows 10, I can try to do so but there seem to be other places where it's been disabled in the paper-edit code. For example (from this patch):
Code later in that file references the imported module:
I saw a lot of commented-out stuff while poking around the code a bit just now to refresh my memory of what's where. Based on that, I doubt I'm up to tracking down all of the things I'd have to change just to try deepspeech under Windows, if it will even run. Plus, I still don't understand how the templating works. You suggested a change that—to me—would enable deepspeech if the platform matches darwin, win32, and ( |
Is your Pull Request request related to another issue in this repository ?
#79
Describe what the PR does
fixed dialog by moving to main process via ipc, fixed ffmpeg and ffprobe bin by adding arm64 arch folder and logic in that module, tested locally with npm start prod and dev, haven't tried packaging yet
State whether the PR is ready for review or whether it needs extra work
Work in progress
Additional context
NA