can i use this free ?
i am not developer and not programmer, I'm looking speech to text for Urdu language videos. i heard whisper v3 it is best for that much accuracy to understand language and give response text form with timestamp simple meaning convert voice to text, that text require into srt file because i want to upload this file for my YouTube videos.
i am not developer and not programmer, I'm looking speech to text for Urdu language videos. i heard whisper v3 it is best for that much accuracy to understand language and give response text form with timestamp simple meaning convert voice to text, that text require into srt file because i want to upload this file for my YouTube videos.
This has apache 2.0 license, so yes you can use this for free. you can even use it in a commercial settings and make money from it.
would you guide me step by step how can i do that ? i mean where can i download and where to install etc etc
would you guide me step by step how can i do that ? i mean where can i download and where to install etc etc
I would recommend checking out YouTube videos.
I checked YouTube after that i posted here
@amgadhasan Sorry for butting in, but I have a hard time understanding the memory limitations on this and other models. Is there any documentation or somewhere I can find this type of information? Because, unfortunately, I don't have a lot of GPU power.
@amgadhasan Sorry for butting in, but I have a hard time understanding the memory limitations on this and other models. Is there any documentation or somewhere I can find this type of information? Because, unfortunately, I don't have a lot of GPU power.
This shouldn't require a lot of gpu power.
You can try this model out on the free colab t4 gpu
Run some tests and note down the vRAM usage and time needed to transcribe etc.
Feel free to check out the demo: https://huggingface.co/spaces/hf-audio/whisper-large-v3
In particular, the YouTube tab might be of interest, since you can transcribe a YouTube audio file simply by pasting the URL