Rotem Dan

I uploaded the model when I tried adding support for the large Whisper models, especially V3. The lack of support to the large models was also reported earlier, on an...

On `v1.0.0` I just released, I added support for [`whisper.cpp`](https://github.com/ggerganov/whisper.cpp), which supports large models (`large-v1`, `large-v2`, `large-v3`), quantized models, and GPU acceleration (CUDA and OpenCL). All `whisper.cpp` models (including quantized...

Large models can't fully be supported until `onnxruntime-node` receives support for models over 2GB, with external data. `whisper.cpp` engine supports them (all of its models, including quantized ones, are supports)....

Thanks for the report. The Whisper tokenization is based on a pure JavaScript implementation I made myself, and uses a standard GPT2 tokenizer. This isn't exactly what is used on...

In the newly released `1.0.x` version, I've integrated the `tiktoken` tokenizer. It can deal with many more types of characters, including some very obscure emojis. The tokenizer implementation is now...

Thanks, Two letter language codes are used throughout all synthesis and recognition operations, I believe. The error message makes it look like the language isn't supported. I should change it...

I believe this issue has been resolved a long time ago. Language codes are normalized and multiple language code formats (including language names like "French") are accepted (documentation may still...

Yes, now I remember looking into SMIL and EPUB 3 in the past. Seems like the links on the W3C site are mostly broken, including the tutorials. The last update...

Thanks for all the details! One question I ask is: who would this tool be intended for? For the average user, it isn't really that easy to give a seamless...

The difficulty I described was in getting the word highlighting working correctly in the DOM within the browser extension, it wasn't about the accuracy of the timing. For synthesized text,...