Nvidia Nemotron 3.5 Transcribe: open ASR for 40 languages

Posted by wondersky 14 hours ago

Counter116Comment20OpenOriginal

Comments

Comment by throwaway81523 7 hours ago

ASR = automatic speech recognition. I think this is the same as STT, speech to text?

It seems to want an NVidia GPU. I didn't look closely to find out if it can run usably without one. For me, usable might mean 10% of realtime, i.e. an overnight run on my old x86 box for 1 hour of transcription. Of course faster is always nice. YMMV.

Comment by Farmadupe2 5 hours ago

Almost all ML models come with reference python code that will run on most architectures.

If you want an implementation optimized for a certain architecture the cost is between 10-30USD of opus tokens and is oneshottable: ("implement Nvidia nemotron 3.5 transcribe for the intel npu on this machine. Use the hf transformers implementation as reference. The finished artifact takes pcm on stdin and utf8 chunks on stdout".)

You can substitute your old intel CPU by modifying the prompt by replacing "intel npu" with "the target cou on this machine ... Your task is complete when the model decode performance is 80% of the memory read speed of this machine"

Comment by throwaway81523 54 minutes ago

Thanks, that's interesting, I hoped maybe it would use pytorch or something. But what I'm wondering is whether the result would run at usable speed. I don't currently use anything like opus and only want to use open weight models, and at the moment I don't have a GPU machine, though I guess I can spin up an hourly VM with one sometime.

Comment by solarkraft 11 hours ago

> Transcribe is built for recorded audio, not live streaming.

Bit unfortunate! Nemotron Streaming in Handy has basically solved dictation for me. It recognizes German mixed with English terms and brand names really well.

Comment by PorciiVorbesc 11 hours ago

Can you share some info how you got that working?

Comment by ndom91 11 hours ago

https://handy.computer/

Select nemotron in the model selection

Comment by PorciiVorbesc 10 hours ago

Oh, since it's called "handy" I expected it would work on mobile phones, as per the name suggests. That seems it only has desktop operating systems.

Also, there's no mention of Nemotron in that handy. Is that the same Handy GGP was talking about?

Comment by killingtime74 9 hours ago

There are apps on the Google Play Store where you can run on-device models or connect to OpenRouter where they are run for you. Im using Qwen 3 ASR locally on my android, it's the second most accurate open model after the one in this post.

Comment by singhkays 5 hours ago

what app do you run on Android for this use-case?

Comment by killingtime74 23 minutes ago

https://github.com/DevEmperor/DictateKeyboard. I forked it to add more models (with Claude). I use openrouter models as well as the local ones

Comment by AkashKaStudio 12 hours ago

From the model card, this supports near-field only speaker transcription, experimental but still something not available in any other OSS Speech LLM.

Comment by xpct 9 hours ago

I played around with Nemotron 3 recently. It's OK, but makes an error every few sentences, at least on the data I worked with (standard vocab). I'm sure we'll get there, but I was pleasantly surprised by how I can recognize words better than it can.

Comment by jiehong 9 hours ago

Agreed. I find Parakeet or Qwen ASR better.

I’m still sad Voxtral 3 wasn’t made open source by Mistral.

Comment by wahnfrieden 11 hours ago

How is it for Japanese

Comment by mewse-hn 9 hours ago

testing it against parakeet right now for ASR, it's slower, heavier, and has a higher error rate. but i'm doing a pass with the embedded translation to see if that's any good, it might be worth it.

e: also i'm testing it with a Q4 quant that might not be full accuracy

Comment by mewse-hn 8 hours ago

ugh, translation is worse than passing a parakeet transcript to a cloud llm. disappointing model.

Comment by mewse-hn 8 hours ago

tried running it against a full anime episode, both vocal isolated and raw. it got confused by english words in the theme song and failed spectacularly three times

Comment by jiehong 9 hours ago

[dead]

Comment by mrkn1 11 hours ago

[flagged]

Comment by llacane 9 hours ago

how is this related with the main topic? also it seems your project (by your bio), it would be nice to disclose it as this is self-promotion.

Comment by vjulian 9 hours ago

I have a few thousand audio interviews as MP3, around an hour each and with specialized vocabulary. Typically there are no more than 2 to 3 speakers. Audio quality ranges from mediocre to fair to good.

I was thinking of creating a local app that calls OpenAI’s VTT model. In my most recent evaluation, the OpenAI model was best for my use case. But is that indeed the case?

Comment by knicholes 9 hours ago

We may never know.