Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Voxtral Transcribe 2

mistral.ai

Post reply on HN
Print view View original

11–20 of 255 posts

Re: Voxtral Transcribe 2

#11

Post by mdrzn »

GaggiX wrote: post #8
mdrzn wrote: post #5

There's no comparison to Whisper Large v3 or other Whisper models.. Is it better? Worse? Why do they only compare to gpt4o mini transcribe?

Gpt4o mini transcribe is better and actually realtime. Whisper is trained to encode the entire audio (or at least 30s chunks) and then decode it.

So "gpt4o mini transcribe" is not just whisper v3 under the hood? Btw it's $0.006 / minute

For Whisper API online (with v3 large) I've found "$0.00125 per compute second" which is the cheapest absolute I've ever found.

Re: Voxtral Transcribe 2

#12

Post by GaggiX »

emmettm wrote: post #10
GaggiX wrote: post #8

Earlier quoted context omitted.

Gpt4o mini transcribe is better and actually realtime. Whisper is trained to encode the entire audio (or at least 30s chunks) and then decode it.

The linked article claims the average word error rate for Voxtral mini v2 is lower than GPT-4o mini transcribe

Gpt4o mini transcribe is better than whisper, the context is the parent comment.

Re: Voxtral Transcribe 2

#13

Post by tekacs »

mdrzn wrote: post #5

There's no comparison to Whisper Large v3 or other Whisper models.. Is it better? Worse? Why do they only compare to gpt4o mini transcribe?

WER is slightly misleading, but Whisper Large v3 WER is classically around 10%, I think, and 12% with Turbo.

The thing that makes it particularly misleading is that models that do transcription to lowercase and then use inverse text normalization to restore structure and grammar end up making a very different class of mistakes than Whisper, which goes directly to final form text including punctuation and quotes and tone.

But nonetheless, they're claiming such a lower error rate than Whisper that it's almost not in the same bucket.

Re: Voxtral Transcribe 2

#14

Post by tekacs »

tekacs wrote: post #13
mdrzn wrote: post #5

There's no comparison to Whisper Large v3 or other Whisper models.. Is it better? Worse? Why do they only compare to gpt4o mini transcribe?

WER is slightly misleading, but Whisper Large v3 WER is classically around 10%, I think, and 12% with Turbo. The thing that makes it particularly misleading is that models that do transcription to lowercase and then use inverse text normalization to restore structure and grammar end up making a very different class of mistakes than Whisper, which goes directly to final form text including punctuation and quotes and t…

On the topic of things being misleading, GPT-4o transcriber is a very _different_ transcriber to Whisper. I would say not better or worse, despite characterizations such. So it is a little difficult to compare on just the numbers.

There's a reason that quite a lot of good transcribers still use V2, not V3.

Re: Voxtral Transcribe 2

#15

Post by Archelaos »

As a rule of thumb for software that I use regularly, it is very useful to consider the costs over a 10-year period in order to compare it with software that I purchase for lifetime to install at home. So that means 1,798.80 $ for the Pro version.

What estimates do others use?

Re: Voxtral Transcribe 2

#16

Post by GaggiX »

mdrzn wrote: post #11
GaggiX wrote: post #8

Earlier quoted context omitted.

Gpt4o mini transcribe is better and actually realtime. Whisper is trained to encode the entire audio (or at least 30s chunks) and then decode it.

So "gpt4o mini transcribe" is not just whisper v3 under the hood? Btw it's $0.006 / minute For Whisper API online (with v3 large) I've found "$0.00125 per compute second" which is the cheapest absolute I've ever found.

>So it's not just whisper v3 under the hood?

Why it should be Whisper v3? They even released an open model: https://huggingface.co/mistralai/Voxtral-Mini-4B-Realtime-26...

Re: Voxtral Transcribe 2

#17

Post by antirez »

Italian represents, I believe, the most phonetically advanced human language. It has the right compromise among information density, understandability, and ability to speech much faster to compensate the redundancy. It's like if it had error correction built-in. Note that it's not just that it has the lower error rate, but is also underrepresented in most datasets.

Re: Voxtral Transcribe 2

#19

Post by Empact »

varispeed wrote: post #6

[flagged]

Many people speak Russian, including many who do not live in Russia, e.g. about 30% of Ukranians.

Beyond that, I don't see how we stand to durably reduce military action by making languages mutually unintelligible.

https://simple.wikipedia.org/wiki/Russian_language#/media/Fi...

Re: Voxtral Transcribe 2

#20

Post by simonw »

This demo is really impressive: https://huggingface.co/spaces/mistralai/Voxtral-Mini-Realtim...

Don't be confused if it says "no microphone", the moment you click the record button it will request browser permission and then start working.

I spoke fast and dropped in some jargon and it got it all right - I said this and it transcribed it exactly right, WebAssembly spelling included:

> Can you tell me about RSS and Atom and the role of CSP headers in browser security, especially if you're using WebAssembly?

Post reply on HN