Comment by woodson
1 day ago
This generic answer from Wikipedia is not very helpful in this context. Zero-shot voice cloning in TTS usually means that data of the target speaker you want the generated speech to sound like does not need to be included in the training data used to train the TTS models. In other words, you can provide an audio sample of the target speaker together with the text to be spoken to generate the audio that sounds like it was spoken by that speaker.
Why wouldn’t that be one-shot voice cloning? The concept of calling it zero shot doesn’t really make sense to me.
Zero-shot means zero-retraining, so think along the lines of "Do you need to modify the weights? Or can you keep the weights fixed and you only need to supply an example?"
As with other replies, yes this is a silly name.
> Zero-shot means zero-retraining, so think along the lines of "Do you need to modify the weights? Or can you keep the weights fixed and you only need to supply an example?"
I would caution that using the term "example" suggests further learning happens at inference-time, which isn't the case.
For LLMs, the entire prompt is the input and conveys both the style and the content vectors. In zero-shot voice cloning, we provide the exact same inputs vectors but just decoupled. Providing reference audio is no different than including "Answer in the style of Sir Isaac Newton" in an LLM's prompt. The model doesn't 'learn' the voice; it simply applies the style vector to the content during the forward pass.
Providing inference-time context (in this case, audio) is no different than giving a prompt to an LLM. Think of it as analogous to an AGENTS.md included in a prompt. You're not retraining the model, you're simply putting the rest of the prompt into context.
If you actually stopped and fine-tuned the model weights on that single clip, that would be one-shot learning.
To me, a closer analogy is In Context Learning.
In the olden days of 2023, you didn’t just find instruct-tuned models sitting on every shelf.
You could use a base model that has only undergone pretraining and can only generate text continuations based on the input it receives. If you provided the model with several examples of a question followed by an answer, and then provided a new question followed by a blank for the next answer, the model understood from the context that it needed to answer the question. This is the most primitive use of ICL, and a very basic way to achieve limited instruction following behavior.
With this few-shot example, I would call that few-shot ICL. Not zero shot, even though the model weights are locked.
But, I am learning that it is technically called zero shot, and I will accept this, even if I think it is a confusingly named concept.
I don't disagree, but that's what people started calling it. Zero-shot doesn't make sense anyway, as how would the model know what voice it should sound like (unless it's a celebrity voice or similar included in the training data where it's enough to specify a name).
> Zero-shot doesn't make sense anyway, as how would the model know what voice it should sound like (unless it's a celebrity voice or similar included in the training data where it's enough to specify a name).
It makes perfect sense; you are simply confusing training samples with inference context. "Zero-shot" refers to zero gradient updates (retraining) required to handle a new class. It does not mean "zero input information."
> how would the model know what voice it should sound like
It uses the reference audio just like a text based model uses a prompt.
> unless it's a celebrity voice or similar included in the training data where it's enough to specify a name
If the voice is in the training data, that is literally the opposite of zero-shot. The entire point of zero-shot is that the model has never encountered the speaker before.
2 replies →
So if you get your target to record (say) 1 hour of audio, that's a one-shot.
If you didn't do that (because you have 100 hours of other people talking), that's zero-shots, no?
> So if you get your target to record (say) 1 hour of audio, that's a one-shot.
No, that would still be zero shot. Providing inference-time context (in this case, audio) is no different than giving a prompt to an LLM. Think of it as analogous to an AGENTS.md included in a prompt. You're not retraining the model, you're simply putting the rest of the prompt into context.
If you actually stopped and fine-tuned the model weights on that single clip, that would be one-shot learning.
2 replies →
It’s nonsensical to call it “zero shot” when a sample of the voice is provided. The term “zero shot cloning” implies you have some representation of the voice from another domain - e.g. a text description of the voice. What they’re doing is ABSOLUTELY one shot cloning. I don’t care if lots of STT folks use the term this way, they’re wrong.
> This generic answer from Wikipedia is not very helpful in this context.
Actually, the general definition fits this context perfectly. In machine learning terms, a specific 'speaker' is simply a 'class.' Therefore, a model generating audio for a speaker it never saw during training is the exact definition of the Zero-Shot Learning problem setup: "a learner observes samples from classes which were not observed during training," as I quoted.
Your explanation just rephrases the very definition you dismissed.
From your definition:
> a learner observes samples from classes which were not observed during training, and needs to predict the class that they belong to.
That's not what happens in zero-shot voice cloning, which is why I dismissed your definition copied from Wikipedia.
> That's not what happens in zero-shot voice cloning
It is exactly what happens. You are confusing the task (classification vs. generation) with the learning paradigm (zero-shot).
In the voice cloning context, the class is the speaker's voice (not observed during training), samples of which are generated by the machine learning model.
The definition applies 1:1. During inference, it is predicting the conditional probability distribution of audio samples that belong to that unseen class. It is "predict[ing] the class that they belong to," which very same class was "not observed during training."
You're getting hung up on the semantics.
1 reply →