3 ms·

What does it mean to reproduce DeepSeek R1-Zero? Like they have a model of equivalent performance? Is there a simple explanation of this post for those who aren

What does it mean to reproduce DeepSeek R1-Zero? Like they have a model of equivalent performance? Is there a simple explanation of this post for those who aren't machine learning experts? Also is the technique here related at all to the technique people think DeepSeek themselves used, where they apparently trained the model using OpenAI outputs?

3nthusia5t 2y ago
Could you provide source for the training the model on OpenAI outputs? I can’t find any news about that.
- blackeyeblitzar 2y ago
  I don't have a source to share, but I saw this claim on social media a few times in the last couple days, where people said their conversation with the model revealed that it thought it was some other OpenAI model. I have no idea how such training can work using another model's output, but I saw comments claiming that this is why their training was so cheap.
suraci 2y ago
> What does it mean to reproduce DeepSeek R1-Zero? means it's reproducible
- coolThingsFirst 2y ago
  Westerners cant reproduce chinese geniuses
evertedsphere 2y ago
reproducing the alphazero-like "model learns to reason on its own without supervised fine-tuning" phenomenon that deepseek-r1-zero exhibited
dvh 2y ago
Reminds me of old polish encyclopedia: horse - everybody knows what horse is https://en.wikipedia.org/wiki/Nowe_Ateny https://en.wikipedia.org/wiki/Nowe_Ateny
coolThingsFirst 2y ago
I think there are 2 levels in the brain. One is used for programming the other for language. Doing them in parallel fails for some reason. A lot of GH projects just don't have solid explanation - i don't know what they built.
zamadatix 2y ago
R1-Zero is trained differently than most reasoning models, such as the "normal" R1 model, in regards what steps are done in training. TinyZero applies the same approach (but only on a subset of use cases) on a much smaller model to show it can apply on much smaller models as well. The details of how it's trained different start to get into "machine learning expert" territory but you can get a decent high level via a casual read through of the DeepSeek link if you want to dive deeper.

3nthusia5t 2y ago

Could you provide source for the training the model on OpenAI outputs? I can’t find any news about that.

blackeyeblitzar 2y ago
I don't have a source to share, but I saw this claim on social media a few times in the last couple days, where people said their conversation with the model revealed that it thought it was some other OpenAI model. I have no idea how such training can work using another model's output, but I saw comments claiming that this is why their training was so cheap.

suraci 2y ago

> What does it mean to reproduce DeepSeek R1-Zero? means it's reproducible

coolThingsFirst 2y ago
Westerners cant reproduce chinese geniuses

evertedsphere 2y ago

reproducing the alphazero-like "model learns to reason on its own without supervised fine-tuning" phenomenon that deepseek-r1-zero exhibited

dvh 2y ago

Reminds me of old polish encyclopedia: horse - everybody knows what horse is https://en.wikipedia.org/wiki/Nowe_Ateny https://en.wikipedia.org/wiki/Nowe_Ateny

coolThingsFirst 2y ago

I think there are 2 levels in the brain. One is used for programming the other for language. Doing them in parallel fails for some reason. A lot of GH projects just don't have solid explanation - i don't know what they built.

zamadatix 2y ago

R1-Zero is trained differently than most reasoning models, such as the "normal" R1 model, in regards what steps are done in training. TinyZero applies the same approach (but only on a subset of use cases) on a much smaller model to show it can apply on much smaller models as well. The details of how it's trained different start to get into "machine learning expert" territory but you can get a decent high level via a casual read through of the DeepSeek link if you want to dive deeper.