cnxhk

This is more like ASI instead of AGI

Any possibility to share some of the eval audio?

Hey Tianyi,

one question is that do you think in the future harness would become more simpler and its behavior should match a guideline or we would add more complexities to make it more robust? Is it important to use the same harness for RL and inference?

You can read far more text than listening. If the voice response is too long, people lose the patience quickly. So it is better to show it on screen if we have too much text.

The paper is quite interesting but efficiency on OCR tasks does not mean it could be plugged into a general llm directly without performance loss. If you train a tokenizer only on OCR text you might be able to get better compression already.

I am wondering if we could use any llm to generate similar graphs lol.

They recently released a subset of LibriSpeech called LibriTTS which consists only clean speech.

It doesn't look correct to me. According to your website "no visa required" for Americans include only "american territories bermuda canada".

imgcat is very useful for me. Glad they are still maintaining it.

Different with western countries, rural area in general cannot provide you things you need for daily life in China. Besides the life is quite boring without many entertainment. That is why many young people left village to cities.