DeepSeek-Coder VS Yi

deepseek-coder-instruct 6.7B still looks like is better than llama 3 8B on HumanEval [0], and deepseek-coder-instruct 33B still within reach to run on 32 GB Macbook M2 Max - Lamma 3 70B on the other hand will be hard to run locally unless you really have 128GB ram or more. But we will see in the following days how it performs in real life.

[0] https://github.com/deepseek-ai/deepseek-coder?tab=readme-ov-...

Deepseek (https://github.com/deepseek-ai/DeepSeek-Coder?tab=readme-ov-...) code is MIT and the model license is available too.

https://github.com/deepseek-ai/deepseek-coder

33B Instruct doesn’t beat 6.7B Instruct by much but maybe those % improvements mean more for your usage.

I run 6.7B since I have 16GB RAM.

Deepseek Coder: https://github.com/deepseek-ai/DeepSeek-Coder (Best open source coding model right now)

The model license:

https://github.com/01-ai/Yi/blob/main/MODEL_LICENSE_AGREEMEN...

1) Your use of the Yi Series Models must comply with the Laws and Regulations as

01.ai's Yi model has performed well and there are several strong fine-tunes on Huggingface also.

I wonder what definition of "open source" the author is using or if he even read the license Yi is released under.

The Yi model license agreement [1] restricts usage and requires compliance with the "laws and administrative regulations of the mainland of the People's Republic of China" and they have a separate license that you can apply for if you want to use Yi commercially. [2]

Kudos to the 01.ai team on a strong LLM model but I do wonder if Wired and others should be a little more careful with the use of "open source" when describing AI models.

1. https://github.com/01-ai/Yi/blob/main/MODEL_LICENSE_AGREEMEN...

I doubt it, training a foundational model (very important to make the difference with language model) make very little economic sense when there's already plenty of open source option. And in fact, almost all AI startup use/fine-tune these models. Only big research center will do foundational research, as it as always been the case. (Mistral and 01.ai are outliers, and I don't see how they're ever gonna recoup their cost)

Yi-34b: https://github.com/01-ai/Yi

The 6B model is unfortunately still a base text completion model. I've been waiting for the Chat version it to be open-sourced :). The 01-ai team is working on it! https://github.com/01-ai/Yi/issues/173