Junbum Lee

icon indicating a website link https://junbuml.ee icon indicating an email [email protected]

icon location Seoul, South Korea icon location AI/ML GDE @ml-gde. Korean AI/NLP Researcher and creator of multiple Korean PLMs. Focused on advancing Open LLMs.

Results 34 comments of Junbum Lee

Great work. quick question on roadmap ETA?

Currently I'm investigating on it. plan to make it compatible with Bitsandbytes. Since I'm not familiar with CUDA or C++, it'll takes some time, so currently I could not guarantee...

Great work. quick question on roadmap ETA?

> you way want to check out https://github.com/IST-DASLab/qmoe. They created some custom cuda functions for sub 1-bit weights. Thanks for information! I'm looking on it 😉

BitNet_Llama_model_test_huggingface_GPU.ipynb

Thanks for notice! Could you provide the colab code you tried to run? it would be helpful to check the issue😄

llama-2-ko-70b-kobest.sh

No specific reason, if the result is same, you could use it with hf causal experimental as default. I'm currently investing if the results are same or not.