Hello OLMo: A truly open LLM

Re:Hello OLMo: A truly open LLM by mysteria on Monday April 08, 2024 @08:41PM (#39975206)

Is this one of the first LLMs of note that was successfully trained on AMD GPUs? I wonder how seamless the process was and if they faced any issues there.

[ Parent | on HN ]

Re:Hello OLMo: A truly open LLM by sanxiyn on Monday April 08, 2024 @09:23PM (#39975450)

Databricks (who also participated in OLMo, it's probably the same codebase) trained on AMD before, see 2023 post https://www.databricks.com/blog/amd-mi250. It was probably seamless, as any issues were fixed by Databricks in 2023.

[ Parent | on HN ]

Re:Hello OLMo: A truly open LLM by otuutti on Tuesday April 09, 2024 @06:57AM (#39978115)

https://huggingface.co/LumiOpen/Poro-34B Also fully trained on LUMI.

(more models here: https://huggingface.co/LumiOpen)

[ Parent | on HN ]