News for hackers. Stuff that matters.

A comment on Qwen2.5-1M: Deploy your own Qwen with context length up to 1M tokens. Parent comment

Re:Qwen2.5-1M: Deploy your own Qwen with context length up t... by seunosewa on Sunday January 26, 2025 @05:41PM (#42834969)
The behaviour you described is what happens when you have small context windows. Perhaps you're feeding the models with more tokens than you think you are. I have enjoyed loading large codebases into AI Studio and getting very satisfying and accurate answers because the models have 1M to 2M token context windows.
[ Parent | on HN ]
Re:Qwen2.5-1M: Deploy your own Qwen with context length up t... by dr_kiszonka on Sunday January 26, 2025 @07:42PM (#42835998)
How do you get those large codebases into AI Studio? Concat everything into one big file?
[ Parent | on HN ]
Re:Qwen2.5-1M: Deploy your own Qwen with context length up t... by adamgordonbell on Sunday January 26, 2025 @08:35PM (#42836387)
Basically yes, I have a helper program, but that's mainly what it does.
[ Parent | on HN ]
Re:Qwen2.5-1M: Deploy your own Qwen with context length up t... by social_quotient on Sunday January 26, 2025 @09:23PM (#42836706)
Concat to a file but it helps to make an ascii tree at the top and then for each merged file out its path and orientation details. I’ve also started playing with adding line ranges to the ascii tree hoping that the LLMs (more specifically the agentic ones) start getting smart enough to jump to the relevant section.
[ Parent | on HN ]
Re:Qwen2.5-1M: Deploy your own Qwen with context length up t... by msoad on Sunday January 26, 2025 @09:25PM (#42836720)
I use yek

https://github.com/bodo-run/yek

[ Parent | on HN ]