Skip to main content

Qwen2.5-Coder-3B SFT on Reasoning Datasets

Created on February 16|Last edited on February 16

50100150200250300train/global_step12345
Select runs that logged eval/loss
to visualize data in this line chart.
50100150200250300train/global_step0.0000020.0000040.0000060.0000080.00001
50100150200250300train/global_step0.811.2
OpenThoughts (Instruct)
4
OpenThoughts (Base)
6
s1k (Instruct)
6