🔥 flame - a fla-hub Collection

fla-hub 's Collections

GSA

HGRN2

GLA

RWKV6

RetNet

Mamba

HGRN

Qwen2.5

🔥 flame

updated Mar 18

A collection of baselines trained by 🔥 flame

fla-hub/transformer-340M-4K-0.5B-20480-lr3e-4-cosine

0.4B • Updated Mar 14 • 22 • 1
fla-hub/transformer-340M-4K-0.5B-20480-lr3e-4-decay0.1-sqrt

0.4B • Updated Mar 14 • 2