Willing to help test the MLA branch on GLM-5.2 (H20)

[lvxiaobu615]

Hi @zzz0906, we're working with GLM-5.2 on 8× H20 and would be happy to help test the MLA branch. If you're able to share the WIP, we can run INT2 vs BF16 accuracy comparisons and report back. Let me know what works for you. Thanks!

[JiashuWu]

Hi @zzz0906 , I am also willing to help test the GLM branch. If the code is not ready to be published onto GitHub, do you mind sending me the code via email at jiwu0083 [at] alumni [dot] sydney [dot] edu [dot] au, look forward to your reply. Many thanks.