🔄 In a Training Loop

Frankenbeast

vmlinux

5

·

AI & ML interests

I like paragliding, and learning about LLM's.

Recent Activity

updated a model 1 day ago

vmlinux/Muse-Glimmer-30B-heretic-ROCmFPX-GGUF

published a model 1 day ago

vmlinux/Muse-Glimmer-30B-heretic-ROCmFPX-GGUF

updated a model 2 days ago

vmlinux/Muse-Glimmer-30B-ROCmFPX-GGUF

View all activity

Organizations

None yet

New activity in nightmedia/Qwen3.6-27B-Architect-Polaris2-Fable-B-F451 3 days ago

Oh.. this writes nicely. I'm going to try and make a ROCmFP8 and ROCmFP4 of it for the STRIX halo!

#5 opened 4 days ago by

New activity in jcbtc/Laguna-S-2.1-Chadrock-ROCmFP4-StrixKVSpine-V4-GGUF 13 days ago

Good job, I'm glad I didn't have to spend days doing this. Are you planning on releasing a QUALITY Q6?

#4 opened 18 days ago by

New activity in vmlinux/Qwen3.5-122B-A10B-Heretic-ROCmFP4-iMatrix-GGUF 18 days ago

Nice Build

#1 opened 28 days ago by

New activity in amd/Instella-MoE-16B-A3B-Think 19 days ago

Congratulations and thank you to AMD on the release!

#1 opened 19 days ago by

commented a paper about 2 months ago

Deeper is Not Always Better: Mitigating the Alignment Tax via Confident Layer Decoding

Paper • 2606.21906 • Published Jun 20 • 26 •