Remix.run Logo
Bringing Up DeepSeek-V4-Flash on AMD MI300X(fergusfinn.com)
72 points by kkm 7 hours ago | 7 comments
maCDzP 4 hours ago | parent | next [-]

I train on AMD MI250X and managed to get Gemma 4 31B to work - but it took a lot of work on the software side.

kkm 4 hours ago | parent [-]

This is very interesting, planning to write about it?

mezark 5 hours ago | parent | prev | next [-]

We at doubleword are bullish for AMD for low-interactivity inference - it does just take a bigger lift on the software side...

brcmthrowaway 3 hours ago | parent [-]

Are you long AMD?

latchkey 10 minutes ago | parent | prev | next [-]

Nice work and thanks for being a customer.

(CEO Hot Aisle)

kkm 5 hours ago | parent | prev | next [-]

Also the vllm patch accompanying the blogpost: https://github.com/doublewordai/vllm-amd-blog-doubleword

benlm 5 hours ago | parent | prev [-]

Nice work! Would DeepSeek V4 Pro on 8xMI300X work with these patches?