FRESH

Hacker News

Bringing Up DeepSeek-V4-Flash on AMD MI300X

117 points by kkm

by maCDzP

1 subcomments

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.

by kkm

0 subcomment

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

by mezark

2 subcomments

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

by edg5000

2 subcomments

Checked out this company about a year ago and they only offered small models. Now I see they have GLM-fp8/Kimi and DeepSeek V4 Pro. Since workloads are predominantly cached input, I'm surprised to see no separate price for cached input vs uncached. I hope the prices will drop significantly; with these prices you'll end up with thousands in monthly costs quickly. Hopefully more hardware companies will be on the market in the coming years. If the Chinese eventually start competing with the current memory makers, maybe that will help.

by benlm

1 subcomments

by latchkey

1 subcomments

by alfiedotwtf

0 subcomment

It’s just weird Deepseek released a model that was not compatible with any of the usual engines. Without derez’s new project just to support DSv4, how long until it’s actually viable in llama :(

by jing09928

0 subcomment