AI is not your buddy, watch out.

[Vermonter]

This is not token heavy, its actually pretty light. The lease analysis that i'm working on will burn through credits....so it'll be interesting to see how it shakes out.

Makes sense. Are you running the Pro or Max plan?

 

[TerryRohrer]

Qwen3.5 dropped and is impressive so far. 35b struggled but 27b works well on my hardware.

Remind me what you are running this on? I was looking at the specs someone said were required and it seemed a bit out of reach.

 

[NP_MAI]

Makes sense. Are you running the Pro or Max plan?

Pro but I suspect that Max is on the horizon. The max subscription costs me half an appraisal for a year's access, completely justifiable even if this automation was the only thing I used.

 

[Vermonter]

Remind me what you are running this on? I was looking at the specs someone said were required and it seemed a bit out of reach.

A ryzen 2700x with 2 rx580s and 64g of ram. That's only 16gb of vram on the gpus, but think having so much system ram helps out

 

[TerryRohrer]

A ryzen 2700x with 2 rx580s and 64g of ram. That's only 16gb of vram on the gpus, but think having so much system ram helps out

Do you have any troubles linking (I assume) the GPUs, or with AMD drivers? I kept running into warnings about both and targeted NVIDIA instead.

 

[Tom D]

OMG, now we have the AI perfectionists on this blog. Just as i was getting over my ansi perfectionists anger. Ansi, seems simplistic after reading the last few threads.

 

[Vermonter]

Do you have any troubles linking (I assume) the GPUs, or with AMD drivers? I kept running into warnings about both and targeted NVIDIA instead.

I have not allocated anything via the config because my gpus are old and technically unsupported, but they both spike when running local models and monitoring via nvtop.

Are you running ollama with nvidia?

 

[TerryRohrer]

I have not allocated anything via the config because my gpus are old and technically unsupported, but they both spike when running local models and monitoring via nvtop.

Are you running ollama with nvidia?

The last parts for my machine are expected to arrive tomorrow. If the weekend goes well, hopefully by Monday!

 

[Vermonter]

Nice. This will help-

Hardware support - Ollama

docs.ollama.com

 

[Alsie35]

What a ****ingly bizarre thread. Outside of the humorous digs at J Grant (whose posts I can't see), the rest of this thread might as well be written in Urdu as far as I can understand it.