Posts: 68
Threads: 3
Joined: Sep 2026
Reputation:
0
Thought I'd post a small recipe because a lot of threads ask similar things.
Just sharing a setup that ended up working reasonably well: DeepSeek, PyTorch, mini PC, with a lot of attention on speculative decoding.
If someone wants the exact settings I can clean them up and post them.
For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.
Posts: 68
Threads: 3
Joined: Sep 2026
Reputation:
0
that fixed it for me too.
Posts: 53
Threads: 4
Joined: Sep 2026
Reputation:
0
What exact setup are you running? Are you testing this on 7900 XTX or something else?
Posts: 93
Threads: 10
Joined: Oct 2026
Reputation:
0
that fixed it for me too.
Posts: 33
Threads: 5
Joined: Aug 2026
Reputation:
0
Need a little more detail here. Do you see the same thing with a smaller batch or shorter prompt?
Posts: 30
Threads: 6
Joined: Aug 2026
Reputation:
0
That matches what I was seeing. The pattern around tokens/sec has been pretty similar for me. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
Posts: 64
Threads: 12
Joined: Sep 2026
Reputation:
0
Good point. I don't think your conclusion is far off.
learning in public
Posts: 37
Threads: 4
Joined: Aug 2026
Reputation:
0
I think the bottleneck is somewhere else. I wouldn't optimize around that first. On my side the bigger swing came from the runtime and not the model.
Posts: 142
Threads: 20
Joined: Sep 2026
Reputation:
0
A couple things I'd check first. Before rebuilding anything, I'd test one clean baseline with Phi and compare from there. Not glamorous advice, but repeatable baselines save a lot of time.
small models can still be useful
Posts: 42
Threads: 4
Joined: Oct 2026
Reputation:
0
I tested this tonight. The issue went away once I changed one setting and restarted the stack cleanly. The surprise was that memory pressure showed up before throughput collapsed.
curious and caffeinated
