small win thread: early impressions

[quiettoken]

[quote]I'd approach it in a boring order. On M3 Max, the easiest win is often to back off the aggressive settings and see whether the problem survives

Need a little more detail here. Do you see the same thing with a smaller batch or shorter prompt?

[marbletrack]

What worked for me was pretty simple. On 4090, the easiest win is often to back off the aggressive settings and see whether the problem survives. Once you know the plain path works, then add the faster path back in piece by piece.

[thetagrid15]

I tested this tonight. My result was mixed: first run looked bad, second run looked normal. I expected the model choice to matter more than it did.

[quiettoken]

Good point. That matches what I saw on 7900 XTX. The biggest change for me was simplifying the stack instead of adding one more optimization layer.