dataset format recommendations: early impressions

[batchember]

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: FLUX, vLLM, M3 Max, with a lot of attention on quant drift.

Maybe useful as a starting point for newer folks.

[phasenode]

Thought I'd post a small recipe because a lot of threads ask similar things.

A couple things I'd check first. If you're on LM Studio, I'd confirm the version first and then reduce variables until the behavior is obvious.

[emberquill]

A couple things I'd check first. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at HBM constraints.

[promptpoint]

I tested this tonight. On M3 Max with Ollama, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput collapsed.

[plainworks]

I tested this tonight. On M3 Max with Ollama, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput col

subscribed, curious where this lands.

[audioforge]

What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at NVFP4. Not glamorous advice, but repeatable baselines save a lot of time.

[phaseport36]

Ran a quick check on my box. On Threadripper box with Ollama, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput collapsed.

[audioforge]

Ran a quick check on my box. On Threadripper box with Ollama, I saw basically the same behavior. The surprise was that memory pressure showed up befor

A couple things I'd check first. On 7900 XTX, 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.

[forgemixer]

Maybe, but I had the opposite result. I wouldn't optimize around that first. On my side the bigger swing came from the runtime and not the model. Worth testing the simpler explanation before adopting the more interesting one. Turns out 'just one more test' was a lie.

[routeline23]

Good point. I don't think your conclusion is far off.

Appreciate the detail. That matches what I saw on 7900 XTX. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.