https://www.datasculpting.ai/inference/ Mon, 26 Oct 2026 11:43:48 +0000 MyBB https://www.datasculpting.ai/inference/showthread.php?tid=1496 Sun, 18 Oct 2026 16:50:00 +0000 quiettoken]]> https://www.datasculpting.ai/inference/showthread.php?tid=1496 https://www.datasculpting.ai/inference/showthread.php?tid=1390 Sat, 17 Oct 2026 22:18:58 +0000 plaingarden18]]> https://www.datasculpting.ai/inference/showthread.php?tid=1390
Maybe useful as a starting point: llama.cpp on M3 Max, keeping long context in mind from the start.

Feel free to tear it apart if you see something dumb.]]>
Maybe useful as a starting point: llama.cpp on M3 Max, keeping long context in mind from the start.

Feel free to tear it apart if you see something dumb.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1396 Sat, 17 Oct 2026 21:25:31 +0000 tensorkite]]> https://www.datasculpting.ai/inference/showthread.php?tid=1396
It feels like the conversation has shifted hard toward inference and serving details lately, and honestly that makes sense once you start dealing with KV cache pressure.

Maybe obvious to some of you, but it clicked for me this week.]]>
It feels like the conversation has shifted hard toward inference and serving details lately, and honestly that makes sense once you start dealing with KV cache pressure.

Maybe obvious to some of you, but it clicked for me this week.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1405 Sat, 17 Oct 2026 17:49:57 +0000 rasterpoint]]> https://www.datasculpting.ai/inference/showthread.php?tid=1405
It feels like the conversation has shifted hard toward inference and serving details lately, and honestly that makes sense once you start dealing with HBM constraints.

Curious whether that lines up with what other people are seeing.

For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.]]>
It feels like the conversation has shifted hard toward inference and serving details lately, and honestly that makes sense once you start dealing with HBM constraints.

Curious whether that lines up with what other people are seeing.

For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1345 Fri, 16 Oct 2026 20:37:17 +0000 rocmplane]]> https://www.datasculpting.ai/inference/showthread.php?tid=1345
One thing I keep noticing is that people underestimate tail latency until they actually try to run this stuff for more than one afternoon.

Curious whether that lines up with what other people are seeing.]]>
One thing I keep noticing is that people underestimate tail latency until they actually try to run this stuff for more than one afternoon.

Curious whether that lines up with what other people are seeing.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1333 Fri, 16 Oct 2026 20:25:50 +0000 Data Sculpting]]> https://www.datasculpting.ai/inference/showthread.php?tid=1333
I'm comparing two approaches here: keep it simple on B200, or squeeze more out of the stack and deal with the complexity.

Trying to avoid rebuilding the whole thing twice.

I care more about stability than squeezing the last possible tokens/sec out of it.]]>
I'm comparing two approaches here: keep it simple on B200, or squeeze more out of the stack and deal with the complexity.

Trying to avoid rebuilding the whole thing twice.

I care more about stability than squeezing the last possible tokens/sec out of it.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1343 Fri, 16 Oct 2026 11:16:21 +0000 vectorshed43]]> https://www.datasculpting.ai/inference/showthread.php?tid=1343
I'm not trying to build a perfect architecture diagram. I just want a sane starting point around LM Studio and GB200.

Simple and boring is a valid answer.]]>
I'm not trying to build a perfect architecture diagram. I just want a sane starting point around LM Studio and GB200.

Simple and boring is a valid answer.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1292 Thu, 15 Oct 2026 10:05:17 +0000 labnest]]> https://www.datasculpting.ai/inference/showthread.php?tid=1292
I've gone back and forth between Ollama and PyTorch for Kimi. The interesting difference for me has been MoE serving.

Interested in tradeoffs more than fan loyalty.

I care more about stability than squeezing the last possible tokens/sec out of it.]]>
I've gone back and forth between Ollama and PyTorch for Kimi. The interesting difference for me has been MoE serving.

Interested in tradeoffs more than fan loyalty.

I care more about stability than squeezing the last possible tokens/sec out of it.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1222 Wed, 14 Oct 2026 23:46:54 +0000 ggufwire22]]> https://www.datasculpting.ai/inference/showthread.php?tid=1222
After a few runs on 4090, I ended up with decent performance, but the setup only really made sense once I adjusted around tool calling.

Happy to post more details if anyone wants them.]]>
After a few runs on 4090, I ended up with decent performance, but the setup only really made sense once I adjusted around tool calling.

Happy to post more details if anyone wants them.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1193 Tue, 13 Oct 2026 23:01:48 +0000 rocmplane]]> https://www.datasculpting.ai/inference/showthread.php?tid=1193
I'm helping a small team set up a local workflow. We don't need a giant stack, but we do care about consumer GPUs.

I'm especially interested in mistakes worth avoiding.

This is mostly for local work rather than public serving.]]>
I'm helping a small team set up a local workflow. We don't need a giant stack, but we do care about consumer GPUs.

I'm especially interested in mistakes worth avoiding.

This is mostly for local work rather than public serving.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1185 Tue, 13 Oct 2026 10:37:42 +0000 llm_wren]]> https://www.datasculpting.ai/inference/showthread.php?tid=1185
Two paths both seem viable, but they break down differently once KV cache pressure becomes important.

What would you pick if you had to live with it for a few months?

I care more about stability than squeezing the last possible tokens/sec out of it.]]>
Two paths both seem viable, but they break down differently once KV cache pressure becomes important.

What would you pick if you had to live with it for a few months?

I care more about stability than squeezing the last possible tokens/sec out of it.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1125 Tue, 13 Oct 2026 04:48:34 +0000 cacheplane98]]> https://www.datasculpting.ai/inference/showthread.php?tid=1125
Just sharing a setup that ended up working reasonably well: Kimi, TensorRT-LLM, A6000, with a lot of attention on consumer GPUs.

Maybe useful as a starting point for newer folks.]]>
Just sharing a setup that ended up working reasonably well: Kimi, TensorRT-LLM, A6000, with a lot of attention on consumer GPUs.

Maybe useful as a starting point for newer folks.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1092 Tue, 13 Oct 2026 04:15:42 +0000 cacheplane98]]> https://www.datasculpting.ai/inference/showthread.php?tid=1092
I tried this with Qwen using MLX on B200. The headline result was better than I expected, but tokens/sec still ended up being the limiting factor.

Happy to post more details if anyone wants them.]]>
I tried this with Qwen using MLX on B200. The headline result was better than I expected, but tokens/sec still ended up being the limiting factor.

Happy to post more details if anyone wants them.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1131 Tue, 13 Oct 2026 03:23:07 +0000 phaseport36]]> https://www.datasculpting.ai/inference/showthread.php?tid=1131
One thing I keep noticing is that people underestimate token cost until they actually try to run this stuff for more than one afternoon.

Wondering if this is just where the whole space is heading now.

The part that surprised me was how much the boring details mattered more than the model choice. At this point the real bottleneck is my patience.]]>
One thing I keep noticing is that people underestimate token cost until they actually try to run this stuff for more than one afternoon.

Wondering if this is just where the whole space is heading now.

The part that surprised me was how much the boring details mattered more than the model choice. At this point the real bottleneck is my patience.]]> https://www.datasculpting.ai/inference/showthread.php?tid=1104 Tue, 13 Oct 2026 00:17:56 +0000 cacheplane98]]> https://www.datasculpting.ai/inference/showthread.php?tid=1104
I'm comparing two approaches here: keep it simple on GB200, or squeeze more out of the stack and deal with the complexity.

Interested in tradeoffs more than fan loyalty.]]>
I'm comparing two approaches here: keep it simple on GB200, or squeeze more out of the stack and deal with the complexity.

Interested in tradeoffs more than fan loyalty.]]>