09-11-2026, 07:41 PM
Tried to reproduce it just now. On mini PC with PyTorch, I saw basically the same behavior. The main thing that moved was latency, not quality.
|
KV cache compression worth it
|
|
« Next Oldest | Next Newest »
|
| Messages In This Thread |
|
KV cache compression worth it - by batchscope72 - 09-06-2026, 12:06 PM
Re: KV cache compression worth it - by batchshed - 09-06-2026, 12:45 PM
Re: KV cache compression worth it - by thetaplane - 09-06-2026, 01:22 PM
Re: KV cache compression worth it - by kernelpeak - 09-06-2026, 05:41 PM
Re: KV cache compression worth it - by bytelab - 09-09-2026, 11:06 PM
Re: KV cache compression worth it - by quantyard - 09-11-2026, 07:41 PM
Re: KV cache compression worth it - by thetaember38 - 09-06-2026, 06:21 PM
Re: KV cache compression worth it - by phaseshed - 09-06-2026, 10:59 PM
Re: KV cache compression worth it - by mangopath4 - 09-11-2026, 02:03 PM
Re: KV cache compression worth it - by kernelgrid - 09-11-2026, 11:24 AM
Re: KV cache compression worth it - by batchshed - 09-11-2026, 06:42 PM
Re: KV cache compression worth it - by mlxgarden - 09-11-2026, 08:03 PM
|
Users browsing this thread: 1 Guest(s)