4 ms·

I think the llamafile[0] system works the best. Binary works on the command line or launches a mini webserver. Llamafile offers builds of Mixtral-8x7B-Instruct,

I think the llamafile[0] system works the best. Binary works on the command line or launches a mini webserver. Llamafile offers builds of Mixtral-8x7B-Instruct, so presumably they may package this one up as well (potentially a quantized format). You would have to confirm with someone deeper in the ecosystem, but I think you should be able to run this new model as is against a llamafile? [0] https://github.com/Mozilla-Ocho/llamafile https://github.com/Mozilla-Ocho/llamafile

noman-land 3y ago
+1 on llamafile. You can point it to a custom model.
jart 3y ago
llamafile author here. I'm downloading Mixtral 8x22b right now. I can't say for certain it'll work until I try it, but let's keep our fingers crossed! If not, we'll be shipping a release as soon as possible that gets it working. My recent work optimizing CPU evaluation https://justine.lol/matmul/ https://justine.lol/matmul/ may have come at just the right time. Mixtral 8x7b always worked best at Q5_K_M and higher, which is 31GB. So unless you've got 4x GeForce RTX 4090's in your computer, CPU inference is going to be the best chance you've got at running 8x22b at top fidelity.
- moffkalast 3y ago
  Correct me if I'm wrong, but in the tests I've run, the matmul optimizations only have an effect if there's no other blas acceleration. If one can at least offload the KV cache to cublas or run with openblas it's not really used, right? At least I didn't see any speedup in with that config when comparing that PR to the main llama.cpp branch.
  - jart 3y ago
    The code that launches my code (see ggml_compute_forward_mul_mat) comes after CLBLAST, Accelerate, and OpenBLAS. The latter take precedence. So if you're not seeing any speedup in enabling them, it's probably because tinyBLAS has reached terms of equality with the BLAS. It's obviously nowhere near as fast as cuBLAS, but maybe PCIE memory transfer overhead explains it. It also really depends on various other factors, like quantization type. For example, the BLAS doesn't support formats like Q4_0 and tinyBLAS does.

noman-land 3y ago

+1 on llamafile. You can point it to a custom model.

jart 3y ago

llamafile author here. I'm downloading Mixtral 8x22b right now. I can't say for certain it'll work until I try it, but let's keep our fingers crossed! If not, we'll be shipping a release as soon as possible that gets it working. My recent work optimizing CPU evaluation https://justine.lol/matmul/ https://justine.lol/matmul/ may have come at just the right time. Mixtral 8x7b always worked best at Q5_K_M and higher, which is 31GB. So unless you've got 4x GeForce RTX 4090's in your computer, CPU inference is going to be the best chance you've got at running 8x22b at top fidelity.

moffkalast 3y ago
Correct me if I'm wrong, but in the tests I've run, the matmul optimizations only have an effect if there's no other blas acceleration. If one can at least offload the KV cache to cublas or run with openblas it's not really used, right? At least I didn't see any speedup in with that config when comparing that PR to the main llama.cpp branch.
- jart 3y ago
  The code that launches my code (see ggml_compute_forward_mul_mat) comes after CLBLAST, Accelerate, and OpenBLAS. The latter take precedence. So if you're not seeing any speedup in enabling them, it's probably because tinyBLAS has reached terms of equality with the BLAS. It's obviously nowhere near as fast as cuBLAS, but maybe PCIE memory transfer overhead explains it. It also really depends on various other factors, like quantization type. For example, the BLAS doesn't support formats like Q4_0 and tinyBLAS does.