Instructions to use replicate/moe-new-models with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Kernels
How to use replicate/moe-new-models with Kernels:
# !pip install kernels from kernels import get_kernel # a version (or an explicit revision) is required; see the "Files and versions" tab for the available ones kernel = get_kernel("replicate/moe-new-models", version=1) - Notebooks
- Google Colab
- Kaggle
Commit History
-
Module name has to correspond to repo name
a8ef99b -
Build
031b5c2 -
Fix Llama expert layout
3c64348 -
Add flake.lock
ef18c25 -
Add Llama MoE
6d19dd0
-
feat: update to include rev in kernel for reproducible symbols
drbhfbbc5a8
-
Update for build.toml changes
29fbb6a
-
Update README.md (#1)
Verified signature- Update README.md (7a48e510d3af0b009c9dee005d12ab20d6070e3f) Co-authored-by: Lysandre <lysandre@users.noreply.huggingface.co>72afabe
-
Build
605a216 -
Remove flake.lock
b736673 -
Vendor `w8a8_block_fp8_matmul` and `per_token_group_quant_fp8`
b41d28a -
Build
0de185e -
Normalize some directory names
3d21f81 -
Make functions visible to make a custom `fused_moe`
c2ee5f1
-
Build
6c5a23a -
Sync with upstream
6eaa88c
-
Build (Torch 2.6)
f235406
-
Build
f1e9385 -
Try to avoid fake op registration issues
91deba9
-
Build
30f310f -
Export ops at the top-level
784afd3 -
Update flake lock
f57bdd6
-
Build
2827526 -
Track shared libraries
3ef2fd7 -
Make compatible with latest kernel-builder
1266e46 -
Add MoE configs
fdd8ac0
-
Add MoE kernels from vLLM
TODO: add MoE configs, but this requires a change to the builder.29e93ec -
initial commit
Verified signature8d5a3ce