Skip to content

Minimax M3 agg #682

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
esmeetu merged 2 commits into from
Jul 28, 2026
+5 −0
Merged

Minimax M3 agg #682

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2f9f8f9
Add --attention_config.indexer_kv_dtype fp8 to MiniMax-M3 NVFP4 variant
faradawn Jul 24, 2026
406b253
Add --attention_config.indexer_kv_dtype fp8 to MiniMax-M3 MXFP8 variant
faradawn Jul 27, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions models/MiniMaxAI/MiniMax-M3.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -112,6 +112,9 @@ variants:
# comfortably, ~4 GPUs for weights alone on Blackwell (B200/B300) or AMD MI350X/MI355X (gfx950)).
vram_minimum_gb: 513
description: "NVIDIA-quantized MXFP8 weights — Blackwell (B200/B300) for native MX tensor cores, and AMD CDNA4 (MI350X/MI355X, gfx950) for native MXFP8 Matrix Cores."
extra_args:
- "--attention_config.indexer_kv_dtype"
- "fp8"
hardware_overrides:
h100:
# 8x H100 80GB (640 GB node) fits the ~513 GB MXFP8 weights but not the
Expand Down Expand Up @@ -152,6 +155,8 @@ variants:
description: "NVIDIA-quantized NVFP4 weights (ModelOpt) for Blackwell (B200/B300) — native FP4 tensor cores, ~half the VRAM of MXFP8."
extra_args:
- "--trust-remote-code"
- "--attention_config.indexer_kv_dtype"
- "fp8"
extra_env:
VLLM_FLOAT32_MATMUL_PRECISION: "high"
mxfp4:
Expand Down