Skip to content

Commit baaf92b

Browse files
seungrokjclaude
andcommitted
[AMD] cap minimaxm3-fp8-mi355x-atom-disagg conc to 256; fix missing newline in models_atom.yaml
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
1 parent 009cff5 commit baaf92b

2 files changed

Lines changed: 3 additions & 3 deletions

File tree

.github/configs/amd-master.yaml

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -2775,7 +2775,7 @@ minimaxm3-fp8-mi355x-atom-disagg:
27752775
osl: 1024
27762776
search-space:
27772777
# 1P1D TP4
2778-
- conc-list: [ 1, 2, 4, 8, 16, 32, 64, 128, 256, 512 ]
2778+
- conc-list: [ 1, 2, 4, 8, 16, 32, 64, 128, 256 ]
27792779
prefill:
27802780
num-worker: 1
27812781
tp: 4
@@ -2795,7 +2795,7 @@ minimaxm3-fp8-mi355x-atom-disagg:
27952795
osl: 1024
27962796
search-space:
27972797
# 1P1D TP4
2798-
- conc-list: [ 1, 2, 4, 8, 16, 32, 64, 128, 256, 512 ]
2798+
- conc-list: [ 1, 2, 4, 8, 16, 32, 64, 128, 256 ]
27992799
prefill:
28002800
num-worker: 1
28012801
tp: 4

benchmarks/multi_node/amd_utils/models_atom.yaml

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -37,4 +37,4 @@ MiniMax-M3-MXFP8:
3737
kv_cache_flags: "--kv_cache_dtype fp8"
3838
tp_dp_flags: "--enable-dp-attention"
3939
ep_dp_flags: "--enable-expert-parallel --enable-dp-attention"
40-
mtp_flags: "--method eagle3 --draft-model Inferact/MiniMax-M3-EAGLE3 --num-speculative-tokens"
40+
mtp_flags: "--method eagle3 --draft-model Inferact/MiniMax-M3-EAGLE3 --num-speculative-tokens"

0 commit comments

Comments
 (0)