Enable TP for Mamba-based models (sgl-project#14811)

roikoren755 · Prozac614 · commit 4090968664ee · 2025-12-17T03:25:40.000Z
Signed-off-by: Roi Koren &lt;roik@nvidia.com&gt;
diff --git a/python/sglang/srt/layers/attention/mamba/mamba.py b/python/sglang/srt/layers/attention/mamba/mamba.py
@@ -339,7 +339,6 @@ def __init__(
             input_is_parallel=True,
             quant_config=quant_config,
             prefix=f"{prefix}.out_proj",
-            reduce_results=False,
         )
 
         self.norm = Mixer2RMSNormGated(

Original file line number	Diff line number	Diff line change
`@@ -339,7 +339,6 @@ def __init__(`
`339`	`339`	`input_is_parallel=True,`
`340`	`340`	`quant_config=quant_config,`
`341`	`341`	`prefix=f"{prefix}.out_proj",`
`342`		`- reduce_results=False,`
`343`	`342`	`)`
`344`	`343`
`345`	`344`	`self.norm = Mixer2RMSNormGated(`