MODEL_TENSOR.SSM_DT_NORM has defined twice (#14991)

* MODEL_TENSOR.SSM_DT_NORM has defined twice, and second overwritten the jamba model's layername

* correct order
This commit is contained in:
Csaba Kecskemeti 2025-07-31 07:59:49 -07:00 committed by GitHub
parent 94933c8c2e
commit 36e5fe7bcd
No known key found for this signature in database
GPG key ID: B5690EEEBB952194

View file

@ -616,6 +616,7 @@ class TensorNameMap:
),
MODEL_TENSOR.SSM_DT_NORM: (
"model.layers.layers.{bid}.mixer.dt_norm.weight", # plamo2
"model.layers.{bid}.mamba.dt_layernorm", # jamba
),
@ -645,10 +646,6 @@ class TensorNameMap:
"model.layers.layers.{bid}.mixer.D", # plamo2
),
MODEL_TENSOR.SSM_DT_NORM: (
"model.layers.layers.{bid}.mixer.dt_norm.weight", # plamo2
),
MODEL_TENSOR.SSM_NORM: (
"model.layers.{bid}.mamba.norm", # falcon-h1 granite-hybrid
"backbone.layers.{bid}.mixer.norm", # mamba2