on my way to gmfss support maybe

TNTwise · Oct 7, 2024 · 1bdd54f · 1bdd54f
1 parent a307ad3
commit 1bdd54f
Showing 1 changed file with 26 additions and 0 deletions.
diff --git a/backend/src/InterpolateArchs/DetectInterpolateArch.py b/backend/src/InterpolateArchs/DetectInterpolateArch.py
@@ -39,6 +39,7 @@ def excluded_keys() -> tuple:
             "module.caltime.8.weight",
             "module.caltime.8.bias",
             "module.block4.lastconv.0.bias",
+            "transformer.layers.4.self_attn.merge.weight",
         ]
 
 
@@ -72,6 +73,7 @@ def excluded_keys() -> tuple:
             "module.caltime.8.weight",
             "module.caltime.8.bias",
             "module.block4.lastconv.0.bias",
+            "transformer.layers.4.self_attn.merge.weight",
         ]
 
 
@@ -102,6 +104,7 @@ def excluded_keys() -> tuple:
             "module.caltime.8.weight",
             "module.caltime.8.bias",
             "module.block4.lastconv.0.bias",
+            "transformer.layers.4.self_attn.merge.weight",
         ]
 
 
@@ -122,6 +125,7 @@ def excluded_keys() -> tuple:
             "module.encode.1.weight",
             "module.encode.1.bias",
             "module.block4.lastconv.0.bias",
+            "transformer.layers.4.self_attn.merge.weight",
         ]
 
 
@@ -142,6 +146,7 @@ def excluded_keys() -> tuple:
             "module.encode.1.weight",
             "module.encode.1.bias",
             "module.block4.lastconv.0.bias",
+            "transformer.layers.4.self_attn.merge.weight",
         ]
 
 
@@ -162,6 +167,7 @@ def excluded_keys() -> tuple:
             "module.encode.1.weight",
             "module.encode.1.bias",
             "module.block4.lastconv.0.bias",
+            "transformer.layers.4.self_attn.merge.weight",
         ]
 
 
@@ -181,8 +187,28 @@ def excluded_keys() -> tuple:
             "module.encode.0.bias",
             "module.encode.1.weight",
             "module.encode.1.bias",
+            "transformer.layers.4.self_attn.merge.weight",
         ]
 
+class GMFSS:
+    def __init__():
+        pass
+
+    def __name__():
+        return "rife413"
+
+    def unique_shapes() -> dict:
+        return {"transformer.layers.4.self_attn.merge.weight": "torch.Size([128, 128])"}
+
+    def excluded_keys() -> tuple:
+        return [
+            "module.encode.0.weight",
+            "module.encode.0.bias",
+            "module.encode.1.weight",
+            "module.encode.1.bias",
+        ]
+
+
 
 archs = [RIFE46, RIFE47, RIFE413, RIFE420, RIFE421, RIFE422lite, RIFE425]