gpu + cpu

carte_ai/src/carte_model.py

@@ -9,17 +9,24 @@
 def _carte_calculate_attention(
     edge_index: Tensor, query: Tensor, key: Tensor, value: Tensor
 ):
+    ## Fix to work on cpu and gpu provided by Ayoub Kachkach
     # Calculate the scaled-dot product attention
     attention = torch.sum(torch.mul(query[edge_index[0], :], key), dim=1)
     attention = attention / math.sqrt(query.size(1))
     attention = softmax(attention, edge_index[0])
-
+
+    # Ensure `attention` and `value` have the same dtype
+    attention = attention.to(value.dtype)
+
     # Generate the output
     src = torch.mul(attention, value.t()).t()
-
+
+    # Ensure `src` and `query` have the same dtype
+    src = src.to(query.dtype)
+
     # Use torch.index_add_ to replace scatter function
     output = torch.zeros_like(query).index_add_(0, edge_index[0], src)
-
+    
     return output, attention
 
 

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "carte-ai"
-version = "0.0.23"
+version = "0.0.25"
 description = "CARTE-AI: Context Aware Representation of Table Entries for AI"
 readme = { file = "README.md", content-type = "text/markdown" }
 requires-python = ">=3.10.12"