convert: split gate_up bias

convert/convert_gptoss.go

@@ -85,6 +85,17 @@ func (m *gptossModel) Tensors(ts []Tensor) []*ggml.Tensor {
 			case "scales":
 				mxfp4s[name].scales = t
 			}
+		} else if strings.HasSuffix(t.Name(), "gate_up_exps.bias") {
+			out = append(out, slices.Collect(splitDim(t, 1,
+				split{
+					Replacer: strings.NewReplacer("gate_up_exps", "gate_exps"),
+					slices: []tensor.Slice{nil, tensor.S(0, int(t.Shape()[1]), 2)},
+				},
+				split{
+					Replacer: strings.NewReplacer("gate_up_exps", "up_exps"),
+					slices: []tensor.Slice{nil, tensor.S(1, int(t.Shape()[1]), 2)},
+				},
+			))...)
 		} else {
 			out = append(out, &ggml.Tensor{
 				Name:     t.Name(),

convert/tensor.go

@@ -16,7 +16,8 @@ import (
 
 type split struct {
 	*strings.Replacer
-	dim int
+	dim    int
+	slices []tensor.Slice
 
 	// fn is an optional function to apply to the tensor after slicing
 	fn func(tensor.Tensor) (tensor.Tensor, error)
@@ -32,9 +33,12 @@ func splitDim(t Tensor, dim int, splits ...split) iter.Seq[*ggml.Tensor] {
 			shape := slices.Clone(t.Shape())
 			shape[dim] = cmp.Or(uint64(split.dim), shape[dim]/uint64(len(splits)))
 
-			slice := slices.Repeat([]tensor.Slice{nil}, len(shape))
+			slice := split.slices
-			slice[dim] = tensor.S(offset, offset+int(shape[dim]))
+			if len(slice) == 0 {
-			offset += int(shape[dim])
+				slice := slices.Repeat([]tensor.Slice{nil}, len(shape))
+				slice[dim] = tensor.S(offset, offset+int(shape[dim]))
+				offset += int(shape[dim])
+			}
 
 			t.SetRepacker(func(_ string, data []float32, shape []uint64) ([]float32, error) {
 				dims := make([]int, len(shape))