matt/ollama37
1
0
Fork 0
mirror of https://github.com/dogkeeper886/ollama37.git synced 2025-12-09 23:37:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

arange

Browse Source
This commit is contained in:
Michael Yang
2025-04-03 10:25:23 -07:00
committed by Michael Yang
parent 1d99451ad7
commit 40b8fdbdca
5 changed files with 42 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
model/models/gemma3/model_vision.go
View File

@@ -92,16 +92,7 @@ func (m *VisionModel) Forward(ctx ml.Context, pixelValues ml.Tensor) ml.Tensor {
hiddenState = hiddenState.Reshape(ctx, numPatches, m.hiddenSize)
hiddenState = hiddenState.Permute(ctx, 1, 0, 2, 3).Contiguous(ctx)
positions := make([]int32, numPatches)
for i := range positions {
positions[i] = int32(i)
}
positionIDs, err := ctx.Input().FromIntSlice(positions, len(positions))
if err != nil {
panic(err)
}
positionIDs := ctx.Arange(0, float32(numPatches), 1, ml.DTypeI32)
hiddenState = hiddenState.Add(ctx, m.PositionEmbedding.Forward(ctx, positionIDs))
for _, layer := range m.Layers {

11
model/models/mllama/model.go
View File

@@ -93,16 +93,7 @@ func (m *Model) EncodeMultimodal(ctx ml.Context, multimodalData []byte) (any, er
return nil, err
}
positions := make([]int32, 1601)
for i := range positions {
positions[i] = int32(i)
}
positionIDs, err := ctx.Input().FromIntSlice(positions, len(positions))
if err != nil {
return nil, err
}
positionIDs := ctx.Arange(0, 1601, 1, ml.DTypeI32)
crossAttentionStates := m.VisionModel.Forward(ctx, pixelValues, positionIDs, aspectRatio)
return m.Projector.Forward(ctx, crossAttentionStates), nil
}