highperfocused/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2025-11-11 19:06:44 +01:00
Code Issues Packages Projects Releases Wiki Activity

batch: use tensors for outputs (#12185)

Browse Source 
this cleans up the model interface slightly without too much impact in
other areas
This commit is contained in:
Michael Yang
2025-09-15 14:33:06 -07:00
committed by GitHub
parent 92b96d54ef
commit 6f7117145f
14 changed files with 27 additions and 37 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
model/input/input.go
View File

@@ -54,10 +54,9 @@ type Batch struct {
// Inputs is the input tokens, including placeholders for multimodal inputs.
Inputs ml.Tensor
// Multimodal is a set of multimodal embeddings previously created by
// EncodeMultimodal, along with an index into Inputs. Unused for text-only
// models or for batches without multimodal elements.
Multimodal []MultimodalIndex
// Outputs are the set of indicies into Inputs for which output data should
// be returned.
Outputs ml.Tensor
// Positions is the position for each Input, relative to its sequence. Equal
// in length to Inputs.
@@ -66,7 +65,8 @@ type Batch struct {
// Sequences is the sequence for each Input. Equal in length to Inputs.
Sequences []int
// Outputs are the set of indicies into Inputs for which output data should
// be returned.
Outputs []int32
// Multimodal is a set of multimodal embeddings previously created by
// EncodeMultimodal, along with an index into Inputs. Unused for text-only
// models or for batches without multimodal elements.
Multimodal []MultimodalIndex
}