jeans/ollama
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

full attn block indexes should be []int32

Browse Source
This commit is contained in:
Bruce MacDonald 2025-05-01 11:21:56 -07:00
parent b3da8a319e
commit ca981c8a49
2 changed files with 12 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
convert/convert_qwen25vl.go
View File

@ -12,18 +12,18 @@ type qwen25VLModel struct {
qwen2Model qwen2Model
VisionModel struct { VisionModel struct {
Depth uint32 `json:"depth"` Depth uint32 `json:"depth"`
HiddenSize uint32 `json:"hidden_size"` HiddenSize uint32 `json:"hidden_size"`
NumHeads uint32 `json:"num_heads"` NumHeads uint32 `json:"num_heads"`
InChannels uint32 `json:"in_chans"` InChannels uint32 `json:"in_chans"`
PatchSize uint32 `json:"patch_size"` PatchSize uint32 `json:"patch_size"`
SpatialMergeSize uint32 `json:"spatial_merge_size"` SpatialMergeSize uint32 `json:"spatial_merge_size"`
SpatialPatchSize uint32 `json:"spatial_patch_size"` SpatialPatchSize uint32 `json:"spatial_patch_size"`
WindowSize uint32 `json:"window_size"` WindowSize uint32 `json:"window_size"`
RMSNormEps float32 `json:"layer_norm_epsilon"` RMSNormEps float32 `json:"layer_norm_epsilon"`
RopeTheta float32 `json:"rope_theta"` RopeTheta float32 `json:"rope_theta"`
FullAttentionBlocks []uint32 `json:"fullatt_block_indexes"` FullAttentionBlocks []int32 `json:"fullatt_block_indexes"`
TemporalPatchSize uint32 `json:"temporal_patch_size"` TemporalPatchSize uint32 `json:"temporal_patch_size"`
} `json:"vision_config"` } `json:"vision_config"`
} }

1
model/models/qwen25vl/model.go
View File

@ -69,7 +69,6 @@ func (m *Model) EncodeMultimodal(ctx ml.Context, multimodalData []byte) (any, er
func (m *Model) PostTokenize(inputs []input.Input) ([]input.Input, error) { func (m *Model) PostTokenize(inputs []input.Input) ([]input.Input, error) {
var result []input.Input var result []input.Input
// Get image token IDs from config
var ( var (
imageToken int32 = 151655 imageToken int32 = 151655
visionStartToken int32 = 151652 visionStartToken int32 = 151652