CERBERUS/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-13 14:27:00 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

x/imagegen: restore image editing dispatch for flux2 models

Browse source 
Commit 6ddd8862 refactored the imagegen runner but accidentally dropped
the ImageEditModel interface and the conditional dispatch in
handleImageCompletion that routes to GenerateImageWithInputs when input
images are present. This caused Flux2 Klein image-to-image editing
(introduced in v0.15.4) to silently fall back to plain text-to-image
generation in v0.15.6+, ignoring any uploaded reference images.

Restore the ImageEditModel interface (extending ImageModel with
GenerateImageWithInputs) and update handleImageCompletion to decode
req.Images and dispatch to the editing path when images are provided,
falling back to GenerateImage for prompt-only requests.
This commit is contained in:
RajeshKumar11 2026-02-25 08:27:05 +05:30
parent f4f0a4a471
commit 934643af18
1 changed files with 32 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

34
x/imagegen/imagegen.go
View file

@ -6,6 +6,7 @@ import (
"context"
"encoding/json"
"fmt"
"image"
"log/slog"
"net/http"
"sync"
@ -22,6 +23,12 @@ type ImageModel interface {
GenerateImage(ctx context.Context, prompt string, width, height int32, steps int, seed int64, progress func(step, total int)) (*mlx.Array, error)
}
// ImageEditModel extends ImageModel with image editing/conditioning capability.
type ImageEditModel interface {
ImageModel
GenerateImageWithInputs(ctx context.Context, prompt string, width, height int32, steps int, seed int64, inputImages []image.Image, progress func(step, total int)) (*mlx.Array, error)
}
var imageGenMu sync.Mutex
// loadImageModel loads an image generation model.
@ -93,8 +100,31 @@ func (s *server) handleImageCompletion(w http.ResponseWriter, r *http.Request, r
flusher.Flush()
}
// Generate image
img, err := s.imageModel.GenerateImage(ctx, req.Prompt, req.Width, req.Height, req.Steps, req.Seed, progress)
// Decode input images for editing, if provided
var inputImages []image.Image
for i, raw := range req.Images {
decoded, err := DecodeImage(raw)
if err != nil {
http.Error(w, fmt.Sprintf("invalid image %d: %v", i, err), http.StatusBadRequest)
return
}
inputImages = append(inputImages, decoded)
}
// Dispatch: image editing when inputs are present, plain generation otherwise
var img *mlx.Array
var err error
if len(inputImages) > 0 {
editModel, ok := s.imageModel.(ImageEditModel)
if !ok {
http.Error(w, "model does not support image editing", http.StatusBadRequest)
return
}
slog.Info("generating image with input conditioning", "images", len(inputImages))
img, err = editModel.GenerateImageWithInputs(ctx, req.Prompt, req.Width, req.Height, req.Steps, req.Seed, inputImages, progress)
} else {
img, err = s.imageModel.GenerateImage(ctx, req.Prompt, req.Width, req.Height, req.Steps, req.Seed, progress)
}
if err != nil {
// Don't send error for cancellation
if ctx.Err() != nil {