Datasets:

Overdog
/

MFRL_Data

license: other
license_name: mixed-upstream
license_link: LICENSE.md
task_categories:
  - image-to-image
language:
  - en
tags:
  - image-editing
  - instruction-based-image-editing
  - feedback
  - self-correction
  - benchmark
pretty_name: MFRL Data
size_categories:
  - 10K<n<100K
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*.parquet
      - split: test
        path: data/test-*.parquet
  - config_name: judge
    data_files:
      - split: test
        path: judge/test-*.parquet
dataset_info:
  - config_name: default
    features:
      - name: id
        dtype: string
      - name: edit_type
        dtype: string
      - name: instruction
        dtype: string
      - name: source_image
        dtype: image
      - name: target_image
        dtype: image
      - name: failed_image
        dtype: image
      - name: feedback
        dtype: string
      - name: feedback_problem
        dtype: string
      - name: feedback_correction
        dtype: string
      - name: failed_image_origin
        dtype: string
      - name: failed_image_model
        dtype: string
      - name: failed_image_seed
        dtype: int64
      - name: target_image_model
        dtype: string
      - name: target_human_confirmed
        dtype: bool
      - name: feedback_model
        dtype: string
      - name: source_dataset
        dtype: string
      - name: source_subset
        dtype: string
      - name: image_key
        dtype: string
      - name: split_origin
        dtype: string
      - name: mfrl_id
        dtype: string
      - name: mfrl_category
        dtype: string
  - config_name: judge
    features:
      - name: id
        dtype: string
      - name: edit_id
        dtype: string
      - name: category
        dtype: string
      - name: subcategory
        dtype: string
      - name: instruction
        dtype: string
      - name: source_image
        dtype: image
      - name: candidate_image
        dtype: image
      - name: label
        dtype: string
      - name: reason
        dtype: string
      - name: confidence
        dtype: string
      - name: generator
        dtype: string
      - name: label_source
        dtype: string

MFRL Data

Instruction-based image edits that went wrong, each paired with what should have happened and how to fix it, plus the MFRL-Bench test set.

default / train: 50,642 records. Each has a source image, an edit instruction, a ground-truth result, a failed result with an obvious error, and feedback that names the problem and then gives a corrective instruction.
default / test: 3,000 records, the edit tasks of MFRL-Bench. Each has the source, the instruction and the benchmark's reference result. 2,738 of them also carry a failed image and feedback built the same way as the training data. The others have none (failed_image and feedback are null), because no obvious failure was found.
judge / test: 2,000 images, the judge track of MFRL-Bench: edited images for the test tasks, half labeled success and half fail, each with a written reason.

No overlap between train and test. The 2,738 test tasks that also had a training record were removed from train, and so were 49 more records that show the same photo as a test task (same Open Images ID, same source file, same OmniEdit scene, or a 256-bit image-hash distance of 25 or less).

All images are 1024×768 PNG and are embedded in the Parquet files.

Loading

from datasets import load_dataset

train = load_dataset("Overdog/MFRL_Data", split="train")
test = load_dataset("Overdog/MFRL_Data", split="test")
judge = load_dataset("Overdog/MFRL_Data", "judge", split="test")

r = train[0]
r["source_image"], r["target_image"], r["failed_image"]   # PIL images
r["instruction"], r["feedback"]

Fields (default config)

Field	Type	Meaning
`id`	string	Record ID; the upstream sample it was built from
`edit_type`	string	One of the 9 content-edit classes
`instruction`	string	Edit instruction given to the editor
`source_image`	image (PNG)	Image to edit
`target_image`	image (PNG)	Ground-truth edit result
`failed_image`	image (PNG)	Edit result with an obvious error
`feedback`	string	Problem in the failed image, then the corrective instruction (at most 150 words)
`feedback_problem`	string	First part of `feedback`: what is wrong and where
`feedback_correction`	string	Second part of `feedback`: what the editor should do instead
`failed_image_origin`	string	`original_candidate` (kept from the candidate pool that produced `target_image`) or `regenerated` (one of three new samples; the model is in `failed_image_model`)
`failed_image_model`	string	Model that generated `failed_image`
`failed_image_seed`	int64	Seed of `failed_image`
`target_image_model`	string	Model that generated `target_image`
`target_human_confirmed`	bool	Whether a person confirmed `target_image` (otherwise chosen by an automatic judge)
`feedback_model`	string	Model that judged `failed_image` and wrote `feedback`
`source_dataset`	string	Dataset the source image and instruction come from
`source_subset`	string	Subset within `source_dataset`
`image_key`	string	Source image ID; split train/test by this column
`split_origin`	string	`ours` (train, or a test task with our quad) or `mfrl_bench` (test task without a quad)
`mfrl_id`	string	MFRL-Bench task id (test split only)
`mfrl_category`	string	MFRL-Bench category (test split only; the 4 text classes are one category, `Text`)

Fields (judge config)

Field	Type	Meaning
`id`	string	MFRL-Bench judge item id
`edit_id`	string	MFRL-Bench edit task the image was made for
`category`	string	MFRL-Bench category
`subcategory`	string	Text sub-type, else null
`instruction`	string	Edit instruction
`source_image`	image (PNG)	Image to edit
`candidate_image`	image (PNG)	Edited image to judge
`label`	string	`success` or `fail`
`reason`	string	Labeling judge reason (Claude Haiku 5.5)
`confidence`	string	Labeling judge confidence
`generator`	string	Model that produced the candidate
`label_source`	string	How the label was obtained

Counts

Train by edit type:

Edit type	Records
Change Object Appearance	21,753
Replace Object	8,646
Add Object	7,648
Remove Object	4,919
Complex Edit	2,595
Add Text	1,925
Move Object	1,405
Replace Text	942
Change Text Style	465
Remove Text	344

Test by MFRL-Bench category:

Category	Tasks
Add Object	500
Change Object Appearance	500
Complex Edit	500
Remove Object	500
Replace Object	500
Text	400
Move Object	100

Train by model of the failed image:

Model	Records
FLUX.2-klein-9B	30,923
FLUX.2-klein-4B	9,393
Qwen-Image-Edit-2511-Lightning	5,088
FLUX.3	5,059
Qwen-Image-2.1	179

Train by source dataset:

Source	Records
GPT-Image-Edit-1.5M	37,602
Pico-Banana-400K	8,753
Open Images (TextOCR)	3,244
Open Images	1,043

How it was built

Sources and ground truth. Source images and instructions come from GPT-Image-Edit-1.5M (OmniEdit subset), Pico-Banana-400K and Open Images V6 (with TextOCR for text edits). Several candidate edits were generated per source with FLUX.3 and Qwen-Image-2.1, and an automatic judge (Claude Haiku 5.5) chose a successful one as the target.
Failed image. If a failed candidate from that pool had an obvious error, meaning one a person notices within seconds at normal size, it was kept. Otherwise FLUX.2-klein-9B (4 steps) generated three new results and the judge picked the one with the most obvious error. If none had one, FLUX.2-klein-4B (4 steps) tried three more, and then Qwen-Image-Edit-2511 with the Lightning 4-step LoRA. If all nine had no obvious error, the record was dropped. Small shifts in color, tone or sharpness, and differences of taste on style edits, do not count as errors.
Feedback. The same judge call wrote the feedback: the problem first, then an imperative correction, in words rather than coordinates, at most 150 words. The judge was Claude Haiku 5.5 for most records and Claude Opus 5.5 for part of the text, move and complex-edit records; feedback_model names it per record.
Test set. MFRL-Bench tasks were selected where two independent Claude Haiku 5.5 judgments agreed on every candidate and on the best one, which is the reference. No Pico-Banana-400K image is in the test set.

Limitations

Labels are automatic. No person reviewed the failed images, the feedback, the references or the judge labels.
Generators differ. In train, most failed images come from FLUX.2-klein, while targets come from FLUX.3. A model trained to tell them apart can learn the generator instead (failed_image_model, target_image_model).
Selection bias. Records where every tried model got the edit right are missing, so easy edits are underrepresented.
The reference is one valid result, not the only one.

License

The data combines material under different terms; see LICENSE.md. In particular, records with source_dataset = Pico-Banana-400K are under CC BY-NC-ND 4.0 (non-commercial, no derivatives). All target_image, failed_image and candidate_image files are AI-generated.