Text-to-image (T2I) self-correction combines a vision-language model (VLM) critic that diagnoses errors and proposes instructions with an image editor that executes them. End-to-end evaluation obscures whether failures arise from incorrect diagnoses or an editor unable to execute valid instructions. We introduce LEDGE (Learning Error Diagnosis in Generation and Editing), a framework that separates these bottlenecks. LEDGE compiles oracle edit instructions from GenEval while holding the generated images and the editor fixed, isolating the editor's correction capability from the critic's diagnostic accuracy. On Stable Diffusion 3.5 Large, oracle instructions raise GenEval from 72.0% to 81.2%, versus 74.1% for the baseline VLM critic. Yet 435 of 644 failed images remain unrepaired, exposing a substantial editor gap. We also use oracle labels from disjoint training prompts to supervise the critic. Oracle supervision raises exact diagnosis from 33.3% (the same VLM with our structured output, untrained) to 70.8%. With the same editor, the resulting critic improves GenEval from 42.2% (uncorrected) to 59.4% on Stable Diffusion 1.5, against 50.9% for the baseline critic, and from 72.0% to 76.0% on Stable Diffusion 3.5 Large, despite training only on Stable Diffusion 1.5. Improvements transfer to unseen object classes, a second editor, and independent T2I-CompBench judges and prompts. A small human evaluation confirms the ranking in estimated net corrections but suggests a smaller gain than detector-based evaluation. LEDGE thus distinguishes two limits of self-correction: knowing what to fix and being able to fix it.
Figures & tables
Figure 1: Overview of LEDGE. The critic lists requirement violations; a fixed renderer turns them into an instruction for a fixed editor. The oracle compiler produces the correct edits from the benchmark’s requirements and the detector’s failure reasons. On test images it is a reference; on training prompts it supervises the critic (blue).
Instructions from
Macro
Micro
Rep.
Reg.
p
None (original)
42.2
40.1
–
–
–
Generic instruction
49.9
48.4
246
63
0.33
Baseline (Qwen)
50.9
49.3
226
23
–
Molmo + rules
52.4
50.9
301
64
0.09
Molmo, oracle-sup.
56.5
55.2
345
12
7×10−13
Ours (oracle-sup.)
59.4
58.3
420
18
6×10−28
Table 1: GenEval on SD1.5 images with the same VAREdit editor and the same 2,212 images. Rep./Reg.: repairs and regressions against the original images; p : exact McNemar test against the baseline critic on the same images. Best non-oracle results in bold.
Figure 2: GenEval overall score on SD1.5 images by instruction source, same editor. Dashed lines mark larger generators without correction. With the oracle-supervised critic, SD1.5 plus correction exceeds SDXL; whether this holds at matched compute is not established.
Figure 3: Editor gap of VAREdit. Bars give the share of failures repaired when VAREdit receives oracle instructions; labels give repairs / failures per task.
Qwen decision
Correct (888)
Failed (1,324)
keep
789 (88.9%)
164 (12.4%)
edit
65 (7.3%)
698 (52.7%)
review
34 (3.8%)
462 (34.9%)
Table 2: Baseline critic decisions against GenEval’s verdict on the original SD1.5 images (left) and its diagnoses of failures against the oracle (right).
Correct images
Failed images
Critic
Keep
Edit ↓
Edit
Exact
Any
Baseline (single error)
88.9
7.3
52.7
–
42.3 †
Structured zero-shot
77.5
22.0
91.4
33.3
43.6
Ours (oracle-sup.)
93.5
6.5
89.0
70.8
79.3
Prompt-only rule
0.0
100.0
100.0
49.1
57.8
Table 3: Diagnosis on official SD1.5 images (888 correct, 1,324 failed) against the oracle, without editing. Exact : same edit set as the oracle; Any : at least one matching edit ( † type-level only, not directly comparable). The gray row ignores the image and is tuned on the test failures, an upper bound for prompt-only critics. Best critic results in bold.
Figure 4: Examples where the baseline critic’s correction fails and the oracle-supervised critic succeeds (same editor). Each row is one example; columns show the original image, the baseline critic’s result and ours, with GenEval’s verdict as the frame color (green: pass, red: fail) and the instruction below each edited image. Examples are hand-selected from the images whose original fails GenEval, whose baseline-critic correction still fails and whose oracle-supervised correction passes (258 on SD1.5, 63 on SD3.5).
Task
Orig.
Qwen
Molmo
Ours
Oracle
Closed
Single object
95.9
96.6
95.0
97.8
99.4
44%
Two objects
34.3
58.1
53.0
69.7
76.5
63%
Counting
37.2
40.0
45.0
43.8
49.1
41%
Colors
73.7
81.9
78.7
84.0
91.5
22%
Position
4.5
10.8
28.5
29.8
32.5
87%
Color attr.
7.8
18.3
14.3
31.5
37.0
71%
Table 4: Per-task accuracy on SD1.5 after editing. Qwen: baseline critic; Molmo: Molmo + rules; closed: share of the baseline-to-oracle gap closed by the oracle-supervised critic. Bold marks the best non-oracle result for counting and for the macro average.
Instructions from
Macro
Micro
Rep.
Reg.
p
None (SD3.5-L)
72.0
70.9
–
–
–
Baseline (Qwen)
74.1
73.1
59
11
4×10−9
Ours (oracle-sup.)
76.0
75.1
109
16
3×10−18
Oracle (reference)
81.2
80.3
209
0
2×10−63
Table 5: Results on SD3.5-Large (downscaled to 5122 ) with the same VAREdit editor. The oracle-supervised critic was trained only on SD1.5 images. p : exact McNemar test against the original images. Rep./Reg.: repairs and regressions. Best non-oracle results in bold.
Instructions
Color
Object
Spatial
Numer.
Mean
None (original)
50.75
68.63
1.46
68.95
47.45
Baseline critic
61.68
79.80
3.53
70.56
53.89
Ours
75.81
85.52
10.36
72.98
61.17
Oracle (GenEval)
77.07
86.58
10.32
72.18
61.54
Table 6: Re-scoring GenEval images with T2I-CompBench’s judges (SD1.5, VAREdit, 2,016 images; scores ×100 ). Color: GenEval’s colors and color attribution; object: single and two objects (BLIP-VQA); spatial: position; numer.: counting (UniDet). Mean: unweighted mean of the four columns. Oracle instructions come from GenEval, not from these judges. Best results in bold.
Editor
Instructions
Macro
Micro
Rep.
Reg.
VAREdit
Baseline
50.9
49.3
226
23
VAREdit
Ours
59.4
58.3
420
18
VAREdit
Oracle
64.3
63.3
512
0
FLUX.1 Kontext
Baseline
51.6
49.9
231
15
FLUX.1 Kontext
Ours
61.5
60.3
450
5
FLUX.1 Kontext
Oracle
64.9
63.7
522
0
Table 7: A second editor on SD1.5 (GenEval, same 2,212 images and instruction files). Rep./Reg.: repairs and regressions against the original images.
Appendix figures & tables11 assets
Supplementary material from the paper’s appendix.
Appendix
Generator
Macro
Micro
Single
Two obj.
Count
Colors
Position
Color attr.
SD1.5
42.23
40.14
95.94
34.34
37.19
73.67
4.50
7.75
SDXL
54.36
52.94
99.38
74.24
39.06
83.24
11.00
19.25
FLUX.1-dev
66.75
65.19
98.75
81.06
74.06
77.39
21.25
48.00
SD3.5-Large
70.99
69.85
99.06
89.65
70.00
83.24
26.00
58.00
SD3.5-Large, 5122 copy
72.02
70.89
99.06
89.90
71.88
84.04
27.25
60.00
Appendix
Table 8: Official GenEval (553 prompts × 4 samples) for each generator at native resolution.
Pilot tag
Images
Keep
Edit
Two edits
Position
448
43
405
0
Counting
376
37
339
175
Color attribution
322
28
294
103
Color + position
272
4
268
89
Colors
74
62
12
0
Two objects
38
17
21
0
Appendix
Table 9: Oracle labels on pilot training images (SD1.5, corrected metadata).
3 epochs, 717 optimizer steps; adapters saved every 50 steps
Selection
Best DEV exact match with ≤ 10% of correct DEV images edited → step 400
Appendix
Table 10: Critic fine-tuning hyperparameters.
Figure 6: Checkpoint selection on pilot DEV. Diagnosis improves steadily with training (top), but after step 400 the oracle-supervised critic edits more correct images (bottom). The rule fixed in advance selects step 400. With 40 correct DEV images, differences of one or two images are within noise.
Figure 7: Per-task GenEval accuracy on SD1.5 after editing, same editor (the data of Table 4 ).
Figure 8: Exact diagnosis by task on official SD1.5 failures. Training most improves color attribution, position and two-object diagnosis. Colors improves little; it is the task least represented in training (37 of 765 pilot prompts).
Figure 9: Per-task accuracy on SD3.5-Large at 512px. The oracle-supervised critic transfers from SD1.5 without retraining, recovering 44% of the oracle’s gain; position transfers best (62%).
Instructions
Color
Numer.
Spatial
Mean
None (original)
36.87
45.50
11.43
31.27
Baseline critic
48.88
45.89
12.27
35.68
Ours
66.38
52.73
18.40
45.83
Appendix
Table 11: T2I-CompBench’s own prompts (SD1.5, VAREdit, 1,800 images; scores ×100 ). Against the original images, the baseline critic improves 325 images and worsens 142 ( p=1.4×10−17 , exact sign test) and ours 636 and 159 ( p=2.7×10−68 ); paired, ours improves 577 images over the baseline critic and worsens 266 ( p=3.4×10−27 ). Best results in bold.
Method
Macro
Micro
Rep.
Reg.
Single (6)
Two obj. (75)
Count. (58)
Colors (30)
Pos. (106)
C. attr. (99)
None (original)
39.7
37.7
–
–
–
–
–
–
–
–
Oracle-flagged resample
49.7
47.7
60
0
66.7
26.7
25.9
40.0
2.8
6.1
Baseline critic, resample
43.6
41.5
33
10
0.0
13.3
22.4
20.0
0.9
3.0
Ours, resample
45.8
43.8
43
6
33.3
22.7
22.4
13.3
2.8
4.0
Baseline critic, edit
49.0
47.5
67
8
33.3
42.7
12.1
40.0
5.7
8.1
Ours, edit
57.6
56.7
118
4
50.0
57.3
10.3
40.0
32.1
20.2
Appendix
Table 12: Resampling vs. editing on a 600-image SD1.5 sample (374 failures). Rep./Reg.: repairs and regressions against the original images. Right: share of failures repaired per task (number of failures in parentheses).
Category
SD1.5
SD3.5
Unrepaired / failures
812 / 1,324
435 / 644
Failed insertion
56.3%
29.0%
Incorrect attribute modification
17.5%
14.7%
Failed move
18.0%
48.7%
Unintended change
9.5%
7.6%
Appendix
Table 13: Unrepaired images under oracle instructions (VAREdit), by failure category. Categories overlap; percentages are of unrepaired images.
Split
Orig.
Base.
Molmo
Ours
Oracle
Closed
Unseen (1,428)
41.7
50.3
55.3
59.1
64.4
63%
Mixed (464)
15.5
30.2
40.4
43.2
47.8
74%
Seen (320)
42.1
49.8
55.2
60.6
65.8
67%
All (2,212)
42.2
50.9
56.5
59.4
64.3
63%
Appendix
Table 14: GenEval macro accuracy by object split (SD1.5, VAREdit). Base.: baseline critic; Molmo: oracle-supervised Molmo (QLoRA); closed: share of the baseline-to-oracle gap closed by the oracle-supervised critic.