saiteja33 commited on
Commit
c0fb38c
·
verified ·
1 Parent(s): 4c31e93

Upload folder using huggingface_hub (part 5)

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_37c514cdd8e2/layers/layer_01.png +3 -0
  2. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_37c514cdd8e2/record.json +62 -0
  3. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_454a6be8e58d/layers/layer_00.png +3 -0
  4. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_454a6be8e58d/layers/layer_01.png +3 -0
  5. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_454a6be8e58d/record.json +63 -0
  6. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/layers/layer_00.png +3 -0
  7. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/layers/layer_01.png +3 -0
  8. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/record.json +61 -0
  9. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_549f7f9da427/layers/layer_00.png +3 -0
  10. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_549f7f9da427/layers/layer_01.png +3 -0
  11. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_549f7f9da427/record.json +63 -0
  12. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_68980ef803ab/layers/layer_00.png +3 -0
  13. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_68980ef803ab/layers/layer_01.png +3 -0
  14. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_68980ef803ab/record.json +64 -0
  15. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_74fc1c35c473/layers/layer_00.png +3 -0
  16. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_74fc1c35c473/layers/layer_01.png +3 -0
  17. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_74fc1c35c473/record.json +64 -0
  18. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/layers/layer_00.png +3 -0
  19. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/layers/layer_01.png +3 -0
  20. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/record.json +63 -0
  21. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/layers/layer_00.png +3 -0
  22. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/layers/layer_01.png +3 -0
  23. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/record.json +61 -0
  24. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/layers/layer_00.png +3 -0
  25. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/layers/layer_01.png +3 -0
  26. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/record.json +61 -0
  27. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/layers/layer_00.png +3 -0
  28. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/layers/layer_01.png +3 -0
  29. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/record.json +61 -0
  30. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_b6b750f3837b/layers/layer_00.png +3 -0
  31. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_b6b750f3837b/layers/layer_01.png +3 -0
  32. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_b6b750f3837b/record.json +62 -0
  33. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/layers/layer_00.png +3 -0
  34. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/layers/layer_01.png +3 -0
  35. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/record.json +61 -0
  36. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/layers/layer_00.png +3 -0
  37. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/layers/layer_01.png +3 -0
  38. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/record.json +62 -0
  39. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/layers/layer_00.png +3 -0
  40. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/layers/layer_01.png +3 -0
  41. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/record.json +62 -0
  42. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_db75b0748b7e/layers/layer_00.png +3 -0
  43. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_db75b0748b7e/layers/layer_01.png +3 -0
  44. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_db75b0748b7e/record.json +61 -0
  45. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f5a708c509ca/layers/layer_00.png +3 -0
  46. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f5a708c509ca/layers/layer_01.png +3 -0
  47. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f5a708c509ca/record.json +61 -0
  48. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/layers/layer_00.png +3 -0
  49. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/layers/layer_01.png +3 -0
  50. vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/record.json +61 -0
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_37c514cdd8e2/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: c1a032178b94b15e23b07f431cfef7b552a1f5daeb1d97804382b4f43cdd6731
  • Pointer size: 131 Bytes
  • Size of remote file: 388 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_37c514cdd8e2/record.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_37c514cdd8e2",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic sign with 'GATE 42' in a clean, high-contrast font on a white board, clear foreground separation",
7
+ "target_evidence": "text 'GATE 42' on sign",
8
+ "expected_visual_elements": [
9
+ "sign",
10
+ "white background",
11
+ "clean font"
12
+ ],
13
+ "negative_prompt": "faded text, cluttered background, real logos, low contrast",
14
+ "diversity_tags": [
15
+ "sign",
16
+ "white background"
17
+ ],
18
+ "llm_model_id": "Qwen/Qwen3-8B",
19
+ "prompt_hash": "37c514cdd8e2",
20
+ "rule_valid": true,
21
+ "rule_issues": [],
22
+ "rule_score": 1.0,
23
+ "opening_key": "synthetic photorealistic sign with gate",
24
+ "max_similarity_to_other_prompt": 0.5217,
25
+ "diversity_score": 0.4783,
26
+ "repeated_opening": false,
27
+ "review": {
28
+ "prompt_id": "ocr_heavy_d5_37c514cdd8e2",
29
+ "accept": true,
30
+ "quality": 5,
31
+ "decomposition_fit": 5,
32
+ "diversity": 4,
33
+ "issues": []
34
+ },
35
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
36
+ "final_rank_score": 0.97,
37
+ "status": "qwen_layered_done",
38
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
39
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
40
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
41
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_37c514cdd8e2/image.png",
42
+ "image_width": 1024,
43
+ "image_height": 1024,
44
+ "num_inference_steps": 28,
45
+ "guidance_scale": 3.5,
46
+ "seed": 27344877,
47
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
48
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
49
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
50
+ "qwen_layer_count_requested": 2,
51
+ "qwen_layer_count_returned": 2,
52
+ "qwen_layer_paths": [
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_37c514cdd8e2/layers/layer_00.png",
54
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_37c514cdd8e2/layers/layer_01.png"
55
+ ],
56
+ "qwen_resolution": 640,
57
+ "qwen_num_inference_steps": 30,
58
+ "qwen_true_cfg_scale": 4.0,
59
+ "qwen_seed": 9764502,
60
+ "qwen_elapsed_seconds": 23.307,
61
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_37c514cdd8e2/record.json"
62
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_454a6be8e58d/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: b51fcab9e77131b459a078839099b9c37f84771ea871e725d63e084fddc1df48
  • Pointer size: 131 Bytes
  • Size of remote file: 385 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_454a6be8e58d/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 035a8b9d0241a9514481c928e38857855486de4512fd3a39313d5cc6ae5adb67
  • Pointer size: 131 Bytes
  • Size of remote file: 397 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_454a6be8e58d/record.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_454a6be8e58d",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic notebook page with clear foreground separation, containing the text 'UNIT 12' in bold black font on a white background, surrounded by clean lines and sharp focus",
7
+ "target_evidence": "text region with 'UNIT 12'",
8
+ "expected_visual_elements": [
9
+ "notebook page",
10
+ "bold black text",
11
+ "white background"
12
+ ],
13
+ "negative_prompt": "blurry text, low contrast, real brands, watermarks, celebrities, famous logos",
14
+ "diversity_tags": [
15
+ "notebook",
16
+ "bold font",
17
+ "white background"
18
+ ],
19
+ "llm_model_id": "Qwen/Qwen3-8B",
20
+ "prompt_hash": "454a6be8e58d",
21
+ "rule_valid": true,
22
+ "rule_issues": [],
23
+ "rule_score": 1.0,
24
+ "opening_key": "synthetic photorealistic notebook page with",
25
+ "max_similarity_to_other_prompt": 0.75,
26
+ "diversity_score": 0.05,
27
+ "repeated_opening": true,
28
+ "review": {
29
+ "prompt_id": "ocr_heavy_d5_454a6be8e58d",
30
+ "accept": true,
31
+ "quality": 5,
32
+ "decomposition_fit": 5,
33
+ "diversity": 4,
34
+ "issues": []
35
+ },
36
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
37
+ "final_rank_score": 0.97,
38
+ "status": "qwen_layered_done",
39
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
40
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
41
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
42
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_454a6be8e58d/image.png",
43
+ "image_width": 1024,
44
+ "image_height": 1024,
45
+ "num_inference_steps": 28,
46
+ "guidance_scale": 3.5,
47
+ "seed": 22820619,
48
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
49
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
50
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
51
+ "qwen_layer_count_requested": 2,
52
+ "qwen_layer_count_returned": 2,
53
+ "qwen_layer_paths": [
54
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_454a6be8e58d/layers/layer_00.png",
55
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_454a6be8e58d/layers/layer_01.png"
56
+ ],
57
+ "qwen_resolution": 640,
58
+ "qwen_num_inference_steps": 30,
59
+ "qwen_true_cfg_scale": 4.0,
60
+ "qwen_seed": 1151137,
61
+ "qwen_elapsed_seconds": 23.35,
62
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_454a6be8e58d/record.json"
63
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: bff64ba9883a609f365b490a083f3e7e134a9883c31b3071a5d669225d156dde
  • Pointer size: 131 Bytes
  • Size of remote file: 336 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 8a379780c6a62b6cde76ec9c2a7e4ae8a5f71e04bf769790ef8e144dc504e0f9
  • Pointer size: 131 Bytes
  • Size of remote file: 453 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/record.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_5278b7ef8c6a",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic screen displaying sharp focus and clear foreground, text 'ORBIT' in white on a dark background, part of a dashboard",
7
+ "target_evidence": "text 'ORBIT' on screen",
8
+ "expected_visual_elements": [
9
+ "screen",
10
+ "dashboard"
11
+ ],
12
+ "negative_prompt": "faded text, low contrast, real brand logos, watermarks",
13
+ "diversity_tags": [
14
+ "screen",
15
+ "dashboard"
16
+ ],
17
+ "llm_model_id": "Qwen/Qwen3-8B",
18
+ "prompt_hash": "5278b7ef8c6a",
19
+ "rule_valid": true,
20
+ "rule_issues": [],
21
+ "rule_score": 1.0,
22
+ "opening_key": "synthetic photorealistic screen displaying sharp",
23
+ "max_similarity_to_other_prompt": 0.6,
24
+ "diversity_score": 0.2,
25
+ "repeated_opening": true,
26
+ "review": {
27
+ "prompt_id": "ocr_heavy_d5_5278b7ef8c6a",
28
+ "accept": true,
29
+ "quality": 5,
30
+ "decomposition_fit": 5,
31
+ "diversity": 4,
32
+ "issues": []
33
+ },
34
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
35
+ "final_rank_score": 0.97,
36
+ "status": "qwen_layered_done",
37
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
38
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
39
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
40
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/image.png",
41
+ "image_width": 1024,
42
+ "image_height": 1024,
43
+ "num_inference_steps": 28,
44
+ "guidance_scale": 3.5,
45
+ "seed": 22310793,
46
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
47
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
48
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
49
+ "qwen_layer_count_requested": 2,
50
+ "qwen_layer_count_returned": 2,
51
+ "qwen_layer_paths": [
52
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/layers/layer_00.png",
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/layers/layer_01.png"
54
+ ],
55
+ "qwen_resolution": 640,
56
+ "qwen_num_inference_steps": 30,
57
+ "qwen_true_cfg_scale": 4.0,
58
+ "qwen_seed": 7937327,
59
+ "qwen_elapsed_seconds": 23.363,
60
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_5278b7ef8c6a/record.json"
61
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_549f7f9da427/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: 01b3f0401781e8e4c8de16cdfe347fe10f90e1c4b332ab535e2559056e760b90
  • Pointer size: 131 Bytes
  • Size of remote file: 389 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_549f7f9da427/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: a75186efa8a840625c29a154fd699173c14eae4d5978c9442e80abb3f7e91718
  • Pointer size: 131 Bytes
  • Size of remote file: 386 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_549f7f9da427/record.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_549f7f9da427",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic notebook page with 'BLUE-7' written in pencil, visible on a clean white sheet, sharp focus",
7
+ "target_evidence": "BLUE-7",
8
+ "expected_visual_elements": [
9
+ "notebook page",
10
+ "pencil writing",
11
+ "clean white sheet"
12
+ ],
13
+ "negative_prompt": "smudged text, real brands, watermarks, blurred edges",
14
+ "diversity_tags": [
15
+ "notebook page",
16
+ "pencil writing",
17
+ "clean white sheet"
18
+ ],
19
+ "llm_model_id": "Qwen/Qwen3-8B",
20
+ "prompt_hash": "549f7f9da427",
21
+ "rule_valid": true,
22
+ "rule_issues": [],
23
+ "rule_score": 1.0,
24
+ "opening_key": "synthetic photorealistic notebook page with",
25
+ "max_similarity_to_other_prompt": 0.4815,
26
+ "diversity_score": 0.3185,
27
+ "repeated_opening": true,
28
+ "review": {
29
+ "prompt_id": "ocr_heavy_d5_549f7f9da427",
30
+ "accept": true,
31
+ "quality": 5,
32
+ "decomposition_fit": 5,
33
+ "diversity": 4,
34
+ "issues": []
35
+ },
36
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
37
+ "final_rank_score": 0.97,
38
+ "status": "qwen_layered_done",
39
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
40
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
41
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
42
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_549f7f9da427/image.png",
43
+ "image_width": 1024,
44
+ "image_height": 1024,
45
+ "num_inference_steps": 28,
46
+ "guidance_scale": 3.5,
47
+ "seed": 20324232,
48
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
49
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
50
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
51
+ "qwen_layer_count_requested": 2,
52
+ "qwen_layer_count_returned": 2,
53
+ "qwen_layer_paths": [
54
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_549f7f9da427/layers/layer_00.png",
55
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_549f7f9da427/layers/layer_01.png"
56
+ ],
57
+ "qwen_resolution": 640,
58
+ "qwen_num_inference_steps": 30,
59
+ "qwen_true_cfg_scale": 4.0,
60
+ "qwen_seed": 6696071,
61
+ "qwen_elapsed_seconds": 23.267,
62
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_549f7f9da427/record.json"
63
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_68980ef803ab/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: fb229c926582e43ab03efb92fdf04c6d39674fae8d0844277e8bb8bddb1081fc
  • Pointer size: 131 Bytes
  • Size of remote file: 592 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_68980ef803ab/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: a05acbe1b083676538f6973c3814445c9f8fa6f335205502c4c137ecf4697c48
  • Pointer size: 131 Bytes
  • Size of remote file: 349 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_68980ef803ab/record.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_68980ef803ab",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic crate with high contrast lighting, featuring the text 'TAXI 5' on a sticker with clean boundaries and sharp focus, placed near a metal tool",
7
+ "target_evidence": "text region with 'TAXI 5'",
8
+ "expected_visual_elements": [
9
+ "crate",
10
+ "sticker",
11
+ "high contrast lighting",
12
+ "metal tool"
13
+ ],
14
+ "negative_prompt": "blurry text, low contrast, real brands, watermarks, celebrities, famous logos",
15
+ "diversity_tags": [
16
+ "crate",
17
+ "sticker",
18
+ "metal tool"
19
+ ],
20
+ "llm_model_id": "Qwen/Qwen3-8B",
21
+ "prompt_hash": "68980ef803ab",
22
+ "rule_valid": true,
23
+ "rule_issues": [],
24
+ "rule_score": 1.0,
25
+ "opening_key": "synthetic photorealistic crate with high",
26
+ "max_similarity_to_other_prompt": 0.6154,
27
+ "diversity_score": 0.3846,
28
+ "repeated_opening": false,
29
+ "review": {
30
+ "prompt_id": "ocr_heavy_d5_68980ef803ab",
31
+ "accept": true,
32
+ "quality": 5,
33
+ "decomposition_fit": 5,
34
+ "diversity": 4,
35
+ "issues": []
36
+ },
37
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
38
+ "final_rank_score": 0.97,
39
+ "status": "qwen_layered_done",
40
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
41
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
42
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
43
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_68980ef803ab/image.png",
44
+ "image_width": 1024,
45
+ "image_height": 1024,
46
+ "num_inference_steps": 28,
47
+ "guidance_scale": 3.5,
48
+ "seed": 21726815,
49
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
50
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
51
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
52
+ "qwen_layer_count_requested": 2,
53
+ "qwen_layer_count_returned": 2,
54
+ "qwen_layer_paths": [
55
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_68980ef803ab/layers/layer_00.png",
56
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_68980ef803ab/layers/layer_01.png"
57
+ ],
58
+ "qwen_resolution": 640,
59
+ "qwen_num_inference_steps": 30,
60
+ "qwen_true_cfg_scale": 4.0,
61
+ "qwen_seed": 9074918,
62
+ "qwen_elapsed_seconds": 23.352,
63
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_68980ef803ab/record.json"
64
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_74fc1c35c473/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: 4955efebc98aabe10a2b7964fe171d07d23b63ef773e8d0b1231da645d51f975
  • Pointer size: 131 Bytes
  • Size of remote file: 540 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_74fc1c35c473/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 5aa47b0b2cb2f90d4faeb12ccdc869e0325422b14660da1656c31950de3928a3
  • Pointer size: 131 Bytes
  • Size of remote file: 327 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_74fc1c35c473/record.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_74fc1c35c473",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic crate with 'RIVER 9' printed on a label, clear boundaries, synthetic lighting, visible from the side",
7
+ "target_evidence": "RIVER 9",
8
+ "expected_visual_elements": [
9
+ "crate",
10
+ "label",
11
+ "synthetic lighting",
12
+ "side view"
13
+ ],
14
+ "negative_prompt": "faded text, real brands, watermarks, low resolution",
15
+ "diversity_tags": [
16
+ "crate",
17
+ "label",
18
+ "side view"
19
+ ],
20
+ "llm_model_id": "Qwen/Qwen3-8B",
21
+ "prompt_hash": "74fc1c35c473",
22
+ "rule_valid": true,
23
+ "rule_issues": [],
24
+ "rule_score": 1.0,
25
+ "opening_key": "synthetic photorealistic crate with river",
26
+ "max_similarity_to_other_prompt": 0.4,
27
+ "diversity_score": 0.4,
28
+ "repeated_opening": true,
29
+ "review": {
30
+ "prompt_id": "ocr_heavy_d5_74fc1c35c473",
31
+ "accept": true,
32
+ "quality": 5,
33
+ "decomposition_fit": 5,
34
+ "diversity": 4,
35
+ "issues": []
36
+ },
37
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
38
+ "final_rank_score": 0.97,
39
+ "status": "qwen_layered_done",
40
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
41
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
42
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
43
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_74fc1c35c473/image.png",
44
+ "image_width": 1024,
45
+ "image_height": 1024,
46
+ "num_inference_steps": 28,
47
+ "guidance_scale": 3.5,
48
+ "seed": 29727816,
49
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
50
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
51
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
52
+ "qwen_layer_count_requested": 2,
53
+ "qwen_layer_count_returned": 2,
54
+ "qwen_layer_paths": [
55
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_74fc1c35c473/layers/layer_00.png",
56
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_74fc1c35c473/layers/layer_01.png"
57
+ ],
58
+ "qwen_resolution": 640,
59
+ "qwen_num_inference_steps": 30,
60
+ "qwen_true_cfg_scale": 4.0,
61
+ "qwen_seed": 7202318,
62
+ "qwen_elapsed_seconds": 23.375,
63
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_74fc1c35c473/record.json"
64
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: 1f43ef0e7207bb843a6d1173fdad5247053082a2daa67d2b93678246593f9751
  • Pointer size: 131 Bytes
  • Size of remote file: 305 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 8f7ac0647789d59b6057608654d54d904cba9b08234f9c596241971d85f72d4f
  • Pointer size: 131 Bytes
  • Size of remote file: 383 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/record.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_9c3ac4f01e0a",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic screen displaying 'UNIT 12' with high contrast text on a dark background, clean boundaries, sharp focus",
7
+ "target_evidence": "UNIT 12",
8
+ "expected_visual_elements": [
9
+ "screen",
10
+ "dark background",
11
+ "high contrast text"
12
+ ],
13
+ "negative_prompt": "blurry text, low contrast, real brands, watermarks, celebrities",
14
+ "diversity_tags": [
15
+ "screen",
16
+ "dark background",
17
+ "high contrast"
18
+ ],
19
+ "llm_model_id": "Qwen/Qwen3-8B",
20
+ "prompt_hash": "9c3ac4f01e0a",
21
+ "rule_valid": true,
22
+ "rule_issues": [],
23
+ "rule_score": 1.0,
24
+ "opening_key": "synthetic photorealistic screen displaying unit",
25
+ "max_similarity_to_other_prompt": 1.0,
26
+ "diversity_score": 0.0,
27
+ "repeated_opening": true,
28
+ "review": {
29
+ "prompt_id": "ocr_heavy_d5_9c3ac4f01e0a",
30
+ "accept": true,
31
+ "quality": 5,
32
+ "decomposition_fit": 5,
33
+ "diversity": 4,
34
+ "issues": []
35
+ },
36
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
37
+ "final_rank_score": 0.97,
38
+ "status": "qwen_layered_done",
39
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
40
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
41
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
42
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/image.png",
43
+ "image_width": 1024,
44
+ "image_height": 1024,
45
+ "num_inference_steps": 28,
46
+ "guidance_scale": 3.5,
47
+ "seed": 21934490,
48
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
49
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
50
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
51
+ "qwen_layer_count_requested": 2,
52
+ "qwen_layer_count_returned": 2,
53
+ "qwen_layer_paths": [
54
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/layers/layer_00.png",
55
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/layers/layer_01.png"
56
+ ],
57
+ "qwen_resolution": 640,
58
+ "qwen_num_inference_steps": 30,
59
+ "qwen_true_cfg_scale": 4.0,
60
+ "qwen_seed": 1953212,
61
+ "qwen_elapsed_seconds": 23.38,
62
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_9c3ac4f01e0a/record.json"
63
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: ac3a993322600c80a1c3be391170ed5972476d465b8dcc72b49228314fc3a27a
  • Pointer size: 131 Bytes
  • Size of remote file: 561 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 137ec41014695a35fab990f125035f5b4ab589404604eca9ad37d7a1b13735b0
  • Pointer size: 131 Bytes
  • Size of remote file: 458 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/record.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_a1d7ea50f51f",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic menu board with clean boundaries and high contrast, text 'MESA' in yellow on a black background, displayed in a restaurant",
7
+ "target_evidence": "text 'MESA' on menu board",
8
+ "expected_visual_elements": [
9
+ "menu board",
10
+ "yellow text"
11
+ ],
12
+ "negative_prompt": "blurry text, low contrast, real brand logos, watermarks",
13
+ "diversity_tags": [
14
+ "menu board",
15
+ "restaurant"
16
+ ],
17
+ "llm_model_id": "Qwen/Qwen3-8B",
18
+ "prompt_hash": "a1d7ea50f51f",
19
+ "rule_valid": true,
20
+ "rule_issues": [],
21
+ "rule_score": 1.0,
22
+ "opening_key": "synthetic photorealistic menu board with",
23
+ "max_similarity_to_other_prompt": 0.5185,
24
+ "diversity_score": 0.2815,
25
+ "repeated_opening": true,
26
+ "review": {
27
+ "prompt_id": "ocr_heavy_d5_a1d7ea50f51f",
28
+ "accept": true,
29
+ "quality": 5,
30
+ "decomposition_fit": 5,
31
+ "diversity": 4,
32
+ "issues": []
33
+ },
34
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
35
+ "final_rank_score": 0.97,
36
+ "status": "qwen_layered_done",
37
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
38
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
39
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
40
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/image.png",
41
+ "image_width": 1024,
42
+ "image_height": 1024,
43
+ "num_inference_steps": 28,
44
+ "guidance_scale": 3.5,
45
+ "seed": 27988147,
46
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
47
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
48
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
49
+ "qwen_layer_count_requested": 2,
50
+ "qwen_layer_count_returned": 2,
51
+ "qwen_layer_paths": [
52
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/layers/layer_00.png",
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/layers/layer_01.png"
54
+ ],
55
+ "qwen_resolution": 640,
56
+ "qwen_num_inference_steps": 30,
57
+ "qwen_true_cfg_scale": 4.0,
58
+ "qwen_seed": 499786,
59
+ "qwen_elapsed_seconds": 23.576,
60
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a1d7ea50f51f/record.json"
61
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: bce020da324b2b6eda7c1b7716aca591158ba082a54fbe731ddfd711ea2152a0
  • Pointer size: 131 Bytes
  • Size of remote file: 469 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 0e1847de08ce831ec4f8efd6cb86d636b08918e5e676284760001ad97743882f
  • Pointer size: 131 Bytes
  • Size of remote file: 301 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/record.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_a9deb5ced6e0",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic screen displaying sharp focus and clear foreground, text 'LIMA' in green on a dark background, embedded in a control panel",
7
+ "target_evidence": "text 'LIMA' on screen",
8
+ "expected_visual_elements": [
9
+ "screen",
10
+ "control panel"
11
+ ],
12
+ "negative_prompt": "faded text, low contrast, real brand logos, watermarks",
13
+ "diversity_tags": [
14
+ "screen",
15
+ "control panel"
16
+ ],
17
+ "llm_model_id": "Qwen/Qwen3-8B",
18
+ "prompt_hash": "a9deb5ced6e0",
19
+ "rule_valid": true,
20
+ "rule_issues": [],
21
+ "rule_score": 1.0,
22
+ "opening_key": "synthetic photorealistic screen displaying sharp",
23
+ "max_similarity_to_other_prompt": 0.6,
24
+ "diversity_score": 0.2,
25
+ "repeated_opening": true,
26
+ "review": {
27
+ "prompt_id": "ocr_heavy_d5_a9deb5ced6e0",
28
+ "accept": true,
29
+ "quality": 5,
30
+ "decomposition_fit": 5,
31
+ "diversity": 4,
32
+ "issues": []
33
+ },
34
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
35
+ "final_rank_score": 0.97,
36
+ "status": "qwen_layered_done",
37
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
38
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
39
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
40
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/image.png",
41
+ "image_width": 1024,
42
+ "image_height": 1024,
43
+ "num_inference_steps": 28,
44
+ "guidance_scale": 3.5,
45
+ "seed": 23580959,
46
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
47
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
48
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
49
+ "qwen_layer_count_requested": 2,
50
+ "qwen_layer_count_returned": 2,
51
+ "qwen_layer_paths": [
52
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/layers/layer_00.png",
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/layers/layer_01.png"
54
+ ],
55
+ "qwen_resolution": 640,
56
+ "qwen_num_inference_steps": 30,
57
+ "qwen_true_cfg_scale": 4.0,
58
+ "qwen_seed": 5119986,
59
+ "qwen_elapsed_seconds": 23.323,
60
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_a9deb5ced6e0/record.json"
61
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: 20454b27551f78c4c6af92d6caaf3448de5c92d1ecba1d5a52565019c32742a0
  • Pointer size: 131 Bytes
  • Size of remote file: 591 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 32d4a43592cfbd65bb31a18f2928caa62d69b1c7a0684ee5924562a9761b276f
  • Pointer size: 131 Bytes
  • Size of remote file: 355 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/record.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_aa56e3c6a137",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic sign with sharp focus and clear foreground separation, text 'DELTA' in red on a dark background, mounted on a wall",
7
+ "target_evidence": "text 'DELTA' on sign",
8
+ "expected_visual_elements": [
9
+ "sign",
10
+ "red text"
11
+ ],
12
+ "negative_prompt": "faded text, real brand logos, watermarks, cluttered background",
13
+ "diversity_tags": [
14
+ "sign",
15
+ "wall-mounted"
16
+ ],
17
+ "llm_model_id": "Qwen/Qwen3-8B",
18
+ "prompt_hash": "aa56e3c6a137",
19
+ "rule_valid": true,
20
+ "rule_issues": [],
21
+ "rule_score": 1.0,
22
+ "opening_key": "synthetic photorealistic sign with sharp",
23
+ "max_similarity_to_other_prompt": 0.64,
24
+ "diversity_score": 0.36,
25
+ "repeated_opening": false,
26
+ "review": {
27
+ "prompt_id": "ocr_heavy_d5_aa56e3c6a137",
28
+ "accept": true,
29
+ "quality": 5,
30
+ "decomposition_fit": 5,
31
+ "diversity": 4,
32
+ "issues": []
33
+ },
34
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
35
+ "final_rank_score": 0.97,
36
+ "status": "qwen_layered_done",
37
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
38
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
39
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
40
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/image.png",
41
+ "image_width": 1024,
42
+ "image_height": 1024,
43
+ "num_inference_steps": 28,
44
+ "guidance_scale": 3.5,
45
+ "seed": 27750557,
46
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
47
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
48
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
49
+ "qwen_layer_count_requested": 2,
50
+ "qwen_layer_count_returned": 2,
51
+ "qwen_layer_paths": [
52
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/layers/layer_00.png",
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/layers/layer_01.png"
54
+ ],
55
+ "qwen_resolution": 640,
56
+ "qwen_num_inference_steps": 30,
57
+ "qwen_true_cfg_scale": 4.0,
58
+ "qwen_seed": 719216,
59
+ "qwen_elapsed_seconds": 23.359,
60
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_aa56e3c6a137/record.json"
61
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_b6b750f3837b/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: c2066d68f87cf1b2b997bddc9ecabc9f5fd59bc753c258aa0f3dabaccc489390
  • Pointer size: 131 Bytes
  • Size of remote file: 513 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_b6b750f3837b/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 3a3301fac180ed25aa03d4fec3cfc96a50c7ac955d203462202fbd8e69bd0cbd
  • Pointer size: 131 Bytes
  • Size of remote file: 742 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_b6b750f3837b/record.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_b6b750f3837b",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic jersey with a label 'TAXI 5' in a contrasting color, clean stitching, and a sports fabric texture",
7
+ "target_evidence": "label 'TAXI 5' on jersey",
8
+ "expected_visual_elements": [
9
+ "jersey",
10
+ "label",
11
+ "sports fabric"
12
+ ],
13
+ "negative_prompt": "low contrast text, real logos, watermarks, distorted stitching",
14
+ "diversity_tags": [
15
+ "jersey",
16
+ "label"
17
+ ],
18
+ "llm_model_id": "Qwen/Qwen3-8B",
19
+ "prompt_hash": "b6b750f3837b",
20
+ "rule_valid": true,
21
+ "rule_issues": [],
22
+ "rule_score": 1.0,
23
+ "opening_key": "synthetic photorealistic jersey with a",
24
+ "max_similarity_to_other_prompt": 0.3571,
25
+ "diversity_score": 0.6429,
26
+ "repeated_opening": false,
27
+ "review": {
28
+ "prompt_id": "ocr_heavy_d5_b6b750f3837b",
29
+ "accept": true,
30
+ "quality": 5,
31
+ "decomposition_fit": 5,
32
+ "diversity": 4,
33
+ "issues": []
34
+ },
35
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
36
+ "final_rank_score": 0.97,
37
+ "status": "qwen_layered_done",
38
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
39
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
40
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
41
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_b6b750f3837b/image.png",
42
+ "image_width": 1024,
43
+ "image_height": 1024,
44
+ "num_inference_steps": 28,
45
+ "guidance_scale": 3.5,
46
+ "seed": 26958848,
47
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
48
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
49
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
50
+ "qwen_layer_count_requested": 2,
51
+ "qwen_layer_count_returned": 2,
52
+ "qwen_layer_paths": [
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_b6b750f3837b/layers/layer_00.png",
54
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_b6b750f3837b/layers/layer_01.png"
55
+ ],
56
+ "qwen_resolution": 640,
57
+ "qwen_num_inference_steps": 30,
58
+ "qwen_true_cfg_scale": 4.0,
59
+ "qwen_seed": 2301543,
60
+ "qwen_elapsed_seconds": 23.388,
61
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_b6b750f3837b/record.json"
62
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: aaa00d0ecca32080477fda965e782dcd1190c6b054ce5b6f0912c00d37fa6440
  • Pointer size: 131 Bytes
  • Size of remote file: 162 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 1cb1d21954901adbf826f9f2d6ed6787794561bd80bdf0c9ac1e075640241e2c
  • Pointer size: 131 Bytes
  • Size of remote file: 391 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/record.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_be8ea0c7f8f5",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic notebook page with clean boundaries and high contrast, text 'UNIT 12' centered at the top, white background, sharp focus",
7
+ "target_evidence": "text 'UNIT 12' on notebook page",
8
+ "expected_visual_elements": [
9
+ "notebook page",
10
+ "centered text"
11
+ ],
12
+ "negative_prompt": "blurry text, low contrast, real brand logos, watermarks",
13
+ "diversity_tags": [
14
+ "notebook",
15
+ "centered text"
16
+ ],
17
+ "llm_model_id": "Qwen/Qwen3-8B",
18
+ "prompt_hash": "be8ea0c7f8f5",
19
+ "rule_valid": true,
20
+ "rule_issues": [],
21
+ "rule_score": 1.0,
22
+ "opening_key": "synthetic photorealistic notebook page with",
23
+ "max_similarity_to_other_prompt": 0.7083,
24
+ "diversity_score": 0.0917,
25
+ "repeated_opening": true,
26
+ "review": {
27
+ "prompt_id": "ocr_heavy_d5_be8ea0c7f8f5",
28
+ "accept": true,
29
+ "quality": 5,
30
+ "decomposition_fit": 5,
31
+ "diversity": 4,
32
+ "issues": []
33
+ },
34
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
35
+ "final_rank_score": 0.97,
36
+ "status": "qwen_layered_done",
37
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
38
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
39
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
40
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/image.png",
41
+ "image_width": 1024,
42
+ "image_height": 1024,
43
+ "num_inference_steps": 28,
44
+ "guidance_scale": 3.5,
45
+ "seed": 27895355,
46
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
47
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
48
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
49
+ "qwen_layer_count_requested": 2,
50
+ "qwen_layer_count_returned": 2,
51
+ "qwen_layer_paths": [
52
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/layers/layer_00.png",
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/layers/layer_01.png"
54
+ ],
55
+ "qwen_resolution": 640,
56
+ "qwen_num_inference_steps": 30,
57
+ "qwen_true_cfg_scale": 4.0,
58
+ "qwen_seed": 673259,
59
+ "qwen_elapsed_seconds": 23.349,
60
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_be8ea0c7f8f5/record.json"
61
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: 57a399807e6cdd879ac3cf629c5e640c4363ae979ae46e6b8e70b453903bba51
  • Pointer size: 130 Bytes
  • Size of remote file: 81.9 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: bb3ea7ca02a4b58d60b2a7ea9e2cd282625f20baefedce60b4e48d39b3bb0dea
  • Pointer size: 131 Bytes
  • Size of remote file: 400 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/record.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_c6a2d9fda628",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic notebook page with clear boundaries, a small text region 'ROOM 18' on a clean white background, sharp focus on the text",
7
+ "target_evidence": "text region with 'ROOM 18'",
8
+ "expected_visual_elements": [
9
+ "notebook page",
10
+ "white background",
11
+ "sharp text focus"
12
+ ],
13
+ "negative_prompt": "blurry text, real brand logos, watermarks, complex background",
14
+ "diversity_tags": [
15
+ "notebook",
16
+ "white background"
17
+ ],
18
+ "llm_model_id": "Qwen/Qwen3-8B",
19
+ "prompt_hash": "c6a2d9fda628",
20
+ "rule_valid": true,
21
+ "rule_issues": [],
22
+ "rule_score": 1.0,
23
+ "opening_key": "synthetic photorealistic notebook page with",
24
+ "max_similarity_to_other_prompt": 0.6,
25
+ "diversity_score": 0.2,
26
+ "repeated_opening": true,
27
+ "review": {
28
+ "prompt_id": "ocr_heavy_d5_c6a2d9fda628",
29
+ "accept": true,
30
+ "quality": 5,
31
+ "decomposition_fit": 5,
32
+ "diversity": 4,
33
+ "issues": []
34
+ },
35
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
36
+ "final_rank_score": 0.97,
37
+ "status": "qwen_layered_done",
38
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
39
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
40
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
41
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/image.png",
42
+ "image_width": 1024,
43
+ "image_height": 1024,
44
+ "num_inference_steps": 28,
45
+ "guidance_scale": 3.5,
46
+ "seed": 23749804,
47
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
48
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
49
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
50
+ "qwen_layer_count_requested": 2,
51
+ "qwen_layer_count_returned": 2,
52
+ "qwen_layer_paths": [
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/layers/layer_00.png",
54
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/layers/layer_01.png"
55
+ ],
56
+ "qwen_resolution": 640,
57
+ "qwen_num_inference_steps": 30,
58
+ "qwen_true_cfg_scale": 4.0,
59
+ "qwen_seed": 4242814,
60
+ "qwen_elapsed_seconds": 23.397,
61
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_c6a2d9fda628/record.json"
62
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: 0ba61d5902e2daefcc8c9f0d0299ee471a6e4dbddcdcef878cc545fd5e543193
  • Pointer size: 131 Bytes
  • Size of remote file: 380 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 2ebae150f57567a2046b8b266ae1a9bfa8d851be27afb0af5529e3b17651de43
  • Pointer size: 131 Bytes
  • Size of remote file: 364 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/record.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_d35a0d0f21f2",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic menu board with 'ALTO' in a clean, high-contrast font on a white board, sharp focus on text",
7
+ "target_evidence": "text 'ALTO' on menu board",
8
+ "expected_visual_elements": [
9
+ "menu board",
10
+ "white background",
11
+ "sharp text"
12
+ ],
13
+ "negative_prompt": "faded text, cluttered background, real logos, low contrast",
14
+ "diversity_tags": [
15
+ "menu board",
16
+ "white background"
17
+ ],
18
+ "llm_model_id": "Qwen/Qwen3-8B",
19
+ "prompt_hash": "d35a0d0f21f2",
20
+ "rule_valid": true,
21
+ "rule_issues": [],
22
+ "rule_score": 1.0,
23
+ "opening_key": "synthetic photorealistic menu board with",
24
+ "max_similarity_to_other_prompt": 0.6087,
25
+ "diversity_score": 0.1913,
26
+ "repeated_opening": true,
27
+ "review": {
28
+ "prompt_id": "ocr_heavy_d5_d35a0d0f21f2",
29
+ "accept": true,
30
+ "quality": 5,
31
+ "decomposition_fit": 5,
32
+ "diversity": 4,
33
+ "issues": []
34
+ },
35
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
36
+ "final_rank_score": 0.97,
37
+ "status": "qwen_layered_done",
38
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
39
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
40
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
41
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/image.png",
42
+ "image_width": 1024,
43
+ "image_height": 1024,
44
+ "num_inference_steps": 28,
45
+ "guidance_scale": 3.5,
46
+ "seed": 29194225,
47
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
48
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
49
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
50
+ "qwen_layer_count_requested": 2,
51
+ "qwen_layer_count_returned": 2,
52
+ "qwen_layer_paths": [
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/layers/layer_00.png",
54
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/layers/layer_01.png"
55
+ ],
56
+ "qwen_resolution": 640,
57
+ "qwen_num_inference_steps": 30,
58
+ "qwen_true_cfg_scale": 4.0,
59
+ "qwen_seed": 9324524,
60
+ "qwen_elapsed_seconds": 23.4,
61
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_d35a0d0f21f2/record.json"
62
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_db75b0748b7e/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: 7a1ddaaaf19a4ab4ae7751f51487ede225510bbe1154c89c2a0b039fe3718c3e
  • Pointer size: 131 Bytes
  • Size of remote file: 353 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_db75b0748b7e/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 924864a123e746f859bcbab67784a7fb82e8bbfe6493a83bcf95e932a9c37dcb
  • Pointer size: 131 Bytes
  • Size of remote file: 705 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_db75b0748b7e/record.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_db75b0748b7e",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic jersey with high contrast and clean boundaries, text 'ALTO' in black on a light blue background, worn fabric texture",
7
+ "target_evidence": "text 'ALTO' on jersey",
8
+ "expected_visual_elements": [
9
+ "jersey",
10
+ "text on fabric"
11
+ ],
12
+ "negative_prompt": "low contrast, blurry text, real logos, watermarks",
13
+ "diversity_tags": [
14
+ "jersey",
15
+ "fabric texture"
16
+ ],
17
+ "llm_model_id": "Qwen/Qwen3-8B",
18
+ "prompt_hash": "db75b0748b7e",
19
+ "rule_valid": true,
20
+ "rule_issues": [],
21
+ "rule_score": 1.0,
22
+ "opening_key": "synthetic photorealistic jersey with high",
23
+ "max_similarity_to_other_prompt": 0.5185,
24
+ "diversity_score": 0.4815,
25
+ "repeated_opening": false,
26
+ "review": {
27
+ "prompt_id": "ocr_heavy_d5_db75b0748b7e",
28
+ "accept": true,
29
+ "quality": 5,
30
+ "decomposition_fit": 5,
31
+ "diversity": 4,
32
+ "issues": []
33
+ },
34
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
35
+ "final_rank_score": 0.97,
36
+ "status": "qwen_layered_done",
37
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
38
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
39
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
40
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_db75b0748b7e/image.png",
41
+ "image_width": 1024,
42
+ "image_height": 1024,
43
+ "num_inference_steps": 28,
44
+ "guidance_scale": 3.5,
45
+ "seed": 26656383,
46
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
47
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
48
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
49
+ "qwen_layer_count_requested": 2,
50
+ "qwen_layer_count_returned": 2,
51
+ "qwen_layer_paths": [
52
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_db75b0748b7e/layers/layer_00.png",
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_db75b0748b7e/layers/layer_01.png"
54
+ ],
55
+ "qwen_resolution": 640,
56
+ "qwen_num_inference_steps": 30,
57
+ "qwen_true_cfg_scale": 4.0,
58
+ "qwen_seed": 8354361,
59
+ "qwen_elapsed_seconds": 23.342,
60
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_db75b0748b7e/record.json"
61
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f5a708c509ca/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: 38a5f5c7fbdc80d28628e5e354782c39e96a0da57389bc3f2cbe6a85ca0514ab
  • Pointer size: 131 Bytes
  • Size of remote file: 409 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f5a708c509ca/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 36ddfb629c9995de682875085294bcc05e452e438b18820784695f1d78bbfab5
  • Pointer size: 131 Bytes
  • Size of remote file: 316 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f5a708c509ca/record.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_f5a708c509ca",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic poster with sharp focus and clear foreground separation, text 'ZEN' in white on a dark blue background, hanging on a wall",
7
+ "target_evidence": "text 'ZEN' on poster",
8
+ "expected_visual_elements": [
9
+ "poster",
10
+ "wall-mounted"
11
+ ],
12
+ "negative_prompt": "faded text, low contrast, real brand logos, watermarks",
13
+ "diversity_tags": [
14
+ "poster",
15
+ "wall-mounted"
16
+ ],
17
+ "llm_model_id": "Qwen/Qwen3-8B",
18
+ "prompt_hash": "f5a708c509ca",
19
+ "rule_valid": true,
20
+ "rule_issues": [],
21
+ "rule_score": 1.0,
22
+ "opening_key": "synthetic photorealistic poster with sharp",
23
+ "max_similarity_to_other_prompt": 0.6957,
24
+ "diversity_score": 0.3043,
25
+ "repeated_opening": false,
26
+ "review": {
27
+ "prompt_id": "ocr_heavy_d5_f5a708c509ca",
28
+ "accept": true,
29
+ "quality": 5,
30
+ "decomposition_fit": 5,
31
+ "diversity": 4,
32
+ "issues": []
33
+ },
34
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
35
+ "final_rank_score": 0.97,
36
+ "status": "qwen_layered_done",
37
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
38
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
39
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
40
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_f5a708c509ca/image.png",
41
+ "image_width": 1024,
42
+ "image_height": 1024,
43
+ "num_inference_steps": 28,
44
+ "guidance_scale": 3.5,
45
+ "seed": 29901775,
46
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
47
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
48
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
49
+ "qwen_layer_count_requested": 2,
50
+ "qwen_layer_count_returned": 2,
51
+ "qwen_layer_paths": [
52
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f5a708c509ca/layers/layer_00.png",
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f5a708c509ca/layers/layer_01.png"
54
+ ],
55
+ "qwen_resolution": 640,
56
+ "qwen_num_inference_steps": 30,
57
+ "qwen_true_cfg_scale": 4.0,
58
+ "qwen_seed": 624180,
59
+ "qwen_elapsed_seconds": 23.433,
60
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f5a708c509ca/record.json"
61
+ }
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/layers/layer_00.png ADDED

Git LFS Details

  • SHA256: b47efe1bd9ec96c961a2c1ccbd550af869a45ddad629b49d58eda27dd42291e0
  • Pointer size: 131 Bytes
  • Size of remote file: 214 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/layers/layer_01.png ADDED

Git LFS Details

  • SHA256: 7159746d5f11f384ddb01d0a6a780dcabd2ce8135293bc4c826fdec81d221f60
  • Pointer size: 131 Bytes
  • Size of remote file: 452 kB
vec_pilot/layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/record.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prompt_id": "ocr_heavy_d5_f75d44bf94ac",
3
+ "run_name": "flux2_qil_decomposition_pilot",
4
+ "category": "ocr_heavy",
5
+ "difficulty": 5,
6
+ "image_prompt": "synthetic photorealistic tag on a clean background with high contrast, text 'RIVER 9' in black on a light yellow label, attached to a box",
7
+ "target_evidence": "text 'RIVER 9' on tag",
8
+ "expected_visual_elements": [
9
+ "tag",
10
+ "label"
11
+ ],
12
+ "negative_prompt": "blurry text, low contrast, real brand logos, watermarks",
13
+ "diversity_tags": [
14
+ "tag",
15
+ "box"
16
+ ],
17
+ "llm_model_id": "Qwen/Qwen3-8B",
18
+ "prompt_hash": "f75d44bf94ac",
19
+ "rule_valid": true,
20
+ "rule_issues": [],
21
+ "rule_score": 1.0,
22
+ "opening_key": "synthetic photorealistic tag on a",
23
+ "max_similarity_to_other_prompt": 0.4815,
24
+ "diversity_score": 0.3185,
25
+ "repeated_opening": true,
26
+ "review": {
27
+ "prompt_id": "ocr_heavy_d5_f75d44bf94ac",
28
+ "accept": true,
29
+ "quality": 5,
30
+ "decomposition_fit": 5,
31
+ "diversity": 4,
32
+ "issues": []
33
+ },
34
+ "review_model_id": "/root/Layered-VQA/models/repos/Qwen_Qwen3-30B-A3B-Instruct-2507",
35
+ "final_rank_score": 0.97,
36
+ "status": "qwen_layered_done",
37
+ "image_model_id": "black-forest-labs/FLUX.2-dev",
38
+ "image_model_source": "/root/Layered-VQA/models/repos/black-forest-labs_FLUX.2-dev",
39
+ "image_model_slug": "black-forest-labs_FLUX.2-dev",
40
+ "image_path": "images/black-forest-labs_FLUX.2-dev/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/image.png",
41
+ "image_width": 1024,
42
+ "image_height": 1024,
43
+ "num_inference_steps": 28,
44
+ "guidance_scale": 3.5,
45
+ "seed": 27626645,
46
+ "qwen_layered_model_id": "Qwen/Qwen-Image-Layered",
47
+ "qwen_layered_model_source": "/root/Layered-VQA/models/repos/Qwen_Qwen-Image-Layered",
48
+ "qwen_layered_prompt": "separate the image into clean transparent layers for foreground objects, readable text, occluders, relation-critical objects, and background",
49
+ "qwen_layer_count_requested": 2,
50
+ "qwen_layer_count_returned": 2,
51
+ "qwen_layer_paths": [
52
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/layers/layer_00.png",
53
+ "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/layers/layer_01.png"
54
+ ],
55
+ "qwen_resolution": 640,
56
+ "qwen_num_inference_steps": 30,
57
+ "qwen_true_cfg_scale": 4.0,
58
+ "qwen_seed": 7367049,
59
+ "qwen_elapsed_seconds": 23.405,
60
+ "qwen_record_path": "layers/qwen_image_layered/black-forest-labs_FLUX.2-dev/k2/ocr_heavy/ocr_heavy_d5_f75d44bf94ac/record.json"
61
+ }