Hugging Face's logo

Datasets:
maveryn
/
vlm-fix

Modalities:
Image
Tabular
Text
Formats:
parquet
Size:
100K - 1M
Libraries:
Datasets
pandas
Polars
Dataset card Data Studio Files
xet
 Community
2
maveryn commited on
Commit
644493f
·
verified ·
1 Parent(s): fd7f818

Add files using upload-large-folder tool

Browse files
Files changed (18) hide show
  1. README.md +52 -0
  2. manifest.json +65 -0
  3. synth_legs_train/export_summary.json +7 -0
  4. synth_legs_train/train.parquet +3 -0
  5. vlm_fix/export_summary.json +9 -0
  6. vlm_fix/images.parquet +3 -0
  7. vlm_fix/main.parquet +3 -0
  8. vlm_fix_posttrain_d1/export_summary.json +11 -0
  9. vlm_fix_posttrain_d1/test.parquet +3 -0
  10. vlm_fix_posttrain_d1/train.parquet +3 -0
  11. vlm_fix_posttrain_d2/export_summary.json +11 -0
  12. vlm_fix_posttrain_d2/test.parquet +3 -0
  13. vlm_fix_posttrain_d2/train.parquet +3 -0
  14. vlm_fix_posttrain_d3/export_summary.json +11 -0
  15. vlm_fix_posttrain_d3/test.parquet +3 -0
  16. vlm_fix_posttrain_d3/train.parquet +3 -0
  17. vlm_fix_text_only/export_summary.json +7 -0
  18. vlm_fix_text_only/main.parquet +3 -0
README.md ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ configs:
3
+ - config_name: vlm_fix
4
+ data_files:
5
+ - split: main
6
+ path: vlm_fix/main.parquet
7
+ - split: images
8
+ path: vlm_fix/images.parquet
9
+ - config_name: vlm_fix_text_only
10
+ data_files:
11
+ - split: main
12
+ path: vlm_fix_text_only/main.parquet
13
+ - config_name: vlm_fix_posttrain_d1
14
+ data_files:
15
+ - split: train
16
+ path: vlm_fix_posttrain_d1/train.parquet
17
+ - split: test
18
+ path: vlm_fix_posttrain_d1/test.parquet
19
+ - config_name: vlm_fix_posttrain_d2
20
+ data_files:
21
+ - split: train
22
+ path: vlm_fix_posttrain_d2/train.parquet
23
+ - split: test
24
+ path: vlm_fix_posttrain_d2/test.parquet
25
+ - config_name: vlm_fix_posttrain_d3
26
+ data_files:
27
+ - split: train
28
+ path: vlm_fix_posttrain_d3/train.parquet
29
+ - split: test
30
+ path: vlm_fix_posttrain_d3/test.parquet
31
+ - config_name: synth_legs_train
32
+ data_files:
33
+ - split: train
34
+ path: synth_legs_train/train.parquet
35
+ ---
36
+
37
+ # vlmfix HF Upload
38
+
39
+ This folder contains local parquet exports prepared for Hugging Face upload.
40
+
41
+ Included configs:
42
+ - `vlm_fix`: main, images
43
+ - `vlm_fix_text_only`: main
44
+ - `vlm_fix_posttrain_d1`: train, test
45
+ - `vlm_fix_posttrain_d2`: train, test
46
+ - `vlm_fix_posttrain_d3`: train, test
47
+ - `synth_legs_train`: train
48
+
49
+ Not included:
50
+ - `VLMBias` training data
51
+
52
+ VLMBias should instead be derived from `anvo25/vlms-are-biased` using the existing prep/build code.
manifest.json ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "export_root": "/home/jovyan/work/vlmfixsubmit/vlmfix/hf_export",
3
+ "configs": [
4
+ {
5
+ "config": "vlm_fix",
6
+ "splits": {
7
+ "main": 115200,
8
+ "images": 3600
9
+ },
10
+ "source_instances": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/vlm_fix/instances.parquet",
11
+ "unique_images": 3600
12
+ },
13
+ {
14
+ "config": "vlm_fix_text_only",
15
+ "splits": {
16
+ "main": 4800
17
+ },
18
+ "source_instances": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/vlm_fix_text_only/instances.parquet"
19
+ },
20
+ {
21
+ "config": "vlm_fix_posttrain_d1",
22
+ "splits": {
23
+ "train": 8192,
24
+ "test": 2400
25
+ },
26
+ "sources": {
27
+ "train": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D1_train_standard_all4_8192.parquet",
28
+ "test": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D1_test_inverse_core2400.parquet"
29
+ }
30
+ },
31
+ {
32
+ "config": "vlm_fix_posttrain_d2",
33
+ "splits": {
34
+ "train": 8192,
35
+ "test": 2400
36
+ },
37
+ "sources": {
38
+ "train": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D2_train_inverse_all4_8192.parquet",
39
+ "test": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D2_test_standard_core2400.parquet"
40
+ }
41
+ },
42
+ {
43
+ "config": "vlm_fix_posttrain_d3",
44
+ "splits": {
45
+ "train": 8192,
46
+ "test": 2400
47
+ },
48
+ "sources": {
49
+ "train": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D3_train_ttt_reversi_both_rules_8192.parquet",
50
+ "test": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D3_test_connect4_dots_both_rules_core2400.parquet"
51
+ }
52
+ },
53
+ {
54
+ "config": "synth_legs_train",
55
+ "splits": {
56
+ "train": 8192
57
+ },
58
+ "source_instances": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/synth_legs/train_8192.parquet"
59
+ }
60
+ ],
61
+ "notes": {
62
+ "vlm_fix": "Normalized into main + images splits to avoid repeating image bytes in every benchmark row.",
63
+ "vlmbias": "Excluded from HF export by request; derive from anvo25/vlms-are-biased instead."
64
+ }
65
+ }
synth_legs_train/export_summary.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": "synth_legs_train",
3
+ "splits": {
4
+ "train": 8192
5
+ },
6
+ "source_instances": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/synth_legs/train_8192.parquet"
7
+ }
synth_legs_train/train.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6812d224589bd59a49ae03f471ad0608d816cc2ea451e51d3f3128036bc56587
3
+ size 193676989
vlm_fix/export_summary.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": "vlm_fix",
3
+ "splits": {
4
+ "main": 115200,
5
+ "images": 3600
6
+ },
7
+ "source_instances": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/vlm_fix/instances.parquet",
8
+ "unique_images": 3600
9
+ }
vlm_fix/images.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:711dc83ff599fc53732fe9594636099dcf9025afe63ef7fd69aa29c5eaa8592a
3
+ size 121261719
vlm_fix/main.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4110f3e1ad19d85d0c6d7799414409fb8bddf33ad93a18f249133c737081dffd
3
+ size 2315836
vlm_fix_posttrain_d1/export_summary.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": "vlm_fix_posttrain_d1",
3
+ "splits": {
4
+ "train": 8192,
5
+ "test": 2400
6
+ },
7
+ "sources": {
8
+ "train": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D1_train_standard_all4_8192.parquet",
9
+ "test": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D1_test_inverse_core2400.parquet"
10
+ }
11
+ }
vlm_fix_posttrain_d1/test.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f205cb0239b5e9685d63201884dcf21f3122366aa4fa5d7c4db3c6a107d679e8
3
+ size 39983791
vlm_fix_posttrain_d1/train.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab7ec3905652f052dc69b92080c86f8d54a8efb33f17c46946b187f9accf067e
3
+ size 137983889
vlm_fix_posttrain_d2/export_summary.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": "vlm_fix_posttrain_d2",
3
+ "splits": {
4
+ "train": 8192,
5
+ "test": 2400
6
+ },
7
+ "sources": {
8
+ "train": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D2_train_inverse_all4_8192.parquet",
9
+ "test": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D2_test_standard_core2400.parquet"
10
+ }
11
+ }
vlm_fix_posttrain_d2/test.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:473c8843c927399f37c7a220b4084c5f6bf2c5704f7d392d081bbe07da6b8bb4
3
+ size 39983737
vlm_fix_posttrain_d2/train.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5bc97a2a353c1e6cb1a02960db3adb78f5af04ac9c4160c81547eda2f8756b9
3
+ size 137984106
vlm_fix_posttrain_d3/export_summary.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": "vlm_fix_posttrain_d3",
3
+ "splits": {
4
+ "train": 8192,
5
+ "test": 2400
6
+ },
7
+ "sources": {
8
+ "train": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D3_train_ttt_reversi_both_rules_8192.parquet",
9
+ "test": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/post_training_v2_disjoint/D3_test_connect4_dots_both_rules_core2400.parquet"
10
+ }
11
+ }
vlm_fix_posttrain_d3/test.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3e5cae5de7ded5c05d336a4d826ffd53a1cdf298e291693c3066fe394f04edd6
3
+ size 45465170
vlm_fix_posttrain_d3/train.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b09c4352b0383e5adc2c7dbb70534710c14882291705d8a27832e8553faea673
3
+ size 121101837
vlm_fix_text_only/export_summary.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": "vlm_fix_text_only",
3
+ "splits": {
4
+ "main": 4800
5
+ },
6
+ "source_instances": "/home/jovyan/work/vlmfixsubmit/vlmfix/data/generated/vlm_fix_text_only/instances.parquet"
7
+ }
vlm_fix_text_only/main.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1f89a26543cea8cda4dfdce81f5d51b72517481b49dedac58f3a9e3da101641
3
+ size 407951