Hugging Face's logo

Datasets:
HaoranLiu
/
CUAGym

Tasks:
Image-Text-to-Text
Modalities:
Image
Text
Formats:
parquet
Size:
10K - 100K
Tags:
cua-lite
gui
sft
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
HaoranLiu commited on
Commit
ffdc97b
·
verified ·
1 Parent(s): c9b9de3

Upload CUAGym batch 1

Browse files
Files changed (50) hide show
  1. README.md +105 -35
  2. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00000-of-00126.parquet +3 -0
  3. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00001-of-00126.parquet +3 -0
  4. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00002-of-00126.parquet +3 -0
  5. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00003-of-00126.parquet +3 -0
  6. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00004-of-00126.parquet +3 -0
  7. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00005-of-00126.parquet +3 -0
  8. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00006-of-00126.parquet +3 -0
  9. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00007-of-00126.parquet +3 -0
  10. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00008-of-00126.parquet +3 -0
  11. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00009-of-00126.parquet +3 -0
  12. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00010-of-00126.parquet +3 -0
  13. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00011-of-00126.parquet +3 -0
  14. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00012-of-00126.parquet +3 -0
  15. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00013-of-00126.parquet +3 -0
  16. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00014-of-00126.parquet +3 -0
  17. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00015-of-00126.parquet +3 -0
  18. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00016-of-00126.parquet +3 -0
  19. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00017-of-00126.parquet +3 -0
  20. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00018-of-00126.parquet +3 -0
  21. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00019-of-00126.parquet +3 -0
  22. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00020-of-00126.parquet +3 -0
  23. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00021-of-00126.parquet +3 -0
  24. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00022-of-00126.parquet +3 -0
  25. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00023-of-00126.parquet +3 -0
  26. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00024-of-00126.parquet +3 -0
  27. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00025-of-00126.parquet +3 -0
  28. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00026-of-00126.parquet +3 -0
  29. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00027-of-00126.parquet +3 -0
  30. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00028-of-00126.parquet +3 -0
  31. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00029-of-00126.parquet +3 -0
  32. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00030-of-00126.parquet +3 -0
  33. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00031-of-00126.parquet +3 -0
  34. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00032-of-00126.parquet +3 -0
  35. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00033-of-00126.parquet +3 -0
  36. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00034-of-00126.parquet +3 -0
  37. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00035-of-00126.parquet +3 -0
  38. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00036-of-00126.parquet +3 -0
  39. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00037-of-00126.parquet +3 -0
  40. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00038-of-00126.parquet +3 -0
  41. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00039-of-00126.parquet +3 -0
  42. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00040-of-00126.parquet +3 -0
  43. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00041-of-00126.parquet +3 -0
  44. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00042-of-00126.parquet +3 -0
  45. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00043-of-00126.parquet +3 -0
  46. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00044-of-00126.parquet +3 -0
  47. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00045-of-00126.parquet +3 -0
  48. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00046-of-00126.parquet +3 -0
  49. desktop/use/train/Qwen3-VL-2B-Instruct/shard-00047-of-00126.parquet +3 -0
  50. stats.json +25 -15
README.md CHANGED
@@ -14,12 +14,30 @@ configs:
14
  - "*/*/train*parquet"
15
  - "*/*/train/*.parquet"
16
  - "*/*/train/*/*.parquet"
17
- - config_name: gpt-5.5
18
  data_files:
19
  - split: train
20
  path:
21
- - "*/*/train/gpt-5.5.parquet"
22
- - "*/*/train/gpt-5.5/*.parquet"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
23
  - config_name: Qwen3.5-27B
24
  data_files:
25
  - split: train
@@ -44,42 +62,84 @@ configs:
44
  path:
45
  - "*/*/train/Qwen3.5-9B.parquet"
46
  - "*/*/train/Qwen3.5-9B/*.parquet"
47
- - config_name: Qwen3-VL-2B-Instruct
48
  data_files:
49
  - split: train
50
  path:
51
- - "*/*/train/Qwen3-VL-2B-Instruct.parquet"
52
- - "*/*/train/Qwen3-VL-2B-Instruct/*.parquet"
53
- - config_name: Qwen3-VL-32B-Instruct
54
  data_files:
55
  - split: train
56
  path:
57
- - "*/*/train/Qwen3-VL-32B-Instruct.parquet"
58
- - "*/*/train/Qwen3-VL-32B-Instruct/*.parquet"
59
- - config_name: Qwen3-VL-4B-Instruct
60
  data_files:
61
  - split: train
62
  path:
63
- - "*/*/train/Qwen3-VL-4B-Instruct.parquet"
64
- - "*/*/train/Qwen3-VL-4B-Instruct/*.parquet"
65
- - config_name: Qwen3-VL-8B-Instruct
66
  data_files:
67
  - split: train
68
  path:
69
- - "*/*/train/Qwen3-VL-8B-Instruct.parquet"
70
- - "*/*/train/Qwen3-VL-8B-Instruct/*.parquet"
71
- - config_name: UI-TARS-1.5-7B
72
  data_files:
73
  - split: train
74
  path:
75
- - "*/*/train/UI-TARS-1.5-7B.parquet"
76
- - "*/*/train/UI-TARS-1.5-7B/*.parquet"
77
- - config_name: UI-TARS-7B-DPO
78
  data_files:
79
  - split: train
80
  path:
81
- - "*/*/train/UI-TARS-7B-DPO.parquet"
82
- - "*/*/train/UI-TARS-7B-DPO/*.parquet"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
83
  ---
84
 
85
  # cua-lite/CUAGym
@@ -99,7 +159,7 @@ from datasets import load_dataset
99
  ds = load_dataset("cua-lite/CUAGym")
100
 
101
  # just one named subset (config)
102
- ds = load_dataset("cua-lite/CUAGym", "gpt-5.5")
103
  ```
104
 
105
  You can also filter by `metadata.platform` / `metadata.task_type` /
@@ -146,17 +206,27 @@ directly yields the folded form.
146
 
147
  | platform | task_type | variant | train | validation |
148
  |---|---|---|---:|---:|
149
- | web | use | gpt-5.5 | 1,330 | 0 |
150
- | web | use | Qwen3.5-27B | 1,328 | 0 |
151
- | web | use | Qwen3.5-2B | 1,330 | 0 |
152
- | web | use | Qwen3.5-4B | 1,332 | 0 |
153
- | web | use | Qwen3.5-9B | 1,324 | 0 |
154
- | web | use | Qwen3-VL-2B-Instruct | 1,334 | 0 |
155
- | web | use | Qwen3-VL-32B-Instruct | 1,333 | 0 |
156
- | web | use | Qwen3-VL-4B-Instruct | 1,334 | 0 |
157
- | web | use | Qwen3-VL-8B-Instruct | 1,330 | 0 |
158
- | web | use | UI-TARS-1.5-7B | 1,332 | 0 |
159
- | web | use | UI-TARS-7B-DPO | 1,329 | 0 |
 
 
 
 
 
 
 
 
 
 
160
 
161
  ## Local mirror & SFT export
162
 
@@ -174,8 +244,8 @@ Rows in the local parquet have `images: list[str]`; bytes are extracted to
174
  the image store. `lite.train.export.export_sft` consumes the local
175
  form directly with `--image-root=$CUA_LITE_DATASETS_ROOT`.
176
 
177
- - Total unique images: **237,207**
178
- - Image store size: **52.09 GB**
179
 
180
  ## Notes
181
 
 
14
  - "*/*/train*parquet"
15
  - "*/*/train/*.parquet"
16
  - "*/*/train/*/*.parquet"
17
+ - config_name: Qwen3-VL-2B-Instruct
18
  data_files:
19
  - split: train
20
  path:
21
+ - "*/*/train/Qwen3-VL-2B-Instruct.parquet"
22
+ - "*/*/train/Qwen3-VL-2B-Instruct/*.parquet"
23
+ - config_name: Qwen3-VL-32B-Instruct
24
+ data_files:
25
+ - split: train
26
+ path:
27
+ - "*/*/train/Qwen3-VL-32B-Instruct.parquet"
28
+ - "*/*/train/Qwen3-VL-32B-Instruct/*.parquet"
29
+ - config_name: Qwen3-VL-4B-Instruct
30
+ data_files:
31
+ - split: train
32
+ path:
33
+ - "*/*/train/Qwen3-VL-4B-Instruct.parquet"
34
+ - "*/*/train/Qwen3-VL-4B-Instruct/*.parquet"
35
+ - config_name: Qwen3-VL-8B-Instruct
36
+ data_files:
37
+ - split: train
38
+ path:
39
+ - "*/*/train/Qwen3-VL-8B-Instruct.parquet"
40
+ - "*/*/train/Qwen3-VL-8B-Instruct/*.parquet"
41
  - config_name: Qwen3.5-27B
42
  data_files:
43
  - split: train
 
62
  path:
63
  - "*/*/train/Qwen3.5-9B.parquet"
64
  - "*/*/train/Qwen3.5-9B/*.parquet"
65
+ - config_name: UI-TARS-1.5-7B
66
  data_files:
67
  - split: train
68
  path:
69
+ - "*/*/train/UI-TARS-1.5-7B.parquet"
70
+ - "*/*/train/UI-TARS-1.5-7B/*.parquet"
71
+ - config_name: UI-TARS-7B-DPO
72
  data_files:
73
  - split: train
74
  path:
75
+ - "*/*/train/UI-TARS-7B-DPO.parquet"
76
+ - "*/*/train/UI-TARS-7B-DPO/*.parquet"
77
+ - config_name: web.use.gpt_5_5
78
  data_files:
79
  - split: train
80
  path:
81
+ - "*/*/train/web.use.gpt_5_5.parquet"
82
+ - "*/*/train/web.use.gpt_5_5/*.parquet"
83
+ - config_name: web.use.qwen3_5_27b
84
  data_files:
85
  - split: train
86
  path:
87
+ - "*/*/train/web.use.qwen3_5_27b.parquet"
88
+ - "*/*/train/web.use.qwen3_5_27b/*.parquet"
89
+ - config_name: web.use.qwen3_5_2b
90
  data_files:
91
  - split: train
92
  path:
93
+ - "*/*/train/web.use.qwen3_5_2b.parquet"
94
+ - "*/*/train/web.use.qwen3_5_2b/*.parquet"
95
+ - config_name: web.use.qwen3_5_4b
96
  data_files:
97
  - split: train
98
  path:
99
+ - "*/*/train/web.use.qwen3_5_4b.parquet"
100
+ - "*/*/train/web.use.qwen3_5_4b/*.parquet"
101
+ - config_name: web.use.qwen3_5_9b
102
+ data_files:
103
+ - split: train
104
+ path:
105
+ - "*/*/train/web.use.qwen3_5_9b.parquet"
106
+ - "*/*/train/web.use.qwen3_5_9b/*.parquet"
107
+ - config_name: web.use.qwen3_vl_2b
108
+ data_files:
109
+ - split: train
110
+ path:
111
+ - "*/*/train/web.use.qwen3_vl_2b.parquet"
112
+ - "*/*/train/web.use.qwen3_vl_2b/*.parquet"
113
+ - config_name: web.use.qwen3_vl_32b
114
+ data_files:
115
+ - split: train
116
+ path:
117
+ - "*/*/train/web.use.qwen3_vl_32b.parquet"
118
+ - "*/*/train/web.use.qwen3_vl_32b/*.parquet"
119
+ - config_name: web.use.qwen3_vl_4b
120
+ data_files:
121
+ - split: train
122
+ path:
123
+ - "*/*/train/web.use.qwen3_vl_4b.parquet"
124
+ - "*/*/train/web.use.qwen3_vl_4b/*.parquet"
125
+ - config_name: web.use.qwen3_vl_8b
126
+ data_files:
127
+ - split: train
128
+ path:
129
+ - "*/*/train/web.use.qwen3_vl_8b.parquet"
130
+ - "*/*/train/web.use.qwen3_vl_8b/*.parquet"
131
+ - config_name: web.use.ui_tars_15_7b
132
+ data_files:
133
+ - split: train
134
+ path:
135
+ - "*/*/train/web.use.ui_tars_15_7b.parquet"
136
+ - "*/*/train/web.use.ui_tars_15_7b/*.parquet"
137
+ - config_name: web.use.ui_tars_7b_dpo
138
+ data_files:
139
+ - split: train
140
+ path:
141
+ - "*/*/train/web.use.ui_tars_7b_dpo.parquet"
142
+ - "*/*/train/web.use.ui_tars_7b_dpo/*.parquet"
143
  ---
144
 
145
  # cua-lite/CUAGym
 
159
  ds = load_dataset("cua-lite/CUAGym")
160
 
161
  # just one named subset (config)
162
+ ds = load_dataset("cua-lite/CUAGym", "Qwen3-VL-2B-Instruct")
163
  ```
164
 
165
  You can also filter by `metadata.platform` / `metadata.task_type` /
 
206
 
207
  | platform | task_type | variant | train | validation |
208
  |---|---|---|---:|---:|
209
+ | desktop | use | Qwen3-VL-2B-Instruct | 1,451 | 0 |
210
+ | desktop | use | Qwen3-VL-32B-Instruct | 1,452 | 0 |
211
+ | desktop | use | Qwen3-VL-4B-Instruct | 1,452 | 0 |
212
+ | desktop | use | Qwen3-VL-8B-Instruct | 1,453 | 0 |
213
+ | desktop | use | Qwen3.5-27B | 1,449 | 0 |
214
+ | desktop | use | Qwen3.5-2B | 1,447 | 0 |
215
+ | desktop | use | Qwen3.5-4B | 1,454 | 0 |
216
+ | desktop | use | Qwen3.5-9B | 1,451 | 0 |
217
+ | desktop | use | UI-TARS-1.5-7B | 1,452 | 0 |
218
+ | desktop | use | UI-TARS-7B-DPO | 1,449 | 0 |
219
+ | web | use | web.use.gpt_5_5 | 1,330 | 0 |
220
+ | web | use | web.use.qwen3_5_27b | 1,328 | 0 |
221
+ | web | use | web.use.qwen3_5_2b | 1,330 | 0 |
222
+ | web | use | web.use.qwen3_5_4b | 1,332 | 0 |
223
+ | web | use | web.use.qwen3_5_9b | 1,324 | 0 |
224
+ | web | use | web.use.qwen3_vl_2b | 1,334 | 0 |
225
+ | web | use | web.use.qwen3_vl_32b | 1,333 | 0 |
226
+ | web | use | web.use.qwen3_vl_4b | 1,334 | 0 |
227
+ | web | use | web.use.qwen3_vl_8b | 1,330 | 0 |
228
+ | web | use | web.use.ui_tars_15_7b | 1,332 | 0 |
229
+ | web | use | web.use.ui_tars_7b_dpo | 1,329 | 0 |
230
 
231
  ## Local mirror & SFT export
232
 
 
244
  the image store. `lite.train.export.export_sft` consumes the local
245
  form directly with `--image-root=$CUA_LITE_DATASETS_ROOT`.
246
 
247
+ - Total unique images: **490,927**
248
+ - Image store size: **168.21 GB**
249
 
250
  ## Notes
251
 
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00000-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14eddc855e8536fd53aa9b8db0b3d457eb2114b3385fe6aacc82fb4f7bf2349a
3
+ size 163105382
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00001-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03b77185fa9f351a2e5144a723c7b950a3f33f9de360280ddfa403ad9a2d84d7
3
+ size 55539518
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00002-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d2620bff5a41587c71840bf2b1a9ea6947a29cb34c9e60dc00e669ee06c88214
3
+ size 150014966
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00003-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1de33f86426f38cdef142e80b181837534240449015a9ad6b60f40a5f455f643
3
+ size 120809819
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00004-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7acd1314b2fb18cea3308f4bd580ba1d7617983a2f5495b3d5d3ef3cf6969588
3
+ size 82817635
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00005-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:377aa91f5d84e0c0995e4579b801d64ccaf19706a3155a11173132af6b03de8c
3
+ size 156402306
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00006-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dcaa3beb54db432907ed4b657100ef84e1bb63ed90a0cb97a685b5bad7b0a2bd
3
+ size 93194078
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00007-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4bfa9a523b46cd23b85f3085529d617b77553e95a812339025649f3979de47ee
3
+ size 54887538
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00008-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a301855fe400409f3099c83fadc0c4f179219b4a37920eb53da49b0e7fe5a11c
3
+ size 69724177
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00009-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebb8fa93d53ad4b598601d7f0da9885a3475572618f4b098b45665869d554f6a
3
+ size 115849401
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00010-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fafc62cd1a08a97a90c385e78e41603322d47f320880207b66593ab5f97e9985
3
+ size 81857096
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00011-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e390812f434d95dd3bbd23c9a5b2f53e53a74f0cd626519414c52596cda23b50
3
+ size 84908185
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00012-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7503c2d5f2dc9b89f114cfcff0fb743889f964092e2f5968871c6bdbc88c1bdb
3
+ size 89176254
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00013-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8fbddd1caf36b0ff3a38c286a9e8eec16a871597b48d6e3c9da2b084c48b2c52
3
+ size 114805076
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00014-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e1b6af96ee0e51160fd2689328b17823ef48ffbf20df6f952444e46167160e7
3
+ size 95475502
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00015-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbfb4a5e2aba580b72160247a5c58b322709a0c1cc190515d422f0edb93550a5
3
+ size 81554858
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00016-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5b341966a526e4635980a922745f694d3f3920434aa6b86bb6f2acb33631640b
3
+ size 110277731
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00017-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4bbcc4144883b3067649dfae8fb68869484fab633e240424e06832ad16349dd2
3
+ size 78704656
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00018-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc7a10ae81f09569915f9b6361432767dc2774c5fd7eb0912c0d16cb0a2100f3
3
+ size 63408234
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00019-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff94d9a812f700b1dce07bc02f945a80fab38040ec74397cbbca55357efa0ea3
3
+ size 105894063
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00020-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b59da0e5e5e0d6fcf0e6a5804b855e0c46ad92b55d8026a48566eccb05176704
3
+ size 76583201
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00021-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a74eee19258cd41f54a441aef7cd056254055c4da5ded64a8c9fe37eb35e461d
3
+ size 110517730
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00022-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ba5796a437a115db154e22f4e9afb830e45f2d0e78a04b03145487ffa783e1c
3
+ size 45812464
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00023-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1f10dfec9f6461ebac79b226e63fe11e6d400e8e02d709bd0723223fe525d9a
3
+ size 55270686
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00024-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2c431bee11c14d9effc53d333990bb7b776ba756b40d496567ce063632da6a0
3
+ size 84701643
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00025-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28330fd7e4ca471280a32211e3094a2f6b558827256ad1c3660d7caa6bd78c08
3
+ size 92988962
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00026-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c95e3dd5c7030426aeef83e72143f760b74955f78b23400059f0c91c26c768d
3
+ size 94071233
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00027-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c6a479f8db72b71be80de627a791df751e15900284629af30729f284e146dd7c
3
+ size 113335020
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00028-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e790aaf193cad67953c13b747418cc2344ee07d60604e16a5aa955ebdc4e94c0
3
+ size 137528740
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00029-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e0ff1e09af4077d3a9ffd3d29527c4ac9a25ec071e87be22f186dd8e964a3bc3
3
+ size 52699481
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00030-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fba91ecff98dc38a8681cf4ddb0c6daf42ab8a4cb2afea7b55827927fe725933
3
+ size 103266172
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00031-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1549c27612fd63732d8a5589ff97278b29954e981177234f405a3c34607019b6
3
+ size 85443895
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00032-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7ea71bf3f02d9b768f8423832ae0a64acb27c68f1b25b404eaf624775b807878
3
+ size 68348857
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00033-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fbd19b0ec4188195392023287ca717b2e17541dd57e557ce90aef925a346f7d3
3
+ size 63785408
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00034-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f99263e26a473795b6d4b9b77f8cee997bf36413bf69a95037c9305d53b15f1c
3
+ size 92808896
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00035-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:61cf4c4c49e1a7abd1fb2b23e230e666b45a98493a55265de43b8553b0aba08e
3
+ size 75662744
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00036-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d384a72fd5a49fb2994495a6156a7b66cfd95cf623a74bdc01639cc129abdee
3
+ size 44990016
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00037-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:73daa3f7b749e92796dae26801ca17b0eef2dac369a6e0cbea912fed58bd1957
3
+ size 79859533
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00038-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b208675a9f084397b14cf543b53093f4480f2bddd83713d08c4fa13179bce24
3
+ size 35042611
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00039-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d11809d9f26c11fb42c917b559ad2a819d69a4fd4c2ca87a126cd81baca70f9f
3
+ size 61144328
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00040-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb51dd8c1cb6ba8e3cb14afa25219279065ef476ca639ac41db628d1c3fb29a4
3
+ size 95291196
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00041-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:034aa6da78f50ffab82d8fca185ce1c6666e1ff1c22f247135e12a612b17546d
3
+ size 79976979
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00042-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f837c2c958c31e741d814ef71176422de5f6b726ed638b7507eb7b288bfe1f4f
3
+ size 82407512
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00043-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e03bc40ce3b16c0a2aa804c1b1e269bd18fec3e0288abc16e9307f9f86b6b3d
3
+ size 94213245
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00044-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:690ff6d6d453ac338e51b3cb79f7d8e6400c37a5bce17afcfe279d9db9127676
3
+ size 102179764
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00045-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3adfe5d1fa77d1bc7a880651098d65958af5fd851403c9472ad1cebdfca087ab
3
+ size 63236112
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00046-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:561f2ebd71dfce5234544ae12870c537a72bf38ecf1b54b894a2994108ada044
3
+ size 97892290
desktop/use/train/Qwen3-VL-2B-Instruct/shard-00047-of-00126.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7c014e3592c7c6584e0e398adc936da6c6b5e5a3550eef098b8dcd41f0f6cb12
3
+ size 141017994
stats.json CHANGED
@@ -1,20 +1,30 @@
1
  {
2
  "rows_in": 0,
3
- "rows_out": 14636,
4
  "rows_dropped": 0,
5
- "unique_images": 237207,
6
- "image_store_bytes": 52094838627,
7
  "by_partition": {
8
- "web::use::train::gpt-5.5": 1330,
9
- "web::use::train::Qwen3.5-27B": 1328,
10
- "web::use::train::Qwen3.5-2B": 1330,
11
- "web::use::train::Qwen3.5-4B": 1332,
12
- "web::use::train::Qwen3.5-9B": 1324,
13
- "web::use::train::Qwen3-VL-2B-Instruct": 1334,
14
- "web::use::train::Qwen3-VL-32B-Instruct": 1333,
15
- "web::use::train::Qwen3-VL-4B-Instruct": 1334,
16
- "web::use::train::Qwen3-VL-8B-Instruct": 1330,
17
- "web::use::train::UI-TARS-1.5-7B": 1332,
18
- "web::use::train::UI-TARS-7B-DPO": 1329
 
 
 
 
 
 
 
 
 
 
19
  }
20
- }
 
1
  {
2
  "rows_in": 0,
3
+ "rows_out": 29146,
4
  "rows_dropped": 0,
5
+ "unique_images": 490927,
6
+ "image_store_bytes": 168205549265,
7
  "by_partition": {
8
+ "desktop::use::train::Qwen3-VL-2B-Instruct": 1451,
9
+ "desktop::use::train::Qwen3-VL-32B-Instruct": 1452,
10
+ "desktop::use::train::Qwen3-VL-4B-Instruct": 1452,
11
+ "desktop::use::train::Qwen3-VL-8B-Instruct": 1453,
12
+ "desktop::use::train::Qwen3.5-27B": 1449,
13
+ "desktop::use::train::Qwen3.5-2B": 1447,
14
+ "desktop::use::train::Qwen3.5-4B": 1454,
15
+ "desktop::use::train::Qwen3.5-9B": 1451,
16
+ "desktop::use::train::UI-TARS-1.5-7B": 1452,
17
+ "desktop::use::train::UI-TARS-7B-DPO": 1449,
18
+ "web::use::train::web.use.gpt_5_5": 1330,
19
+ "web::use::train::web.use.qwen3_5_27b": 1328,
20
+ "web::use::train::web.use.qwen3_5_2b": 1330,
21
+ "web::use::train::web.use.qwen3_5_4b": 1332,
22
+ "web::use::train::web.use.qwen3_5_9b": 1324,
23
+ "web::use::train::web.use.qwen3_vl_2b": 1334,
24
+ "web::use::train::web.use.qwen3_vl_32b": 1333,
25
+ "web::use::train::web.use.qwen3_vl_4b": 1334,
26
+ "web::use::train::web.use.qwen3_vl_8b": 1330,
27
+ "web::use::train::web.use.ui_tars_15_7b": 1332,
28
+ "web::use::train::web.use.ui_tars_7b_dpo": 1329
29
  }
30
+ }