XuHuang commited on
Commit
812e683
·
verified ·
1 Parent(s): 8e17209

Upload folder using huggingface_hub

Browse files
dataset_dict.json CHANGED
@@ -1 +1 @@
1
- {"splits": ["train", "test"]}
 
1
+ {"splits": ["train"]}
train/cache-216aef9dfde83398.arrow ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ddac6739f6619002e7d2845c56b164fe1224aec5fd8bd149f959cbc3f4088d9
3
+ size 258481888
train/cache-21b92fc929a767e9.arrow ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a1b0eff29e32c30ccdca04a30898f1d224f98fa6e29e1ce42550dca54b9fb4b
3
+ size 258481888
train/cache-26c2b0cfe812003e.arrow ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c06f7f9f944a9195c65fee50ac532bd5e68f2712123977fb363744d2f8b404f5
3
+ size 258481888
train/cache-b0e753f295f2ae1e.arrow ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:057d540923c7fab7fdd68e2a065b1cc746ea9ed24c8e0322159718701629c876
3
+ size 258481888
train/cache-d7b45486acc59d4d.arrow ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d18f02112c79f0fc5b857711c76098f3aa0b6cc9256e06a0879b19a53456814
3
+ size 258481888
train/data-00000-of-00001.arrow ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b9f2f49738dbffb8e83a77b420bc1e65b2b094f6d8e83449868e7d9b7ef4bf48
3
+ size 242670872
train/dataset_info.json CHANGED
@@ -1,6 +1,17 @@
1
  {
 
2
  "citation": "",
 
 
 
3
  "description": "",
 
 
 
 
 
 
 
4
  "features": {
5
  "prompt": {
6
  "dtype": "string",
@@ -11,14 +22,14 @@
11
  "dtype": "string",
12
  "_type": "Value"
13
  },
14
- "_type": "Sequence"
15
  },
16
  "all_rm_scores": {
17
  "feature": {
18
  "dtype": "float64",
19
  "_type": "Value"
20
  },
21
- "_type": "Sequence"
22
  },
23
  "chosen": {
24
  "dtype": "string",
@@ -33,42 +44,42 @@
33
  "dtype": "int64",
34
  "_type": "Value"
35
  },
36
- "_type": "Sequence"
37
  },
38
  "chosen_attention_mask": {
39
  "feature": {
40
  "dtype": "int64",
41
  "_type": "Value"
42
  },
43
- "_type": "Sequence"
44
  },
45
  "chosen_labels": {
46
  "feature": {
47
  "dtype": "int64",
48
  "_type": "Value"
49
  },
50
- "_type": "Sequence"
51
  },
52
  "rejected_input_ids": {
53
  "feature": {
54
  "dtype": "int64",
55
  "_type": "Value"
56
  },
57
- "_type": "Sequence"
58
  },
59
  "rejected_attention_mask": {
60
  "feature": {
61
  "dtype": "int64",
62
  "_type": "Value"
63
  },
64
- "_type": "Sequence"
65
  },
66
  "rejected_labels": {
67
  "feature": {
68
  "dtype": "int64",
69
  "_type": "Value"
70
  },
71
- "_type": "Sequence"
72
  },
73
  "reference_chosen_logps": {
74
  "dtype": "float32",
@@ -85,8 +96,31 @@
85
  "history0_rejected_logps": {
86
  "dtype": "float32",
87
  "_type": "Value"
 
 
 
 
 
 
 
 
88
  }
89
  },
90
  "homepage": "",
91
- "license": ""
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
92
  }
 
1
  {
2
+ "builder_name": "json",
3
  "citation": "",
4
+ "config_name": "default",
5
+ "dataset_name": "json",
6
+ "dataset_size": 88858120,
7
  "description": "",
8
+ "download_checksums": {
9
+ "/hai/scratch/fangwu97/xu/SimPO_slurm/datasets/gemma2_ultrafeedback/mnpo_iter2_skywork_scored.jsonl": {
10
+ "num_bytes": 91582104,
11
+ "checksum": null
12
+ }
13
+ },
14
+ "download_size": 91582104,
15
  "features": {
16
  "prompt": {
17
  "dtype": "string",
 
22
  "dtype": "string",
23
  "_type": "Value"
24
  },
25
+ "_type": "List"
26
  },
27
  "all_rm_scores": {
28
  "feature": {
29
  "dtype": "float64",
30
  "_type": "Value"
31
  },
32
+ "_type": "List"
33
  },
34
  "chosen": {
35
  "dtype": "string",
 
44
  "dtype": "int64",
45
  "_type": "Value"
46
  },
47
+ "_type": "List"
48
  },
49
  "chosen_attention_mask": {
50
  "feature": {
51
  "dtype": "int64",
52
  "_type": "Value"
53
  },
54
+ "_type": "List"
55
  },
56
  "chosen_labels": {
57
  "feature": {
58
  "dtype": "int64",
59
  "_type": "Value"
60
  },
61
+ "_type": "List"
62
  },
63
  "rejected_input_ids": {
64
  "feature": {
65
  "dtype": "int64",
66
  "_type": "Value"
67
  },
68
+ "_type": "List"
69
  },
70
  "rejected_attention_mask": {
71
  "feature": {
72
  "dtype": "int64",
73
  "_type": "Value"
74
  },
75
+ "_type": "List"
76
  },
77
  "rejected_labels": {
78
  "feature": {
79
  "dtype": "int64",
80
  "_type": "Value"
81
  },
82
+ "_type": "List"
83
  },
84
  "reference_chosen_logps": {
85
  "dtype": "float32",
 
96
  "history0_rejected_logps": {
97
  "dtype": "float32",
98
  "_type": "Value"
99
+ },
100
+ "history1_chosen_logps": {
101
+ "dtype": "float32",
102
+ "_type": "Value"
103
+ },
104
+ "history1_rejected_logps": {
105
+ "dtype": "float32",
106
+ "_type": "Value"
107
  }
108
  },
109
  "homepage": "",
110
+ "license": "",
111
+ "size_in_bytes": 180440224,
112
+ "splits": {
113
+ "train": {
114
+ "name": "train",
115
+ "num_bytes": 88858120,
116
+ "num_examples": 6598,
117
+ "dataset_name": "json"
118
+ }
119
+ },
120
+ "version": {
121
+ "version_str": "0.0.0",
122
+ "major": 0,
123
+ "minor": 0,
124
+ "patch": 0
125
+ }
126
  }
train/state.json CHANGED
@@ -1,13 +1,10 @@
1
  {
2
  "_data_files": [
3
  {
4
- "filename": "data-00000-of-00002.arrow"
5
- },
6
- {
7
- "filename": "data-00001-of-00002.arrow"
8
  }
9
  ],
10
- "_fingerprint": "16715b50e5b3cc4a",
11
  "_format_columns": null,
12
  "_format_kwargs": {},
13
  "_format_type": null,
 
1
  {
2
  "_data_files": [
3
  {
4
+ "filename": "data-00000-of-00001.arrow"
 
 
 
5
  }
6
  ],
7
+ "_fingerprint": "2a7a396ab75a6dad",
8
  "_format_columns": null,
9
  "_format_kwargs": {},
10
  "_format_type": null,