SwathiManikya commited on
Commit
6bb25a7
·
verified ·
1 Parent(s): 5b03f11

Delete dataset_card.yaml

Browse files
Files changed (1) hide show
  1. dataset_card.yaml +0 -152
dataset_card.yaml DELETED
@@ -1,152 +0,0 @@
1
- dataset_name: hlo-feature-dataset
2
- pretty_name: HLO Feature Dataset for Deep Learning Resource Estimation
3
- dataset_type: graph-and-tabular
4
- license: MIT
5
- task_categories:
6
- - time-series-forecasting
7
- - regression
8
- - graph-machine-learning
9
- language: en
10
- tags:
11
- - HPC
12
- - resource-prediction
13
- - XLA
14
- - compiler-features
15
- - deep-learning
16
- - graph-learning
17
- - scheduling
18
- size_categories:
19
- - 1K<n<10K
20
- source_datasets:
21
- - custom
22
- dataset_summary: The HLO Feature Dataset contains High-Level Optimizer (HLO) graph
23
- features and metadata extracted from deep learning training workloads. It is designed
24
- for tasks such as runtime prediction, resource estimation, and graph-based machine
25
- learning in HPC environments.
26
- structured_data:
27
- features:
28
- - name: name
29
- type: string
30
- description: ''
31
- - name: samples
32
- type: float
33
- description: ''
34
- - name: input_dim_w
35
- type: float
36
- description: ''
37
- - name: input_dim_h
38
- type: float
39
- description: ''
40
- - name: input_dim_c
41
- type: float
42
- description: ''
43
- - name: output_dim
44
- type: float
45
- description: ''
46
- - name: optimizer
47
- type: string
48
- description: ''
49
- - name: epochs
50
- type: float
51
- description: ''
52
- - name: batch
53
- type: float
54
- description: ''
55
- - name: learn_rate
56
- type: float
57
- description: ''
58
- - name: tf_version
59
- type: string
60
- description: ''
61
- - name: cuda_version
62
- type: string
63
- description: ''
64
- - name: batch_time
65
- type: float
66
- description: ''
67
- - name: epoch_time
68
- type: float
69
- description: ''
70
- - name: fit_time
71
- type: float
72
- description: ''
73
- - name: npz_path
74
- type: string
75
- description: ''
76
- - name: gpu_make
77
- type: string
78
- description: ''
79
- - name: gpu_name
80
- type: string
81
- description: ''
82
- - name: gpu_arch
83
- type: string
84
- description: ''
85
- - name: gpu_cc
86
- type: string
87
- description: ''
88
- - name: gpu_core_count
89
- type: string
90
- description: ''
91
- - name: gpu_sm_count
92
- type: string
93
- description: ''
94
- - name: gpu_memory_size
95
- type: string
96
- description: ''
97
- - name: gpu_memory_type
98
- type: string
99
- description: ''
100
- - name: gpu_memory_bw
101
- type: string
102
- description: ''
103
- - name: gpu_tensor_core_count
104
- type: string
105
- description: ''
106
- - name: max_memory_util
107
- type: float
108
- description: ''
109
- - name: avg_memory_util
110
- type: float
111
- description: ''
112
- - name: max_gpu_util
113
- type: string
114
- description: ''
115
- - name: avg_gpu_util
116
- type: string
117
- description: ''
118
- - name: max_gpu_temp
119
- type: string
120
- description: ''
121
- - name: avg_gpu_temp
122
- type: string
123
- description: ''
124
- graph_data:
125
- node_features: node_feat
126
- edge_index: edge_index
127
- additional_keys:
128
- - node_opcode
129
- - node_config_ids
130
- - node_splits
131
- usage_example: '```python
132
-
133
- from datasets import load_dataset
134
-
135
- import numpy as np
136
-
137
-
138
- dataset = load_dataset(''your-username/hlo-feature-dataset'')
139
-
140
- sample = dataset[''train''][0]
141
-
142
-
143
- graph_data = np.load(sample[''npz_path''])
144
-
145
- node_features = graph_data[''node_feat'']
146
-
147
- edges = graph_data[''edge_index'']
148
-
149
- ```'
150
- citation: "@misc{hlofeatures2025,\n title={HLO Feature Dataset for AI Resource Estimation},\n\
151
- \ author={Your Name},\n year={2025},\n url={https://huggingface.co/datasets/your-username/hlo-feature-dataset}\n\
152
- }"