mpfbench99 commited on
Commit
125712c
Β·
verified Β·
1 Parent(s): 56eddaa

Updating ReadMe

Browse files
Files changed (1) hide show
  1. README.md +295 -3
README.md CHANGED
@@ -1,3 +1,295 @@
1
- ---
2
- license: cc-by-nc-sa-4.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-nc-4.0
3
+ language:
4
+ - en
5
+ tags:
6
+ - multiphase flows
7
+ size_categories:
8
+ - 10K<n<100K
9
+ ---
10
+ **Overview**
11
+ Scientific machine learning (SciML) is a promising strategy for designing multiphase flow solvers, but it requires a large dataset. This paper presents a comprehensive dataset created from 11,000 simulations, both in 2D and 3D, using the Lattice Boltzmann method~(LBM). The dataset captures intricate physics by varying factors such as surface tension, density, and viscosity of fluids. These simulations, comprising 1 million time snapshots, provide extensive data on two-fluid behavior. By making this dataset publicly available, we aim to encourage SciML research and its applications in complex fluid systems, facilitating the creation of more precise and efficient SciML frameworks for multiphysics applications. Our dataset spans multiple orders of magnitude of Reynolds and Bond numbers, density, and viscosity ratios, making it exceptionally rich and valuable for understanding multiphase flows.
12
+
13
+
14
+ **Dataset Information**
15
+
16
+ Our library contains simulated data from both 2d and 3d simulaions. The library has four categories (with corresponding sample sizes):
17
+
18
+ ```2d bubble - 5000```
19
+ ```2d drop - 5000```
20
+ ```3d bubble - 500```
21
+ ```3d drop - 500```
22
+
23
+ Due to size of the datasets and to facilitate easier downloads, we have divided the simulations under each category into groups. Further, we have created tar files (for 2d) and splitted tar files (for 3d) to reduce the download sizes. The following steps show how to extract these into ```*.npz.tar.gz``` files.
24
+
25
+ ***2d***
26
+ ```
27
+ tar -xzvf dropGroup1_npz_*.tar.gz
28
+ tar -xzvf dropGroup2_npz_*.tar.gz
29
+ tar -xzvf dropGroup3_npz_*.tar.gz
30
+ tar -xzvf dropGroup4_npz_*.tar.gz
31
+ tar -xzvf dropGroup5_npz_*.tar.gz
32
+
33
+ tar -xzvf bubbleGroup1_npz_*.tar.gz
34
+ tar -xzvf bubbleGroup2_npz_*.tar.gz
35
+ tar -xzvf bubbleGroup3_npz_*.tar.gz
36
+ tar -xzvf bubbleGroup4_npz_*.tar.gz
37
+ tar -xzvf bubbleGroup5_npz_*.tar.gz
38
+ ```
39
+
40
+
41
+ ***3d***
42
+ ```
43
+ Step - 1 : Merge split tarred files into tar.gz files
44
+
45
+ cat 3DbubbleGroup1_npz_* > 3DbubbleGroup1_npz.tar.gz
46
+ cat 3DbubbleGroup2_npz_* > 3DbubbleGroup2_npz.tar.gz
47
+ cat 3DbubbleGroup3_npz_* > 3DbubbleGroup3_npz.tar.gz
48
+ cat 3DbubbleGroup4_npz_* > 3DbubbleGroup4_npz.tar.gz
49
+ cat 3DbubbleGroup5_npz_* > 3DbubbleGroup5_npz.tar.gz
50
+
51
+ cat 3DdropGroup1_npz_* > 3DdropGroup1_npz.tar.gz
52
+ cat 3DdropGroup2_npz_* > 3DdropGroup2_npz.tar.gz
53
+ cat 3DdropGroup3_npz_* > 3DdropGroup3_npz.tar.gz
54
+ cat 3DdropGroup4_npz_* > 3DdropGroup4_npz.tar.gz
55
+ cat 3DdropGroup5_npz_* > 3DdropGroup5_npz.tar.gz
56
+
57
+ Step 2 : Create the *npz.tar.gz files
58
+
59
+ tar -xzvf 3DdropGroup1_npz.tar.gz
60
+ tar -xzvf 3DdropGroup2_npz.tar.gz
61
+ tar -xzvf 3DdropGroup3_npz.tar.gz
62
+ tar -xzvf 3DdropGroup4_npz.tar.gz
63
+ tar -xzvf 3DdropGroup5_npz.tar.gz
64
+
65
+ tar -xzvf 3DbubbleGroup1_npz.tar.gz
66
+ tar -xzvf 3DbubbleGroup2_npz.tar.gz
67
+ tar -xzvf 3DbubbleGroup3_npz.tar.gz
68
+ tar -xzvf 3DbubbleGroup4_npz.tar.gz
69
+ tar -xzvf 3DbubbleGroup5_npz.tar.gz
70
+
71
+ ```
72
+
73
+ Refer to the directory structure section below to understand how these files have been staged inside huggingface.
74
+
75
+ **Data Preparation**
76
+
77
+ Please refer to our [github](https://github.com/baskargroup/mpfbench-tools/tree/master/DataPrep) page that provides codes for doing data preparation for SciML models:
78
+
79
+ ```
80
+ https://github.com/baskargroup/mpfbench-tools/tree/master/DataPrep
81
+ ```
82
+
83
+ **License**
84
+ ```
85
+ CC-BY-NC-4.0
86
+ ```
87
+
88
+ **How to download files from Huggingface**
89
+
90
+ To run the example code, you need to install the following package:
91
+
92
+ ```bash
93
+ pip install huggingface_hub
94
+ ```
95
+
96
+ The following script demonstrates how to download a directory from the Hugging Face Hub:
97
+
98
+ ```python
99
+ from huggingface_hub import HfApi, hf_hub_download
100
+ import os
101
+ import shutil
102
+
103
+ REPO_ID = "BGLab/mpf-bench"
104
+ DIRECTORY = "2Dbubble/bubbleGroup1_npz"
105
+
106
+ # Initialize the Hugging Face API
107
+ api = HfApi()
108
+
109
+ # List files in the directory
110
+ files_list = api.list_repo_files(repo_id=REPO_ID, repo_type="dataset")
111
+
112
+ # Filter the files in the specified directory
113
+ files_to_download = [f for f in files_list if f.startswith(DIRECTORY)]
114
+
115
+ # Create local directory if it doesn't exist
116
+ os.makedirs(DIRECTORY, exist_ok=True)
117
+
118
+ # Download each file
119
+ for file in files_to_download:
120
+ file_path = hf_hub_download(repo_id=REPO_ID, filename=file, repo_type="dataset")
121
+ # Copy the file to the local directory using shutil.copy2
122
+ shutil.copy2(file_path, os.path.join(DIRECTORY, os.path.basename(file_path)))
123
+
124
+ print("Files downloaded successfully.")
125
+
126
+ ```
127
+
128
+ **Directory Structure**
129
+ ```
130
+ main/
131
+ β”œβ”€β”€ 2Dbubble/
132
+ β”‚ β”œβ”€β”€ bubbleGroup1_npz/
133
+ β”‚ β”‚ β”œβ”€β”€ bubbleGroup1_npz_1.tar.gz
134
+ β”‚ β”‚ β”œβ”€β”€ bubbleGroup1_npz_2.tar.gz
135
+ β”‚ β”‚ β”œβ”€β”€ .
136
+ β”‚ β”‚ β”œβ”€β”€ .
137
+ β”‚ β”‚ β”œβ”€β”€ .
138
+ β”‚ β”‚ └── bubbleGroup1_npz_10.tar.gz
139
+ β”‚ β”œβ”€β”€ bubbleGroup2_npz/
140
+ β”‚ β”‚ β”œβ”€β”€ bubbleGroup2_npz_1.tar.gz
141
+ β”‚ β”‚ β”œβ”€β”€ bubbleGroup2_npz_2.tar.gz
142
+ β”‚ β”‚ β”œβ”€β”€ .
143
+ β”‚ β”‚ β”œβ”€β”€ .
144
+ β”‚ β”‚ β”œβ”€β”€ .
145
+ β”‚ β”‚ └── bubbleGroup2_npz_10.tar.gz
146
+ .
147
+ .
148
+ .
149
+ β”‚ β”œβ”€β”€ bubbleGroup5_npz/
150
+ β”‚ β”‚ β”œβ”€β”€ bubbleGroup5_npz_1.tar.gz
151
+ β”‚ β”‚ β”œβ”€β”€ bubbleGroup5_npz_2.tar.gz
152
+ β”‚ β”‚ β”œβ”€β”€ .
153
+ β”‚ β”‚ β”œβ”€β”€ .
154
+ β”‚ β”‚ β”œβ”€β”€ .
155
+ β”‚ β”‚ └── bubbleGroup5_npz_10.tar.gz
156
+ β”œβ”€β”€ 2Ddrop/
157
+ β”‚ β”œβ”€β”€ dropGroup1_npz/
158
+ β”‚ β”‚ β”œβ”€β”€ dropGroup1_npz_1.tar.gz
159
+ β”‚ β”‚ β”œβ”€β”€ dropGroup1_npz_2.tar.gz
160
+ β”‚ β”‚ β”œβ”€β”€ .
161
+ β”‚ β”‚ β”œβ”€β”€ .
162
+ β”‚ β”‚ β”œβ”€β”€ .
163
+ β”‚ β”‚ └── dropGroup1_npz_10.tar.gz
164
+ β”‚ β”œβ”€β”€ dropGroup2_npz/
165
+ β”‚ β”‚ β”œβ”€β”€ dropGroup2_npz_1.tar.gz
166
+ β”‚ β”‚ β”œβ”€β”€ dropGroup2_npz_2.tar.gz
167
+ β”‚ β”‚ β”œβ”€β”€ .
168
+ β”‚ β”‚ β”œβ”€β”€ .
169
+ β”‚ β”‚ β”œβ”€β”€ .
170
+ β”‚ β”‚ └── dropGroup2_npz_10.tar.gz
171
+ .
172
+ .
173
+ .
174
+ β”‚ β”œβ”€β”€ dropGroup5_npz/
175
+ β”‚ β”‚ β”œβ”€β”€ dropGroup5_npz_1.tar.gz
176
+ β”‚ β”‚ β”œβ”€β”€ dropGroup5_npz_2.tar.gz
177
+ β”‚ β”‚ β”œβ”€β”€ .
178
+ β”‚ β”‚ β”œβ”€β”€ .
179
+ β”‚ β”‚ β”œβ”€β”€ .
180
+ β”‚ β”‚ └── dropGroup5_npz_10.tar.gz
181
+ β”œβ”€β”€ 3Dbubble/
182
+ β”‚ β”œβ”€β”€ split/
183
+ β”‚ β”‚ β”œβ”€β”€ 3DbubbleGroup1_npz_aa
184
+ β”‚ β”‚ β”œβ”€β”€ 3DbubbleGroup1_npz_ab
185
+ β”‚ β”‚ β”œβ”€β”€ .
186
+ β”‚ β”‚ β”œβ”€β”€ .
187
+ β”‚ β”‚ β”œβ”€β”€ .
188
+ β”‚ β”‚ └── 3DbubbleGroup1_npz_af
189
+ β”‚ β”‚ β”œβ”€β”€ 3DbubbleGroup2_npz_aa
190
+ β”‚ β”‚ β”œβ”€β”€ 3DbubbleGroup2_npz_ab
191
+ β”‚ β”‚ β”œβ”€β”€ .
192
+ β”‚ β”‚ β”œβ”€β”€ .
193
+ β”‚ β”‚ β”œβ”€β”€ .
194
+ β”‚ β”‚ └── 3DbubbleGroup2_npz_af
195
+ .
196
+ .
197
+ .
198
+ β”‚ β”‚ β”œβ”€β”€ 3DbubbleGroup5_npz_aa
199
+ β”‚ β”‚ β”œβ”€β”€ 3DbubbleGroup5_npz_ab
200
+ β”‚ β”‚ β”œβ”€β”€ .
201
+ β”‚ β”‚ β”œβ”€β”€ .
202
+ β”‚ β”‚ β”œβ”€β”€ .
203
+ β”‚ β”‚ └── 3DbubbleGroup5_npz_af
204
+ β”œβ”€β”€ 3Ddrop/
205
+ β”‚ β”œβ”€β”€ split/
206
+ β”‚ β”‚ β”œβ”€β”€ 3DdropGroup1_npz_aa
207
+ β”‚ β”‚ β”œβ”€β”€ 3DdropGroup1_npz_ab
208
+ β”‚ β”‚ β”œβ”€β”€ .
209
+ β”‚ β”‚ β”œβ”€β”€ .
210
+ β”‚ β”‚ β”œβ”€β”€ .
211
+ β”‚ β”‚ └── 3DdropGroup1_npz_af
212
+ β”‚ β”‚ β”œβ”€β”€ 3DdropGroup2_npz_aa
213
+ β”‚ β”‚ β”œβ”€β”€ 3DdropGroup2_npz_ab
214
+ β”‚ β”‚ β”œβ”€β”€ .
215
+ β”‚ β”‚ β”œβ”€β”€ .
216
+ β”‚ β”‚ β”œβ”€β”€ .
217
+ β”‚ β”‚ └── 3DdropGroup2_npz_af
218
+ .
219
+ .
220
+ .
221
+ β”‚ β”‚ β”œβ”€β”€ 3DdropGroup5_npz_aa
222
+ β”‚ β”‚ β”œβ”€β”€ 3DdropGroup5_npz_ab
223
+ β”‚ β”‚ β”œβ”€β”€ .
224
+ β”‚ β”‚ β”œβ”€β”€ .
225
+ β”‚ β”‚ β”œβ”€β”€ .
226
+ β”‚ β”‚ └── 3DdropGroup5_npz_af
227
+ β”œβ”€β”€ Sample Dataset/
228
+ β”‚ β”œβ”€β”€ 2D/
229
+ β”‚ β”‚ β”œβ”€β”€ bubble/
230
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample1/
231
+ β”‚ β”‚ β”‚ β”‚ |── X_bubble_padded_dataset_1.npz
232
+ β”‚ β”‚ β”‚ β”‚ |── Y_bubble_padded_dataset_1.npz
233
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample2/
234
+ β”‚ β”‚ β”‚ β”‚ |── X_bubble_padded_dataset_2.npz
235
+ β”‚ β”‚ β”‚ β”‚ |── Y_bubble_padded_dataset_2.npz
236
+ .
237
+ .
238
+ .
239
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample10/
240
+ β”‚ β”‚ β”‚ β”‚ |── X_bubble_padded_dataset_10.npz
241
+ β”‚ β”‚ β”‚ β”‚ |── Y_bubble_padded_dataset_10.npz
242
+ β”‚ β”‚ β”œβ”€β”€ drop/
243
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample1/
244
+ β”‚ β”‚ β”‚ β”‚ |── X_drop_padded_dataset_1.npz
245
+ β”‚ β”‚ β”‚ β”‚ |── Y_drop_padded_dataset_1.npz
246
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample2/
247
+ β”‚ β”‚ β”‚ β”‚ |── X_drop_padded_dataset_2.npz
248
+ β”‚ β”‚ β”‚ β”‚ |── Y_drop_padded_dataset_2.npz
249
+ .
250
+ .
251
+ .
252
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample10/
253
+ β”‚ β”‚ β”‚ β”‚ |── X_drop_padded_dataset_10.npz
254
+ β”‚ β”‚ β”‚ β”‚ |── Y_drop_padded_dataset_10.npz
255
+ β”‚ β”œβ”€β”€ 3D/
256
+ β”‚ β”‚ β”œβ”€β”€ bubble/
257
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample1/
258
+ β”‚ β”‚ β”‚ β”‚ |── X_bubble_padded_dataset_3d_1.npz
259
+ β”‚ β”‚ β”‚ β”‚ |── Y_bubble_padded_dataset_3d_1.npz
260
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample2/
261
+ β”‚ β”‚ β”‚ β”‚ |── X_bubble_padded_dataset_3d_2.npz
262
+ β”‚ β”‚ β”‚ β”‚ |── Y_bubble_padded_dataset_3d_2.npz
263
+ .
264
+ .
265
+ .
266
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample10/
267
+ β”‚ β”‚ β”‚ β”‚ |── X_bubble_padded_dataset_3d_10.npz
268
+ β”‚ β”‚ β”‚ β”‚ |── Y_bubble_padded_dataset_3d_10.npz
269
+ β”‚ β”‚ β”œβ”€β”€ drop/
270
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample1/
271
+ β”‚ β”‚ β”‚ β”‚ |── X_drop_padded_dataset_3d_1.npz
272
+ β”‚ β”‚ β”‚ β”‚ |── Y_drop_padded_dataset_3d_1.npz
273
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample2/
274
+ β”‚ β”‚ β”‚ β”‚ |── X_drop_padded_dataset_3d_2.npz
275
+ β”‚ β”‚ β”‚ β”‚ |── Y_drop_padded_dataset_3d_2.npz
276
+ .
277
+ .
278
+ .
279
+ β”‚ β”‚ β”‚ β”œβ”€β”€ Sample10/
280
+ β”‚ β”‚ β”‚ β”‚ |── X_drop_padded_dataset_3d_10.npz
281
+ β”‚ β”‚ β”‚ β”‚ |── Y_drop_padded_dataset_3d_10.npz
282
+ β”œβ”€β”€ README.md
283
+ β”œβ”€β”€ .gitattributes
284
+ └── info.txt
285
+ ```
286
+
287
+ **Citation**
288
+ If you find this dataset useful in your research, please consider citing our paper as follows:
289
+ ```
290
+ @article{shadkhah2024MPFBench,
291
+ title = "MPFBench: A Large Scale Dataset for SciML of Multi-Phase-Flows: Droplet and Bubble Dynamics",
292
+ author = "Shadkhah, Mehdi and Tali, Ronak and Rabeh, Ali and Herron, Ethan and Yang, Cheng-Hau and Upadhyaya, Abhisek and Krishnamurthy, Adarsh and Hegde, Chinmay and Balu, Aditya and Ganapathysubramanian, Baskar"
293
+ year = "2024"
294
+ }
295
+ ```