arminmehrabian commited on
Commit
1a10727
·
verified ·
1 Parent(s): f59ceed

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. README.md +67 -23
  2. config.pt +2 -2
  3. pytorch_model.bin +1 -1
README.md CHANGED
@@ -1,17 +1,3 @@
1
- ---
2
- license: apache-2.0
3
- language:
4
- - en
5
- base_model:
6
- - nasa-impact/nasa-smd-ibm-st-v2
7
- tags:
8
- - GNN
9
- - NASA
10
- - EOSDIS
11
- - Earth Science
12
- datasets:
13
- - nasa-gesdisc/nasa-eo-knowledge-graph
14
- ---
15
  # EOSDIS Graph Neural Network Model Card
16
 
17
  ## Model Overview
@@ -23,8 +9,65 @@ datasets:
23
  **Base Language Model**: nasa-impact/nasa-smd-ibm-st-v2
24
 
25
  ## Model Architecture
 
26
  ### Architecture Diagram
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
27
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
28
 
29
  ### Core Components
30
  - **Base Text Encoder**: NASA-SMD-IBM Language Model (768-dimensional embeddings)
@@ -46,10 +89,10 @@ datasets:
46
  - **Source**: NASA EOSDIS Knowledge Graph
47
  - **Node Types and Counts**:
48
  - Datasets: Earth science datasets from NASA DAACs
49
- - Publications: Publications citing NASA EOSDIS datsets and publications citing those publications
50
  - Instruments: Earth observation instruments
51
  - Platforms: Satellite and other observation platforms
52
- - Science Keywords: NASA GCMD Science Keywords
53
 
54
  ### Training Process
55
  - **Optimization**: Adam optimizer
@@ -103,7 +146,7 @@ from gnn_model import EOSDIS_GNN
103
  # Load models
104
  tokenizer = AutoTokenizer.from_pretrained("nasa-impact/nasa-smd-ibm-st-v2")
105
  text_model = AutoModel.from_pretrained("nasa-impact/nasa-smd-ibm-st-v2")
106
- gnn_model = EOSDIS_GNN.from_pretrained("arminmehrabian/nasa-eosdis-heterogeneous-gnn")
107
 
108
  # Process query
109
  def get_embedding(text):
@@ -167,15 +210,16 @@ results = searcher.search(
167
  ### Citation
168
 
169
  ```bibtex
170
- @misc{eosdis-gnn-2025,
171
  title={EOSDIS Graph Neural Network Model},
172
- author={Armin Mehrabian},
173
- year={2025},
174
  publisher={Hugging Face},
175
- howpublished={\url{https://huggingface.co/arminmehrabian/nasa-eosdis-heterogeneous-gnn}}
176
  }
177
  ```
178
 
179
  ## Contact Information
180
- - **Maintainer**: [Armin Mehrabian]
181
- - **Email**: [[email protected]]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  # EOSDIS Graph Neural Network Model Card
2
 
3
  ## Model Overview
 
9
  **Base Language Model**: nasa-impact/nasa-smd-ibm-st-v2
10
 
11
  ## Model Architecture
12
+
13
  ### Architecture Diagram
14
+ ```mermaid
15
+ graph TB
16
+ subgraph Input Layer
17
+ NASA[NASA Language Model<br/>768d] --> EMB[Text Embeddings]
18
+ EMB --> DS[Dataset Nodes]
19
+ EMB --> PUB[Publication Nodes]
20
+ EMB --> INST[Instrument Nodes]
21
+ EMB --> PLAT[Platform Nodes]
22
+ EMB --> KEY[Science Keyword Nodes]
23
+ end
24
+
25
+ subgraph GNN Layers
26
+ L1[GNN Layer 1<br/>768d → 256d]
27
+ L2[GNN Layer 2<br/>256d → 256d]
28
+ L3[GNN Layer 3<br/>256d → 768d]
29
+
30
+ DS --> L1
31
+ PUB --> L1
32
+ INST --> L1
33
+ PLAT --> L1
34
+ KEY --> L1
35
+
36
+ L1 --> L2
37
+ L2 --> L3
38
+ end
39
+
40
+ subgraph Output Layer
41
+ L3 --> DSO[Dataset Embeddings]
42
+ L3 --> PUBO[Publication Embeddings]
43
+ L3 --> INSTO[Instrument Embeddings]
44
+ L3 --> PLATO[Platform Embeddings]
45
+ L3 --> KEYO[Keyword Embeddings]
46
+ end
47
+
48
+ style NASA fill:#f9f,stroke:#333,stroke-width:2px
49
+ style L1 fill:#bbf,stroke:#333,stroke-width:2px
50
+ style L2 fill:#bbf,stroke:#333,stroke-width:2px
51
+ style L3 fill:#bbf,stroke:#333,stroke-width:2px
52
+ ```
53
 
54
+ ### Edge Types
55
+ ```mermaid
56
+ graph LR
57
+ DS[Dataset] -->|references| PUB[Publication]
58
+ DS -->|uses| INST[Instrument]
59
+ DS -->|platform| PLAT[Platform]
60
+ DS -->|tagged_with| KEY[Science Keyword]
61
+ INST -->|on| PLAT
62
+ PUB -->|mentions| INST
63
+ PUB -->|discusses| KEY
64
+
65
+ style DS fill:#f9f,stroke:#333,stroke-width:2px
66
+ style PUB fill:#bbf,stroke:#333,stroke-width:2px
67
+ style INST fill:#bfb,stroke:#333,stroke-width:2px
68
+ style PLAT fill:#fbb,stroke:#333,stroke-width:2px
69
+ style KEY fill:#ffb,stroke:#333,stroke-width:2px
70
+ ```
71
 
72
  ### Core Components
73
  - **Base Text Encoder**: NASA-SMD-IBM Language Model (768-dimensional embeddings)
 
89
  - **Source**: NASA EOSDIS Knowledge Graph
90
  - **Node Types and Counts**:
91
  - Datasets: Earth science datasets from NASA DAACs
92
+ - Publications: Related scientific papers
93
  - Instruments: Earth observation instruments
94
  - Platforms: Satellite and other observation platforms
95
+ - Science Keywords: NASA Earth Science taxonomy
96
 
97
  ### Training Process
98
  - **Optimization**: Adam optimizer
 
146
  # Load models
147
  tokenizer = AutoTokenizer.from_pretrained("nasa-impact/nasa-smd-ibm-st-v2")
148
  text_model = AutoModel.from_pretrained("nasa-impact/nasa-smd-ibm-st-v2")
149
+ gnn_model = EOSDIS_GNN.from_pretrained("your-username/eosdis-gnn")
150
 
151
  # Process query
152
  def get_embedding(text):
 
210
  ### Citation
211
 
212
  ```bibtex
213
+ @misc{eosdis-gnn-2024,
214
  title={EOSDIS Graph Neural Network Model},
215
+ author={Your Name},
216
+ year={2024},
217
  publisher={Hugging Face},
218
+ howpublished={\url{https://huggingface.co/your-username/eosdis-gnn}}
219
  }
220
  ```
221
 
222
  ## Contact Information
223
+ - **Maintainer**: [Your Name]
224
+ - **Email**: [Your Email]
225
+ - **Organization**: [Your Organization]
config.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e838a96fe36ddcd3e4d2cf8be2a6fa41c27b40cc3933194d5877f94cf8a3cc0b
3
- size 860
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b56569d9bd26838846528abb38bf37f374748b9ee790e5a37e189657824cc0de
3
+ size 1372
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7fb26177ceb863df2c3b1fc94c0ecf691cff19abcc515c163f8eee2cba81e080
3
  size 27120886
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04a0203feb172b10c369032acfa1cf46d419bf0ba744dfcf9f7644e67f0191bb
3
  size 27120886