ClergeF commited on
Commit
e0b9ba6
·
verified ·
1 Parent(s): 0c1bb28

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +112 -1
README.md CHANGED
@@ -5,6 +5,117 @@ colorFrom: blue
5
  colorTo: gray
6
  sdk: docker
7
  pinned: false
 
8
  ---
 
9
 
10
- Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5
  colorTo: gray
6
  sdk: docker
7
  pinned: false
8
+ license: apache-2.0
9
  ---
10
+ # MVT Community Value Model API
11
 
12
+ This API powers the **MyVillage Token (MVT)** scoring system.
13
+ It evaluates real-world community actions and predicts:
14
+
15
+ ---
16
+
17
+ ## 1. Estimated Dollar Value
18
+ How financially impactful the action is.
19
+
20
+ ## 2. Impact Level
21
+ Overall significance of the action (0–1 scaled).
22
+
23
+ ## 3. Category Strength Scores (0–1)
24
+ Each action is analyzed across 10 categories:
25
+
26
+ - **Family**
27
+ - **Community**
28
+ - **Education**
29
+ - **Health**
30
+ - **Environment**
31
+ - **Business**
32
+ - **Finance**
33
+ - **History / Heritage**
34
+ - **Spirituality**
35
+ - **Innovation**
36
+ - **Impact (standalone model)**
37
+
38
+ These models were trained using a custom dataset of **2,000+ human-created community actions**, labeled by the MVT team.
39
+
40
+ ---
41
+
42
+ # How the API Works
43
+
44
+ Each model is stored separately on the HuggingFace Hub.
45
+ This Space loads:
46
+
47
+ - A universal text embedder
48
+ - 12 linear regression models trained on SentenceTransformer embeddings
49
+ - Outputs predictions for a single input sentence
50
+
51
+ ---
52
+
53
+ # API Usage
54
+
55
+ ## POST /predict
56
+
57
+ ### Input:
58
+ ```json
59
+ {
60
+ "text": "I helped my neighbor fix her fence after the storm."
61
+ }
62
+ ```
63
+
64
+ ### Output:
65
+ ```json
66
+ {
67
+ "input": "I helped my neighbor fix her fence after the storm.",
68
+ "predictions": {
69
+ "estimated_value": 42.18,
70
+ "impact_level": 0.67,
71
+ "impact": 0.61,
72
+ "family": 0.20,
73
+ "community": 0.88,
74
+ "education": 0.05,
75
+ "health": 0.11,
76
+ "environment": 0.22,
77
+ "business": 0.03,
78
+ "finance": 0.02,
79
+ "history": 0.01,
80
+ "spirituality": 0.04,
81
+ "innovation": 0.08
82
+ }
83
+ }
84
+ ```
85
+
86
+ ---
87
+
88
+ # Running Locally
89
+
90
+ ```
91
+ uvicorn app:app --reload --host 0.0.0.0 --port 7860
92
+ ```
93
+
94
+ Then open:
95
+
96
+ ```
97
+ http://localhost:7860/docs
98
+ ```
99
+
100
+ ---
101
+
102
+ # Models Used
103
+
104
+ Each model is individually hosted under the `ClergeF/` namespace:
105
+
106
+ - `value-impact-model`
107
+ - `impact-model`
108
+ - `family-model`
109
+ - `community-model`
110
+ - `education-model`
111
+ - `health-model`
112
+ - `environment-model`
113
+ - `business-model`
114
+ - `finance-model`
115
+ - `history-model`
116
+ - `spirituality-model`
117
+ - `innovation-model`
118
+
119
+ Embedder sourced from:
120
+
121
+ - `MVT-models/universal_embedder`