Update model card
Browse files
README.md
CHANGED
|
@@ -88,6 +88,14 @@ columns the model expects. The canonical definition lives in
|
|
| 88 |
`src/engagement_predictor/features.py` in the application repository
|
| 89 |
(https://github.com/<your-org>/viralst).
|
| 90 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 91 |
## How to use
|
| 92 |
|
| 93 |
```python
|
|
@@ -105,7 +113,8 @@ prediction = model.predict(X)
|
|
| 105 |
|
| 106 |
## Limitations
|
| 107 |
|
| 108 |
-
-
|
|
|
|
| 109 |
- Feature quality depends on upstream LLM extraction from brief documents;
|
| 110 |
prediction quality degrades if that extraction is inaccurate or incomplete.
|
| 111 |
- Not yet validated against real campaign data from brands outside the
|
|
|
|
| 88 |
`src/engagement_predictor/features.py` in the application repository
|
| 89 |
(https://github.com/<your-org>/viralst).
|
| 90 |
|
| 91 |
+
## Training dataset
|
| 92 |
+
|
| 93 |
+
The raw synthetic dataset used to train this model is included in this repo
|
| 94 |
+
under `dataset/` (`brands.csv`, `briefs.csv`, `contents.csv` — 30
|
| 95 |
+
brands, 300 briefs, 2000 content rows). It was generated by
|
| 96 |
+
`training/generate_poc_data.py`; see "Training data and methodology" above
|
| 97 |
+
for why it is synthetic and what it does and does not represent.
|
| 98 |
+
|
| 99 |
## How to use
|
| 100 |
|
| 101 |
```python
|
|
|
|
| 113 |
|
| 114 |
## Limitations
|
| 115 |
|
| 116 |
+
- Trained on synthetic data; does not reflect real-world engagement
|
| 117 |
+
distributions, platform trends, or actual creator/audience behavior.
|
| 118 |
- Feature quality depends on upstream LLM extraction from brief documents;
|
| 119 |
prediction quality degrades if that extraction is inaccurate or incomplete.
|
| 120 |
- Not yet validated against real campaign data from brands outside the
|