---
language: en
license: mit
tags:
- regression
- soulprint
- tamu
- xgboost
- embeddings
datasets:
- custom
metrics:
- mse
- r2
model-index:
- name: Tamu-xgb-model
  results:
  - task:
      type: regression
      name: Predicting Tamu Scores
    dataset:
      name: Soulprint Tamu Dataset
      type: custom
      size: 912
    metrics:
    - name: MSE
      type: mse
      value: 0.0167
    - name: R²
      type: r2
      value: 0.803
---




# Tamu XGBoost Regression Model

## Overview
The **Tamu Regression Model** is part of the Soulprint archetype system, designed to measure expressions of *lightness, uplift, and shared resonance* in text.  
It was trained on a **balanced dataset of 912 rows**, evenly distributed across three continuous output bins:

- **Low (0.00–0.33)**: minimal energy, muted or subdued responses  
- **Mid (0.34–0.66)**: moderate energy, rhythmic or collective responses  
- **High (0.67–1.00)**: elevated energy, loud or vibrant expressions  

The model outputs a **continuous score between 0.00 and 1.00**, where higher values correspond to stronger expressions of Tamu energy.

---

## Training Details
- **Dataset size:** 912 rows (balanced: 304 per bin)  
- **Embedding model:** `sentence-transformers/all-mpnet-base-v2`  
- **Regressor:** XGBoost Regressor (`reg:squarederror`)  
- **Metrics achieved:**  
  - **MSE:** 0.0167  
  - **R²:** 0.803  

---

## Usage
### Inference Example
```python
import xgboost as xgb
from sentence_transformers import SentenceTransformer
from huggingface_hub import hf_hub_download

# -----------------------------
# 1. Download model from Hugging Face Hub
# -----------------------------
REPO_ID = "mjpsm/Tamu-xgb-model"
FILENAME = "Tamu_xgb_model.json"

model_path = hf_hub_download(repo_id=REPO_ID, filename=FILENAME)

# -----------------------------
# 2. Load model + embedder
# -----------------------------
model = xgb.XGBRegressor()
model.load_model(model_path)

embedder = SentenceTransformer("all-mpnet-base-v2")

# -----------------------------
# 3. Example prediction
# -----------------------------
text = "Inside the library, the pages turned slowly as students whispered."
embedding = embedder.encode([text])
score = model.predict(embedding)[0]

print("Predicted Tamu Score:", round(float(score), 3))
```