File size: 1,742 Bytes
26d068d
 
 
 
d2232af
26d068d
 
 
 
d2232af
 
 
405d07f
26d068d
 
d2232af
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
---
library_name: transformers
license: apache-2.0
base_model: distilbert-base-uncased
language: en
tags:
  - text-classification
  - bloom
  - check-in-quality
  - transformers
  - fastapi
datasets:
  - user6295018/checkin-quality-dataset
metrics:
  - accuracy
  - f1
  - precision
  - recall
pipeline_tag: text-classification
model-index:
  - name: Bloom Check-in Quality Classifier
    results: []
---

# 🌸 Bloom Check-in Quality Classifier

The **Bloom Check-in Quality Classifier** is a fine-tuned `DistilBERT` model designed to analyze daily check-ins from the *Coding in Color* program and classify them into one of three categories:

- **Descriptive** — Clear, thoughtful, and specific check-ins  
- **Neutral** — Somewhat informative but missing depth  
- **Vague** — Minimal or unclear updates  

This model powers Bloom AI’s productivity assistant, which helps students reflect on their daily work habits and communicate effectively.

---

## 🧠 Model Details

- **Base model:** `distilbert-base-uncased`
- **Framework:** 🤗 Transformers + PyTorch
- **Language:** English
- **Task:** Text Classification
- **Labels:** `["vague", "neutral", "descriptive"]`

---

## 📊 Training Information

- **Dataset:** 1,200+ anonymized check-ins from the Coding in Color program  
- **Split:** 80% train / 10% validation / 10% test  
- **Epochs:** 3  
- **Batch size:** 16  
- **Optimizer:** AdamW  
- **Learning rate:** 5e-5  

---

## ⚙️ Inference Example

```python
from transformers import pipeline

classifier = pipeline("text-classification", model="user6295018/checkin-quality-classifier")

classifier("Had a really productive day working on my API and debugging the UI.")
# [{'label': 'descriptive', 'score': 0.94}]