How pedros2705/ino544-2026i--carros-deportivos-japoneses- works
DO YOU TENSORFLOW? is a JDM-car image classifier with a browser interface, FastAPI inference endpoints, and a separate PyTorch training script. Users can upload an image or analyze live camera frames; the API passes decoded image bytes to the classifier and returns a label, probabilities, and vehicle specifications. Uploaded predictions and submitted reference images are saved with JSON metadata. Training code documents a dataset-to-model workflow, but its dataset is not in the tree. Edges reflect the shown calls; the HTML’s wiring and any ONNX runtime are not verified.
Main entry points: app.py FastAPI Routes · index.html JDM Web Page · connection.py Image and Record Store · script.js Image Upload
Components 11
Browser Interface
- JDM Web Page
static/index.html - Image Upload
static/script.js - Live Camera
static/camara.js
Inference Service
- FastAPI Routes
app.py
Model Lifecycle
- Image Classifier
utils.py - Model Training
entrenar_ia.py - Model Weights
modelo_jdm_v2.pt
Reference Storage
- Reference Submission
app.py - Image and Record Store
database/connection.py - JSON Records
database/records.json
Other components
- User
Connections 12
- User → JDM Web Page: uses
- JDM Web Page → Image Upload: loads
- JDM Web Page → Live Camera: loads
- Image Upload → FastAPI Routes: posts image
static/script.js - Live Camera → FastAPI Routes: posts frames
static/camara.js - FastAPI Routes → Image Classifier: predicts with
app.py - Image Classifier → Model Weights: loads
- Model Training → Model Weights: saves
entrenar_ia.py - FastAPI Routes → Image and Record Store: saves predictions
app.py - User → Reference Submission: submits reference
- Reference Submission → Image and Record Store: saves image
app.py - Image and Record Store → JSON Records: updates
database/connection.py