# CSAILVision/places365

The Places365-CNNs for Scene Classification

Repository: https://github.com/CSAILVision/places365
Canonical: https://ross.abutalabs.com/products/places365
Homepage: http://places2.csail.mit.edu/
Language: Python
License: MIT
License Family: permissive
Topics: cnn, baseline-cnns
Last push: 2025-10-17T23:52:35+00:00

## Health v2 (maintenance only)
Score: 53/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 47, release rhythm 35, longevity 100
- inputs: {"age_days": 3765, "days_push": 320, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2082, forks 538 (observed 2026-08-28T04:06:11.749664+00:00)

## What it is
Places365-CNNs is a collection of pretrained convolutional neural network models (AlexNet, GoogLeNet, VGG16, ResNet50/152) trained on the Places365 scene classification dataset of ~1.8 million images across 365 scene categories. It accompanies the larger Places2 database of 10+ million scene images from MIT CSAIL for scene recognition research.

## Use cases
- classify images by scene category like bedroom, street, or kitchen
- extract deep scene features from photos for transfer learning
- benchmark scene recognition models against Places365 baselines
- generate scene attributes and class activation maps for images
- train models for indoor vs outdoor environment recognition
- use pretrained scene CNNs as a backbone for downstream vision tasks

## When to choose
- you need scene/environment classification rather than object recognition
- you want pretrained CNN baselines for scene recognition research
- you need a large-scale scene dataset for training deep visual features
- you are reproducing Places2 Challenge or scene-centric benchmark results

## When to avoid
- you need general object detection or fine-grained object classification
- you want a lightweight modern architecture rather than legacy Caffe/Torch models
- your project requires actively maintained code with recent framework support
- you need non-commercial licensing since the dataset is for academic research and education

## Facets
- artifact type: dataset
- maturity: stable
- function: machine-learning, image-processing, computer-vision
- domain: computer-vision, deep-learning, image-processing, machine-learning
- platform: python, cross-platform
- tags: scene-classification, pretrained-models, cnn, places365, scene-recognition, caffe, pytorch

## Member repositories
- CSAILVision/places365 (main) score 53

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:11.749664+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:55:54.628405+00:00, confidence not recorded.
  - readme: https://github.com/CSAILVision/places365 (fetched 2026-08-28T04:06:11.749664+00:00, sha c27ba9db7950)
  - homepage: http://places2.csail.mit.edu/ (fetched 2026-08-29T10:35:57.094766+00:00, sha 6f81d2820cfc)
- Data as of 2026-08-30T08:39:29.467469+00:00.