---
tags:
- chest-xray
- radiology
- contrastive-learning
- mimic-cxr
- vision-encoder
license: apache-2.0
---

# LAPVQA — Pretrain (Sigmoid)

Part of the [LAPVQA collection](https://huggingface.co/collections/dmusingu/lapvqa).

## Description

A **ViT-L/14** vision encoder trained from scratch on [MIMIC-CXR](https://physionet.org/content/mimic-cxr)
using a **sigmoid (multi-label binary cross-entropy) contrastive loss** — an alternative to InfoNCE that
treats each image-text pair independently rather than competing within the batch.

## Architecture

| Component | Detail |
|---|---|
| Vision backbone | ViT-L/14, 24-layer, 1024-dim, 16-head, patch 14, 384 px |
| Text encoder | 6-layer, 512-dim bidirectional transformer, GPT-2 vocab (50 257) |
| Projection | Linear → 512-dim shared embedding space |
| Loss | Per-pair sigmoid BCE (SigLIP-style) |
| Training data | MIMIC-CXR (physionet.org/content/mimic-cxr) |
| Epochs | 50 |

## Downstream Evaluation (frozen encoder + linear probe)

| Dataset | Mean AUC |
|---|---|
| NIH CXR-14 (14-class) | 0.650 |
| CheXpert-5 (5-class) | 0.785 |

## Files

| File | Description |
|---|---|
| `encoder_final.pt` | Vision encoder weights at end of training |
| `model_best.pt` | Full model at best validation loss |
| `model_epochXXX.pt` | Periodic epoch snapshots (every 10 epochs) |

## Usage

```python
import torch
from lapvqa.pretrain.model import ContrastiveModel

ckpt = torch.load("encoder_final.pt", map_location="cpu")
model = ContrastiveModel()
model.vision_encoder.load_state_dict(ckpt)
model.eval()
```