ModelPublicadoExterno

Lapvqa Diffvqa Native

Part of the LAPVQA collection.

Fuentehuggingface.co/dmusingu/lapvqa-diffvqa-native/tree/217640ae10b2e63be150c449ee3ffa089022e739

TipoOtherVersiónClip vit l14 bestCcalidad

Fuente del modelo

Tipo: Other
Tamaño del archivo: 196–197 MB
Versión: vClip vit l14 best
Actualizado: 10 oct 2026
Fuente: Hugging Face

Descripción de la fuente

LAPVQA — Differential VQA (Native / End-to-end)

Part of the LAPVQA collection.

Description

DiffVQA models trained end-to-end (encoder + head jointly). Each .pt file is a plain state dict of DiffVQAHead. MAE-ViT-L/16 is the primary encoder studied.

Results (test set, MAE-ViT-L/16)

BLEU-4	ROUGE-2	RadGraph-s	BERTScore F1
0.472	0.573	0.288	0.938

File	Encoder	vis_dim
`clip-vit-l14_best.pt`	CLIP ViT-L/14	1024
`coca_best.pt`	CoCa	768
`florence2_best.pt`	Florence-2	1024
`mae-vit-l16_best.pt`	MAE ViT-L/16	1024
`siglip_best.pt`	SigLIP	1152

Loading

import torch
from lapvqa.diffvqa.model import DiffVQAHead

ckpt = torch.load("mae-vit-l16_best.pt", map_location="cpu")
head = DiffVQAHead(vis_dim=1024)
head.load_state_dict(ckpt)
head.eval()

Fuentes

1 fuente

Hugging FacePrincipalVersión 217640ae10b2e63be150c449ee3ffa089022e739

huggingface.co/dmusingu/lapvqa-diffvqa-native/tree/217640ae10b2e63be150c449ee3ffa089022e739

Verificado 10 oct

Artefactos del modelo

5 artefactos

clip-vit-l14_best.pt

pt · 197 MB · SHA-256 c56a34e5e4c2…04fd · Hugging Face

Descargar

coca_best.pt

pt · 196 MB · SHA-256 276f4534f69e…7122 · Hugging Face

Descargar

florence2_best.pt

pt · 197 MB · SHA-256 2f996f577226…f084 · Hugging Face

Descargar

mae-vit-l16_best.pt

pt · 197 MB · SHA-256 285981ac78b6…1b95 · Hugging Face

Descargar

siglip_best.pt

pt · 197 MB · SHA-256 589b0c01ee99…9095 · Hugging Face

Descargar

Extractos de fuentes

2 extractos

Readme1 agohuggingface.co/dmusingu/lapvqa-diffvqa-native

--- tags: - chest-xray - radiology - visual-question-answering - differential-vqa - mimic-cxr license: apache-2.0 --- # LAPVQA — Differential VQA (Native / End-to-end) Part of the [LAPVQA collection](https://huggingface.co/collections/dmusingu/lapvqa). ## Description DiffVQA models trained **end-to-end** (encoder + head jointly). Each `.pt` file is a plain state dict of `DiffVQAHead`. MAE-ViT-L/16 is the primary encoder studied. ## Results (test set, MAE-ViT-L/16) | BLEU-4 | ROUGE-2 | RadGraph-s | BERTScore F1 | |---|---|---|---| | 0.472 | 0.573 | 0.288 | 0.938 | | File | Encoder | vis_dim | |---|---|---| | `clip-vit-l14_best.pt` | CLIP ViT-L/14 | 1024 | | `coca_best.pt` | CoCa | 768 | | `florence2_best.pt` | Florence-2 | 1024 | | `mae-vit-l16_best.pt` | MAE ViT-L/16 | 1024 | | `siglip_best.pt` | SigLIP | 1152 | ## Loading ```python import torch from lapvqa.diffvqa.model import DiffVQAHead ckpt = torch.load("mae-vit-l16_best.pt", map_location="cpu") head = DiffVQAHead(vis_dim=1024) head.load_state_dict(ckpt) head.eval() ```

Note1 agohuggingface.co/dmusingu/lapvqa-diffvqa-native

Source context: 0 downloads · 0 likes · Pipeline visual-question-answering · Repo dmusingu/lapvqa-diffvqa-native