Edit model card

Trained for 4 epochs.

Usage:

image_processor = AutoImageProcessor.from_pretrained("microsoft/dit-large")
model = BeitForSemanticSegmentation.from_pretrained("jzju/dit-doclaynet")
image = Image.open('img.png').convert('RGB')
inputs = image_processor(images=image, return_tensors="pt")
outputs = model(**inputs)
# logits are of shape (batch_size, num_labels, height, width)
logits = outputs.logits
out = logits[0].detach()
out.size()
for i in range(11):
    plt.imshow(out[i])
    plt.show()

Labels:

1: Caption
2: Footnote
3: Formula
4: List-item
5: Page-footer
6: Page-header
7: Picture
8: Section-header
9: Table
10: Text
11: Title

Data label convert:

model = BeitForSemanticSegmentation.from_pretrained("microsoft/dit-base", num_labels=11)
ds = load_dataset("ds4sd/DocLayNet-v1.1")
mask = np.zeros([11, 1025, 1025])
for b, c in zip(d["bboxes"], d["category_id"]):
    b = [np.clip(int(bb), 0, 1025) for bb in b]
    mask[c - 1][b[1]:b[1]+b[3], b[0]:b[0]+b[2]] = 1
mask = [cv2.resize(a, dsize=(56, 56), interpolation=cv2.INTER_AREA) for a in mask]
d["label"] = np.stack(mask)
Downloads last month
180
Safetensors
Model size
163M params
Tensor type
F32
·
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Dataset used to train jzju/dit-doclaynet