-
Is TrOCR good choice for handwritten text Recognition for images with large W/H ratio of 5-6. E.g. 600*100 size images ?
TrOCR resizes input image to square image of 384*384 which distorts the imag…
-
File "project_dir/code/repos/unilm/layoutlmft/layoutlmft/models/layoutlmv2/modeling_layoutlmv2.py", line 668, in _calc_img_embeddings
embeddings = visual_embeddings + position_embeddings + spatia…
-
**Describe the bug**
Model I am using (UniLM, MiniLM, LayoutLM ...): BEIT 2
The problem arises when using:
* [-] the official example scripts: (give details below)
* [ ] my own modified scripts:…
-
**Describe the bug**
Model I am using (UniLM, MiniLM, LayoutLM ...): EdgeLM
The problem arises when using:
* [X] README link to checkpoint model
A clear and concise description of what the bu…
-
**Describe the bug**
Trying to install Kosmos-G, there's a disclaimer that says:
> Make sure you have Pytorch 1.13.0 installed.
However on
```shell
bash vl_setup.sh
```
We got
```she…
-
#! /bin/bash
python extract_convert.py
python extract_vectorize.py
for ((i=0; i
-
Model I am using is LayoutLMv2:
(Link of the demo for reference: https://huggingface.co/spaces/nielsr/LayoutLMv2-FUNSD )
I do get 'questions' & 'answers' as separate colored boxes in output imag…
-
**Describe the bug**
LayoutXLM
The problem arises when using:
* [ ] my own modified scripts: (give details below)
A clear and concise description of what the bug is.
**Expected behavior…
-
Hi, for anyone who interested in the implementation of LayoutLMv3. Transfomers have updated the code for mask image modeling and the code is based on DEIT. You can inherit the code to implement the Ma…
-
I am using DiT for text detection and am having difficulty finding a way to perform inference on my documents. Has anyone successfully created an inference script for this model in this task?
Valent…