AdvancedLiterateMachinery

AdvancedLiterateMachinery copied to clipboard

Omniparser inference

4

Hi, Do you provide pretrained models for inference in Omniparser? I couldn't find the inference and pretrained models.

aneeqahsan01

Hi, thanks for the great work on layoutllm. I only see zero-shot evaluation on the published paper, is there also a finetuned compression between layoutllm and other methods that is...

jianglong-he-Infrrd

bash scripts/infer/demo_wired.sh 推理有线表格时，报了个关于加载模型权重的错误（但无线表格可以成功推理），请问该怎么解决呀？ Traceback (most recent call last): File "demo.py", line 82, in demo(opt) File "demo.py", line 21, in demo detector = Detector(opt) File "/data/xyj/space/AdvancedLiterateMachinery/DocumentUnderstanding/LORE-TSR/src/lib/detectors/ctdet.py", line 32, in __init__ super(CtdetDetector,...

Selvaggiar

converts image considering format provided in pdf2img

1

Solves issue #150 . The pdf2img script previously accepted both jpg and png as output format arguments, but always converted to png regardless of the argument provided. This PR fixes...

Harsh19012003

command line argument takes image format input as jpg and png but always converts into png

Harsh19012003

When code for the TSR will be released?

2

When will code for TR task be released？ Do you have a plan? #132

ygzhu

Hi, is there a way to fine-tune MGP-STR with custom data? Also, for demo.py in MGP-STR, does fusion model exist when joining char_pred, bpe_pred, and wp_pred? Thank you so much.

Youngkwan-Cho

Inference Geolayoutlm ??

6

how do i inference Geolayoutlm for kIE task ?

zokkomon

如题： Application/DocXChain中的TableParsing，有现成工具，将其识别结果转为latex格式吗？

SidneyRey

关于resnet-18骨干以及预训练模型加载问题

2

1.我在尝试使用自己数据集并加载wireless预训练模型时发现使用respfn-18骨干训练会存在网络不一致情况，加载wtw和ptn预训练模型并使用dla-34骨干训练就不存在问题； 2.demo_wireless里使用的arch是resfpnhalf,查看了代码里面加载骨干的代码被注释了，所以是不是可以理解为加载作者提供的wireless预训练模型和resnet骨干网络是冲突的呢？ 3.作者提供的三个预训练模型中，只有wireless是字面意思上的无线表格，但是在论文里好像并没有提到无框线表格的相关信息，还是说这三个预训练模型其实可以用于无框线表格预测呢？

Jiangyangya