未验证 提交 113c9bb5 编写于 作者: 文幕地方's avatar 文幕地方 提交者: GitHub

Update README.md

上级 11d6814f
......@@ -136,7 +136,7 @@ wget https://paddleocr.bj.bcebos.com/dataset/XFUND.tar
```shell
python3.7 train_ser.py \
--model_name_or_path "layoutxlm-base-uncased" \
--ser_model_type "LayoutLM" \
--ser_model_type "LayoutXLM" \
--train_data_dir "XFUND/zh_train/image" \
--train_label_path "XFUND/zh_train/xfun_normalize_train.json" \
--eval_data_dir "XFUND/zh_val/image" \
......@@ -316,6 +316,7 @@ export CUDA_VISIBLE_DEVICES=0
python3.7 infer_ser_re_e2e.py \
--model_name_or_path "./PP-Layout_v1.0_ser_pretrained/" \
--re_model_name_or_path "./PP-Layout_v1.0_re_pretrained/" \
--ser_model_type "LayoutXLM" \
--max_seq_length 512 \
--output_dir "output/ser_re_e2e/" \
--infer_imgs "images/input/zh_val_21.jpg"
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册