提交 a5dbd8bb 编写于 作者: 文幕地方's avatar 文幕地方

rm unsused code

上级 5e7ad64c
...@@ -182,16 +182,6 @@ python3 eval_ser.py \ ...@@ -182,16 +182,6 @@ python3 eval_ser.py \
``` ```
最终会打印出`precision`, `recall`, `f1`等指标 最终会打印出`precision`, `recall`, `f1`等指标
```shell
export CUDA_VISIBLE_DEVICES=0
python3.7 infer_ser.py \
--model_name_or_path "./PP-Layout_v1.0_ser_pretrained/" \
--output_dir "output_res/" \
--infer_imgs "XFUND/zh_val/image/" \
--ocr_json_path "XFUND/zh_val/xfun_normalize_val.json"
```
* 使用评估集合中提供的OCR识别结果进行预测 * 使用评估集合中提供的OCR识别结果进行预测
```shell ```shell
...@@ -237,7 +227,7 @@ python3 train_re.py \ ...@@ -237,7 +227,7 @@ python3 train_re.py \
--eval_data_dir "XFUND/zh_val/image" \ --eval_data_dir "XFUND/zh_val/image" \
--eval_label_path "XFUND/zh_val/xfun_normalize_val.json" \ --eval_label_path "XFUND/zh_val/xfun_normalize_val.json" \
--label_map_path 'labels/labels_ser.txt' \ --label_map_path 'labels/labels_ser.txt' \
--num_train_epochs 2 \ --num_train_epochs 200 \
--eval_steps 10 \ --eval_steps 10 \
--output_dir "output/re/" \ --output_dir "output/re/" \
--learning_rate 5e-5 \ --learning_rate 5e-5 \
......
...@@ -139,7 +139,7 @@ def train(args): ...@@ -139,7 +139,7 @@ def train(args):
global_step = 0 global_step = 0
tr_loss = 0.0 tr_loss = 0.0
set_seed(ags.seed) set_seed(args.seed)
best_metrics = None best_metrics = None
train_reader_cost = 0.0 train_reader_cost = 0.0
......
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