PaddleOCR/main/applications/%E6%89%8B%E5%86%99%E6%96%87%E5%AD%97%E8%AF%86%E5%88%AB

[giscus[bot]]

PaddleOCR/main/applications/%E6%89%8B%E5%86%99%E6%96%87%E5%AD%97%E8%AF%86%E5%88%AB

Awesome multilingual OCR toolkits based on PaddlePaddle (practical ultra lightweight OCR system, support 80+ languages recognition, provide data annotation and synthesis tools, support training and deployment among server, mobile, embedded and IoT devices)

https://paddlepaddle.github.io/PaddleOCR/main/applications/%E6%89%8B%E5%86%99%E6%96%87%E5%AD%97%E8%AF%86%E5%88%AB.html

[hecheng64]

我们使用configs/rec/PP-OCRv3/ch_PP-OCRv3_rec_distillation.yml 这个是蒸馏，但是修改地方都没体现到，百度文档就是写不够仔细，前后矛盾
官网上面：
Architecture:
model_type: &model_type "rec"
name: DistillationModel
algorithm: Distillation
Models:
Teacher:
pretrained:
freeze_params: false ----这里到底是是true还是false， 如果是true，根本无法收敛
return_all_feats: true
model_type: *model_type
algorithm: SVTR
Transform:
Backbone:

[hecheng64]

怎么联系，issue提几个类似问题，回答也是前后矛盾

[SWHL]

识别模型需要很大的数据量才能有效收敛的。你用的多少数据集呀