feat: 添加多个数据集的支持,包括Gigaspeech、TextVQA、OCR-VQA-200K、RefCOCO系列,更新数据集工厂和处理逻辑,优化图像处理功能

This commit is contained in:
2025-05-15 20:33:29 +08:00
parent 9ca588224d
commit 24a6c3c114
17 changed files with 568 additions and 78 deletions
+5
View File
@@ -0,0 +1,5 @@
images/test_images/*
images/train_images/*
TextVQA_0.5.1_test.json
TextVQA_0.5.1_train.json
TextVQA_0.5.1_val.json