未验证 提交 9ffbacc4 编写于 作者: D Double_V 提交者: GitHub

Merge pull request #6616 from Rm1n90/release/2.5

Fix TensorRT failed due to some TRT inputs dynamic shape info not set
......@@ -276,6 +276,7 @@ def create_predictor(args, mode, logger):
min_input_shape = {"x": [1, 3, imgH, 10]}
max_input_shape = {"x": [args.rec_batch_num, 3, imgH, 2304]}
opt_input_shape = {"x": [args.rec_batch_num, 3, imgH, 320]}
config.exp_disable_tensorrt_ops(["transpose2"])
elif mode == "cls":
min_input_shape = {"x": [1, 3, 48, 10]}
max_input_shape = {"x": [args.rec_batch_num, 3, 48, 1024]}
......
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