huangjun12

download link 404

2

这三个好像都失效了

[NOT Merge] add new model example pr

1

This pr just shows how to add new model into paddlevideo. more details can be refer to [add_new_algorithm doc](https://github.com/PaddlePaddle/PaddleVideo/blob/develop/docs/zh-CN/contribute/add_new_algorithm.md) Try to run, just ```bash git clone https://github.com/huangjun12/PaddleVideo.git git checkout contribute_pr...

loss_kd_neg计算的意义

3

https://github.com/HikariTJU/LD/blob/2dda5c043e733c7cc40f6ce41fb37d0e76c44eeb/mmdet/models/dense_heads/ld_head.py#L267 请教下 这一项loss恒为0，为什么还要计算呢

耦合头效果怎么样

6

请问下 LD蒸馏有在耦合头上实验过效果吗？如TAL

Add TSM compile args

1

TSM脚本添加compile模式接口 更新子模块

add fleet for TSM

3

add fleet for TSM

update bmn to 2.0rc

total text评估精度比论文稍低

6

使用提供的预训练模型，按README文档进行评估，精度比paper稍低一些，请问下符合预期吗？ |metric | P | R | F | | :------ | :----------: | :----: | :----: | | code | 88.07 | 80.48 | 84.10 | | paper |...

Errors when training with dataloader_num_workers > 0

2

error message ``` RuntimeError: Cannot re-initialize CUDA in forked subprocess. To use CUDA with multiprocessing, you must use the 'spawn' start method ```