[upd] READMEs, add new recorded video on bilibili

README.md

@@ -26,9 +26,10 @@ https://user-images.githubusercontent.com/39692511/226858919-dd4ccf7e-a5ba-4a33-
 ## 🔥 News
 
 - Welcome to our ICLR poster! (https://iclr.cc/virtual/2023/poster/12227).
+- On **May. 11th** another livestream on OpenMMLab & ReadPaper (bilibili)! [[`📹Recorded Video`](https://www.bilibili.com/video/BV11s4y1M7qL/)]
 - A brief introduction (in English) is available on [SlidesAlive](https://recorder-v3.slideslive.com/?share=81463&s=e4098919-55dc-431e-83dd-e1979e5c0faa) now!
-- On **Apr. 27th (UTC+8 8pm)** another livestream would be held at [OpenMMLab (Bilibili)](https://space.bilibili.com/1293512903)!
-- On **Mar. 22nd (UTC+8 8pm)** another livestream would be held at 极市平台 (Bilibili)! [[`📹Recorded Video`](https://www.bilibili.com/video/BV1Da4y1T7mr/)]
+- On **Apr. 27th (UTC+8 8pm)** another livestream would be held at [OpenMMLab (bilibili)](https://space.bilibili.com/1293512903)!
+- On **Mar. 22nd (UTC+8 8pm)** another livestream would be held at 极市平台 (bilibili)! [[`📹Recorded Video`](https://www.bilibili.com/video/BV1Da4y1T7mr/)]
 - The share on [TechBeat (将门创投)](https://www.techbeat.net/talk-info?id=758) is scheduled on **Mar. 16th (UTC+8 8pm)** too! [[`📹Recorded Video`](https://www.techbeat.net/talk-info?id=758)]
 - We are honored to be invited by Synced ("机器之心机动组 视频号" on WeChat) to give a talk about SparK on **Feb. 27th (UTC+0 11am, UTC+8 7pm)**, welcome! [[`📹Recorded Video`](https://www.bilibili.com/video/BV1J54y1u7U3/)]
 - This work got accepted to ICLR 2023 as a Spotlight (notable-top-25%).
@@ -157,7 +158,7 @@ import torch, timm
 res50 = timm.create_model('resnet50')
 res50.load_state_dict(torch.load('resnet50_1kpretrained_timm_style.pth', 'cpu'), strict=False)
 ```
-  
+
 - **Pretraining**
   - any ResNet or ConvNeXt on ImageNet-1k:  see [pretrain/](pretrain)
   - **your own CNN model**:  see [pretrain/](pretrain), especially [pretrain/models/custom.py](pretrain/models/custom.py)

downstream_imagenet/README.md

@@ -47,6 +47,7 @@ See files under `--exp_dir` to track your experiment:
     
     It also reports training loss/acc, best evaluation acc, and remaining time at each epoch.
 
+- `tensorboard_log/`: saves a lot of tensorboard logs, you can visualize accuracies, loss values, learning rates, gradient norms and more things via `tensorboard --logdir /path/to/this/tensorboard_log/ --port 23333`.
 
 ## Resuming
 

pretrain/README.md

@@ -37,7 +37,7 @@ python3 main.py --exp_name=debug --data_path=/path/to/imagenet --model=resnet50
 ## Pretraining Any Model on ImageNet-1k (224x224)
 
 For pretraining, run [/pretrain/main.py](/pretrain/main.py) with `torchrun`.
-**It is required to specify** the ImageNet data folder (`--data_path`), your experiment name & log dir (`--exp_name` and `--exp_dir`, automatically created if not exists), and the model name (`--model`, valid choices see the keys of 'pretrain_default_model_kwargs' in [/pretrain/models/__init__.py line34](/pretrain/models/__init__.py#L34)).
+**It is required to specify** the ImageNet data folder (`--data_path`), your experiment name & log dir (`--exp_name` and `--exp_dir`, automatically created if not exists), and the model name (`--model`, valid choices see the keys of 'pretrain_default_model_kwargs' in [/pretrain/models/\_\_init\_\_.py line34](/pretrain/models/__init__.py#L34)).
 
 We use the **same** pretraining configurations (lr, batch size, etc.) for all models (ResNets and ConvNeXts) in 224 pretraining.
 Their **names** and **default values** are in [/pretrain/utils/arg_util.py line23-44](/pretrain/utils/arg_util.py#L23-L44).
@@ -82,6 +82,7 @@ See files under `--exp_dir` to track your experiment:
     
     It also reports the loss and remaining pretraining time at each epoch.
 
+- `tensorboard_log/`: saves a lot of tensorboard logs, you can visualize loss values, learning rates, gradient norms and more things via `tensorboard --logdir /path/to/this/tensorboard_log/ --port 23333`.
 - `stdout_backup.txt` and `stderr_backup.txt`: will save all output to stdout/stderr