**Note: for network definitions, we directly use `timm.models.ResNet` and [official ConvNeXt](https://github.com/facebookresearch/ConvNeXt/blob/048efcea897d999aed302f2639b6270aedf8d4c8/models/convnext.py).**
For **all** other configurations/hyperparameters, their names and **default values** can be found in [utils/arg_util.py line24-47](https://github.com/keyu-tian/SparK/blob/main/pretrain/utils/arg_util.py#L24).
If you do not specify them like `--ep=800`, those default configurations would be used.
Note that the first argument `<experiment_name>` is the name of your experiment, which would be used to create an output directory named `output_<experiment_name>`.
-`<model>_still_pretraining.pth`: saves model and optimizer states, current epoch, current reconstruction loss, etc; can be used to resume pre-training
We do not use sparse convolutions in this pytorch implementation, due to their limited optimization on modern hardwares.
As can be found in [encoder.py](https://github.com/keyu-tian/SparK/blob/main/pretrain/encoder.py), we use masked dense convolution to simulate submanifold sparse convolution.
We also define some sparse pooling or normalization layers in [encoder.py](https://github.com/keyu-tian/SparK/blob/main/pretrain/encoder.py).
All these "sparse" layers are implemented through pytorch built-in operators.
## Some details: how we mask images and how to set the patch size
In SparK, the mask patch size **equals to** the downsample ratio of the CNN model (so there is no configuration like `--patch_size=32`).
Here is the reason: when we do mask, we:
1. first generate the binary mask for the **smallest** resolution feature map, i.e., generate the `_cur_active` or `active_b1ff` in [line86-87](https://github.com/keyu-tian/SparK/blob/main/pretrain/spark.py#L86), which is a `torch.BoolTensor` shaped as `[B, 1, fmap_size, fmap_size]`, and would be used to mask the smallest feature map.
3. then progressively upsample it (i.e., expand its 2nd and 3rd dimensions by calling `repeat_interleave(..., 2)` and `repeat_interleave(..., 3)` in [line16](https://github.com/keyu-tian/SparK/blob/main/pretrain/encoder.py#L16)), to mask those feature maps ([`x` in line21](https://github.com/keyu-tian/SparK/blob/main/pretrain/encoder.py#L21)) with larger resolutions .
So if you want a patch size of 16 or 8, you should actually define a new CNN model with a downsample ratio of 16 or 8.
Note that the `forward` function of this CNN should have an arg named `hierarchy`. You can look at https://github.com/keyu-tian/SparK/blob/main/pretrain/models/convnext.py#L78 to see what `hierarchy` means and how to handle it.
After that, you can simply run `main.sh` with `--hierarchy=3` and see if it works.