Yun Chen

Question on License

1

Hi @sunset1995. Great work and impressive results. I'm a PhD student at University of Toronto, and a researcher on a commercial company. I'm interested in trying your ideas for some...

Hi there, I felt that there are some recolor filters applied on the image, which makes the color pattern a bit weird. Is it possible if you provide more detail...

random_crop results in image with zero target.

2

[random_crop](https://github.com/kuangliu/torchcv/blob/90f923dc354b83bf4b61fa640d13add965eff575/examples/ssd/train.py#L56) op may crop only background part, which will result in the `cls_targets` to all be zero. This will raise `ZeroDivisionError: float division by zero` in [here](https://github.com/kuangliu/torchcv/blob/90f923dc354b83bf4b61fa640d13add965eff575/torchcv/loss/ssd_loss.py#L68) when all item...

Hi, Nice Work. But I think there is actually no need for CUDA code -- they could be implemented vectorized, which is more intuitive. For example: [`__global__ void InitDistanceMetricKernel`](https://github.com/yeezhu/SPN.pytorch/blob/f8ebdca128ce6b550d090b45dae7ef43d54d00e1/spnlib/spn/src/libspn_kernel.cu#L4) could...

## ❓ Questions and Help I'm confused by the `AdjustSmoothL1Loss` using `running_mean-running_var` in the paper. Running mean and running_var have different dimensional quantity. Subtracting has no meaning on them. e.g.,...

mtime of taskchampion.sqlite3 got changed when calling `task list`

2

When I run `task list` the ` taskchampion.sqlite3`'s mtime changed. This is inconvenience as I use the mtime to sync between workstations. So is there a way to turn off...

### Describe the solution you'd like It would be great if we could link to other memos by searching while typing. I found that memos already [support add reference by...

Question on gradient of depth

16

Thanks for the great work! This work is awesome. I note that the gradient is not backpropaged from the pred_depth, do you plan to add it? This would be super...

Test depth grad on xyz and alpha. Note: I found the initialization of alpha is important: 1. To optimize the alpha: if there are too many points, the initial alpha...