Tadej Svetina
Tadej Svetina
### My actions before raising this issue - [x] Read/searched [the docs](https://github.com/cvat-ai/cvat/tree/master#documentation) - [x] Searched [past issues](/issues) ### Expected Behaviour Upload MOT annotations normally. ### Current Behaviour I have MOT...
I have a dataset that has about 55 shards, each is ~500MB `.tgz` file, they are cached on disk (originals stored on S3). The sample contains a `npy` array and...
## 🚀 Feature Request I propose to separate concerns between Accelerate and Catalyst: Catalyst should not try to spawn/launch anything - this should all be taken care of by `accelerate...
Currently, only a base64 encoded string can be passed to `-cloud-startup-script`. It would be more convenient if a path to a file (script) could be specified
Fix the dimension size of the operator, and describe that it is filled with -1.
Right now, `cml runner` command can only launch `amd64` based EC2 instances, as there is only an AMI for this architecture available. Would be great if there was also support...
### Version 1.36.0 ### Describe the bug. A `TensorGPU` can be created from any object conforming to a CUDA Array Interface. Version 3 of this interface (accodring to [numba docs](https://numba.readthedocs.io/en/stable/cuda/cuda_array_interface.html#cuda-array-interface-version-3))...
I would suggest that the CVAT container automatically creates a superuser on startup, with username/password/email set by the `DJANGO_SUPERUSER_*` [env variables](https://docs.djangoproject.com/en/3.2/ref/django-admin/#createsuperuser). Right now, users have to manually create the superuser...
You may want to build the binary with `musl` instead of `glibc`, because `glibc` does not allow you to build a truly static binary - which may cause conflicts if...
Would you be ok if the dataset was uploaded to huggingface hub? That would make it much easier to use. I can do that myself, just asking for permission because...