-
**Context**
Hey, I'm setting up a python backend and I am using `dlpack` to keep the tensors on GPU. As described in [its](https://github.com/triton-inference-server/python_backend?tab=readme-ov-file…
Thytu updated
10 months ago
-
After adding Openblas, OpenCV, leptonica-devel and tesseract-devel I have tried to compile mxnet from source. On both my machines this process fails.
## Environment info
Operating System:
Fedora …
-
right now, everything is either an input or output. Reason being TC does not do any allocation by itself. We should find a better way to handle this
cc Albert Cohen who is interested in this
-
I would support JAX as another backend.
Would probably be worth opening another issue for tracking JAX as a backend. For IO, I was thinking that could be done with #659
_Originally posted by @iv…
-
This issue will be used to track the work for zero-copy between CUDA and XLA.
Inspired by
- https://github.com/pytorch/pytorch/blob/f20e3ae0c36146c962a5665018e9ad662a7cf211/aten/src/ATen/DLConver…
-
I am decoding a list of videos with:
```
video = VideoReader(str(video_path), ctx=gpu(0))
frame_ids = list(range(300))
frames = video.get_batch(frame_ids).asnumpy()
```
on every iteratio…
-
### Describe the enhancement requested
[JAX](https://github.com/jax-ml/jax) is a popular numeric computing library. People googling for how to convert pyarrow arrays to JAX arrays might not infer by …
-
Hello,
I'm trying to build examples and tests, using AVX2 and OpenCL:
```
cmake .. -DBUILD_EXAMPLES=ON -DBUILD_TESTS=ON -DUSE_AVX2=ON -DUSE_OPENCL=ON
```
However compilation fails with:
…
-
## 🚀 Feature
Having `torch.from_numba(cuda_arr)` can be useful in many cases.
## Motivation
With the easy to use APIs in numba, I would like to do the preprocessing on gpu and then pass resulta…
-
## 🐛 Bug
When I deploy my own 2B model using MLC on Android, the model interface initializes successfully and displays the "Ready to chat" prompt after opening. However, the app crashes after sendi…