-
### 🚀 The feature, motivation and pitch
I have tried a few times to add Unified Memory support to Pytorch, so as to leverage as many resources of my computer as possible while running training and in…
-
Super sorry if this is obvious, but -- how do I copy a tensor from CPU -> GPU and vice versa? I've been looking through the documentation, and can't seem to find how to do this?
sflc6 updated
6 years ago
-
### 🐛 Describe the bug
when backward tensor of sng operator is converted from torch tensor to numpy then output is not correct.
it was expected that output should be 0's as torch tensor has filled …
-
This would enlarge the accepted set of inputs of `torch.as_tensor` and would support PIL images / h5py arrays. I think this feature request goes well in the theme of standardizing support for methods …
-
Please:
- [x] Check for duplicate requests.
- [x] Describe your goal, and if possible provide a code snippet with a motivating example.
TVM can generate highly optimized operators. Is it possib…
-
Hello,
I'm trying to implement the pytorch version of the neural receiver model. I have converted the tensorflow tensors into pytorch tensors (edno data). The project details cannot be discussed. But…
-
## 🐛 Bug
## To Reproduce
Steps to reproduce the behavior:
1. mlc_llm convert_weight ./dist/models/Qwen2-1.5B-Instruct/ --model-type qwen2 --quantization q0f16 -o dist/Qwen2-1.5B-Instruct-…
-
### Description
For D2H (gpu to cpu) transfer, `jax.device_put` has very low throughput. `device_put` yields ~2.7GB/s transfer speed; in contrast, a very simple CUDA program yields ~25GB/s. Is ther…
-
The eager mode is described in the doc as " mostly used to debug and check intermediate results are as expected",
however it seems it has a much greater potential than just this: with support for GP…
-
Hello,
Did you test it with torch::tensor? I mean were you able to read a wav file into a C++ torch::tensor?
ghost updated
3 years ago