-
**Is your feature request related to a problem? Please describe.**
We should allow Feature Views to return matrices/tensors natively. For example, `torch.tensors`.
At the moment, for some feature…
-
### 🚀 The feature, motivation and pitch
I want to implement gradient checkpointing API where I want to specify some tensors which are not stored for backpropagation. The current `checkpoint` API does…
-
Hi, I inspected the collected depth values which are used to train the ray prediction network. However, the depth images tensors are all populated with the max depth values, when the ray ground truth …
-
how can i use the multiple GPU reasoning the cli? and when i used the single gpu i always got the error :RuntimeError: Expected all tensors to be on the same device, but found at least two devices, c…
-
### 🐛 Describe the bug
```
import torch
from triton.testing import do_bench
from torch.nn.attention.flex_attention import create_block_mask, flex_attention, noop_mask, BlockMask
import torch.nn.f…
-
### 🚀 The feature, motivation and pitch
I'm working on exporting an object detection model which casts a list of tensor shapes into a tensor, and I'm trying to hint the compiler that it should trea…
dgcnz updated
4 weeks ago
-
It would be really nice if System.Numerics.Tensors.TensorPrimitives had partial derivative functions (gradient) for SoftMax and Sigmoid as they are commonly needed in machine learing.
-
Using openai official text model, **text embedding** dim is **768**, mismatching with **llm2clip img embedding** dim **1280**.
```
text_model = CLIPModel.from_pretrained("openai/clip-vit-large-patch1…
-
## Context
Suppose I have the following code (essentially [this test file](https://github.com/google/heir/blob/main/tests/Dialect/LinAlg/Conversions/linalg_to_tensor_ext/float_vector_square_matrix_mat…
-
### 🐛 Describe the bug
Hi,
Not sure if I'm using nested tensors incorrectly, but I would like to pad some variable-length sequences and feed the resulting padded tensor into a DataLoader.
```py
…