-
### Checklist
- I have searched related issues but cannot get the expected help.
- I have read related documents and don't know what to do.
### Describe the question you meet
Hi, I'm trying …
-
article link: https://www.sciencedirect.com/science/article/pii/S0028393218300265?via%3Dihub
coordinate extracted by recent run of ACE: 10.0 48.0 38.0
article on neurosynth-compose: https://co…
-
min repro:
```
import torch
@torch.compile(fullgraph=True)
def f(x, y, z, w):
y = torch.ops.aten.addmm(x, y, z)
return y.view(-1).sin()
x = torch.randn(1308, requires_grad=True, dev…
-
**Is your feature request related to a problem? Please describe.**
Currently, an activation code can be used arbitrarily often to activate a device.
This enables owners of a card to "share" the same…
-
It seems there is a bug in the example 'capture_activations.py' script. There are two
```
parser=argparse.ArgumentParser()
```
statements making it only use the arguments from the last statement. …
-
**Is your feature request related to a problem? Please describe.**
As we are human, sometimes we forget to deactivate the license on machines before we wipe them, sell them, etc, Occupying license a…
-
(Ideally I don't think it should be auto-broadcasting in the first place). But if we just get rid of O-D array broadcasting that solves our problem over at https://github.com/EnzymeAD/Reactant.jl/issu…
-
Report is BOB view by corp week with counts of Passes, reservations made, and passes upgraded.
Filters:
- [ ] Corp Week - multi select
- [ ] Pass Type - Free, Paid, All
Include a second table that sh…
-
Unit test are failing for
- tests/test_activations.py
- tests/test_losses.py
- tests/test_model.py
on the latest main branch
-
https://github.com/pytorch/ao/tree/main/torchao/prototype/quantized_training
Currently INT8 training recipes only support **row-wise scaling** for weight. This should be strictly better than (or at…