-
Hi, we found two implementation issues in triton's `do_bench` when enabling max-autotune feature in Pytorch.
1. `do_bench` should return the number of `ms` , but currently is `ns`.
https://gi…
-
Just an idea. With all these great changes to the tuning aspects of iNav. What do folks think about autotune setting the TPA breakpoint, based on an average during the tune?
-
I am looking for an example of how to structure the `queries` and `qrels` parameters of the autotune function because I searched in the repo and didn't find any example for that. Precisely, what shoul…
-
We have had Sourcegraph.com and customer issues due to opening too many connections to postgresql. We should consider to either:
1. Autotune our db pool sizes. We can query max_connections.
2. Aut…
-
Previously we do this
```python
from torchao.quantization.quant_api import change_linear_weights_to_int8_woqtensors
model = torch.compile(model, mode="max-autotune", fullgraph=True)
change_lin…
-
I confess to knowing NOTHING about coding.
I have built the excellent looking filament jointer project. I am about to start the software side now that I think I have the hardware built correctly.
U…
-
Hi,
I've been playing aroud with the 39.7 Version and the Autotune seems to stop working when I put any of these parameters with a specific value:
-lhr-reduce-value
-lhr-reduce-time
-lhr-reduce-…
-
No comments necessary
-
Would be nice to tune my the PID of my heated elements.
Or is this product dead?
-
Hello!
I was checking the Autotune implementation and I'm trying to figure out the strategy used by fastText for the search.
Checking the [code we can find the search strategy for the Autotune f…