-
Hi,
The 2.2 release of Torch added a great integration to speed up transformer attention based architectures.
> FlashAttentionV2 backend for scaled dot product attention
An example of how it…
-
One of the things @gchanan keeps telling me we should do (and we still haven't done, all these years later), is get rid of torch.set_default_tensor_type. Why is it bad? Let me count the ways:
* It …
-
# Issue Metrics
| Metric | Average | Median | 90th percentile |
| --- | --- | --- | ---: |
| Time to first response | 1 day, 21:30:09 | 14:21:20 | 5 days, 3:51:05 |
| Time to close | 7 days, 1:43:44 …
-
## Feature
## Is your feature request related to a problem? Please describe.
In Michelangelo, Neuropod JNI and backends are built into different places: Neuropod JNI is built into a michelangelo…
-
Hi there. I am upgrading my bindings for the lord of llms tool and I now need to be able to vectorize text to embedding space of the current model. Is there a way to have access to the latent space of…
-
As of now, users do not have the freedom to pass hyperparameters during the run.
-
I'm creating this issue because I couldn't figure it out on my own.
1. **Can rust-bert support WASM**?
The issue I had trying to refactor this crate on my own is that the idea of resources reli…
-
# Request For Comment: The future of Kaldi-compatible features
## Problems
`torchaudio.compliance.kaldi` implements functionalities that tries to reproduce Kaldi's feature extractions, and this mo…
-
as soon we start to train a model after processing an images with comlap.
below isssues are coming
```
ERROR:root:Path not found: /configs/360.gin
Traceback (most recent call last):
File "…
-
## 🚀 Feature
Incorporate support for Posit number format as another dtype of PyTorch
## Motivation
The Posit number format (introduced in [Beating Floating Point at its Own Game](http://www.johng…