-
I took a benchmark of `torch.compile`'d `Adam.step` with Thunder backend. Surprisingly, the compiled `Adam.step` was even slower than the eager mode.
||compilation (s)|execution (ms)|
|:-|-:|-:|
…
-
Simply slapping parens on makes it warn but without them, nothing:
```
m: eager (1..10);
rakudo-moar 72948e847: OUTPUT: «WARNINGS for :Useless use of ".." in expression "1..10" in sink context…
-
## Issue Description
The current implementation of the UserGroup does not utilize Eloquent. This limitation makes it hard to manage relationships, run complex queries efficiently and perform eager lo…
-
This code should print `(1, 2)` forever, and on Julia 1.0, it does. However, on 1.6 and 1.12 it throws.
```julia
accumulate([1,2,3]) do args...
while true
println(args)
end
end
…
-
### Requirements
- [x] I've searched [platform-specific issues](https://github.com/jtroo/kanata/blob/main/docs/platform-known-issues.adoc), [issues](https://github.com/jtroo/kanata/issues) and [dis…
-
# linear search algorithm in java.
@ghostmkg "I would love the opportunity to contribute and work on this project.
Could you please assign the issue to me? I'm eager to dive in and make a meani…
-
### System Info
- `transformers` version: 4.44.2
- Platform: Linux-6.5.0-35-generic-x86_64-with-glibc2.35
- Python version: 3.10.14
- Huggingface_hub version: 0.24.6
- Safetensors version: 0.4.5
…
lapp0 updated
3 weeks ago
-
### Description
Indices can be set up to load specific files in an mmap directory ahead of time. This is particularly handy for fast searches that need all the data to be readily available in the pag…
-
### 🐛 Describe the bug
In deepspeed workloads, Zero parameter offload is implemented via module hooks. We find under torch.compile scenrio, if there any graph breaks happen in the pre/post hook of a …
-
When using eager loading, all relations are loaded. It would be useful, especially with a large number of relations, to be able to filter the returned relations. Perhaps something like this:
```s…