-
This seems like a very interesting approach that seems very beneficial in theory. Are there any performance benchmarks were it can be directly compared to other libraries. Maybe something like the [JS…
-
Define and implement a few benchmark models that can be used to measure (improvements in) model assembly time.
Suggested criteria:
1. There should only be a few (say, 3-4) benchmarks.
2. The total s…
-
Latest `transformers=4.44.2` packages from HuggingFace uses `torch.isin()` which prevents graph break in certain control flows. It also adds the condition branch to avoid `copy.deepcopy` which is not …
-
I found a big problem with whitespace characters, which increase the size of the final server response by 25% (197kb vs 149kb)
With this change, I increased the average number of requests from 968r…
-
### Run Information
Name | Value
-- | --
Architecture | arm64
OS | ubuntu 22.04
Queue | AmpereUbuntu
Baseline | [e733c2f22710b6382c9aadb8139833fbf395c332](https://github.com/dotnet/runtime/commit/…
-
Some notes for later as I go along with writing type hints, on things that I can't tackle because they require a newer Python version or a larger refactoring than I want to do right now (I want some t…
-
Replace next/react/etc with a simpler stack. This will be a try-out for [comma connect's](https://github.com/commaai/connect) next framework.
We care about simplicity, brevity, and performance. So…
-
to fix https://github.com/eknkc/ssr-benchmark/issues/27 and to prevent any noisy neighbors and global monkey patching each framework should run in a worker or at least isolated in some way
-
### Run Information
Name | Value
-- | --
Architecture | x64
OS | ubuntu 22.04
Queue | TigerUbuntu
Baseline | [5ba3c05f3bf0ea789421c6e69d3983ecdc086f36](https://github.com/dotnet/runtime/commit/5ba…
-
I followed the documents to perform resnet50 benchmark.
```bash
cm run script --tags=run-mlperf,inference,_find-performance,_full,_r4.1 --model=resnet50 --implementation=nvidia --framework=tensorrt …