2023-03-30 07:26:27 -06:00
|
|
|
<div align="center">
|
|
|
|
|
|
|
|
# Text Generation Inference benchmarking tool
|
|
|
|
|
|
|
|
![benchmark](../assets/benchmark.png)
|
|
|
|
|
|
|
|
</div>
|
|
|
|
|
2024-02-16 03:58:58 -07:00
|
|
|
A lightweight benchmarking tool based inspired by [oha](https://github.com/hatoo/oha)
|
2023-03-30 07:26:27 -06:00
|
|
|
and powered by [tui](https://github.com/tui-rs-revival/ratatui).
|
|
|
|
|
2024-02-16 03:58:58 -07:00
|
|
|
## Install
|
2023-03-30 07:26:27 -06:00
|
|
|
|
2024-02-16 03:58:58 -07:00
|
|
|
```shell
|
2023-03-30 07:26:27 -06:00
|
|
|
make install-benchmark
|
|
|
|
```
|
|
|
|
|
|
|
|
## Run
|
|
|
|
|
|
|
|
First, start `text-generation-inference`:
|
|
|
|
|
|
|
|
```shell
|
|
|
|
text-generation-launcher --model-id bigscience/bloom-560m
|
|
|
|
```
|
|
|
|
|
|
|
|
Then run the benchmarking tool:
|
|
|
|
|
|
|
|
```shell
|
|
|
|
text-generation-benchmark --tokenizer-name bigscience/bloom-560m
|
2024-02-16 03:58:58 -07:00
|
|
|
```
|