# tritony - Tiny configuration for Triton Inference Server
[](https://coveralls.io/github/rtzr/tritony?branch=main)
## What is this?
If you see [the official example](https://github.com/triton-inference-server/client/tree/main/src/python/examples), it is really confusing to use where to start.
Use tritony! You will get really short lines of code like example below.
```python
import argparse
import os
from glob import glob
import numpy as np
from PIL import Image
from tritony import InferenceClient
def preprocess(img, dtype=np.float32, h=224, w=224, scaling="INCEPTION"):
sample_img = img.convert("RGB")
resized_img = sample_img.resize((w, h), Image.Resampling.BILINEAR)
resized = np.array(resized_img)
if resized.ndim == 2:
resized = resized[:, :, np.newaxis]
scaled = (resized / 127.5) - 1
ordered = np.transpose(scaled, (2, 0, 1))
return ordered.astype(dtype)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--image_folder", type=str, help="Input folder.")
FLAGS = parser.parse_args()
client = InferenceClient.create_with("densenet_onnx", "0.0.0.0:8001", input_dims=3, protocol="grpc")
client.output_kwargs = {"class_count": 1}
image_data = []
for filename in glob(os.path.join(FLAGS.image_folder, "*")):
image_data.append(preprocess(Image.open(filename)))
result = client(np.asarray(image_data))
for output in result:
max_value, arg_max, class_name = output[0].decode("utf-8").split(":")
print(f"{max_value} ({arg_max}) = {class_name}")
```
## Release Notes
- 24.07.11 Upgrade minimum tritonclient version to 2.34.0
- 23.08.30 Support `optional` with model input, `parameters` on config.pbtxt
- 23.06.16 Support tritonclient>=2.34.0
- Loosely modified the requirements related to tritonclient
## Key Features
- [x] Simple configuration. Only `$host:$port` and `$model_name` are required.
- [x] Generating asynchronous requests with `asyncio.Queue`
- [x] Simple Model switching
- [ ] Support async tritonclient
## Requirements
$ pip install tritonclient[all]
## Install
$ pip install tritony
## Test
### With Triton
```bash
./bin/run_triton_tritony_sample.sh
```
```bash
pytest -s --cov-report term-missing --cov=tritony tests/
```
### Example with image_client.py
- Follow steps
in [the official triton server documentation](https://github.com/triton-inference-server/server#serve-a-model-in-3-easy-steps)
```bash
# Download Images from https://github.com/triton-inference-server/server.git
python ./example/image_client.py --image_folder "./server/qa/images"
```
Raw data
{
"_id": null,
"home_page": "https://github.com/rtzr/tritony",
"name": "tritony",
"maintainer": null,
"docs_url": null,
"requires_python": null,
"maintainer_email": null,
"keywords": "grpc, http, triton, tensorrt, inference, server, service, client, nvidia, rtzr",
"author": "Arthur Kim, RTZR team",
"author_email": "arthur@rtzr.ai",
"download_url": "https://files.pythonhosted.org/packages/73/e2/e87a66d934e9f7a48c4bf5967045eb1e5042e52504d589ac8ac3d0481fc6/tritony-0.0.19.tar.gz",
"platform": null,
"description": "# tritony - Tiny configuration for Triton Inference Server\n\n\n\n[](https://coveralls.io/github/rtzr/tritony?branch=main)\n\n## What is this?\n\nIf you see [the official example](https://github.com/triton-inference-server/client/tree/main/src/python/examples), it is really confusing to use where to start.\n\nUse tritony! You will get really short lines of code like example below.\n\n```python\nimport argparse\nimport os\nfrom glob import glob\nimport numpy as np\nfrom PIL import Image\n\nfrom tritony import InferenceClient\n\n\ndef preprocess(img, dtype=np.float32, h=224, w=224, scaling=\"INCEPTION\"):\n sample_img = img.convert(\"RGB\")\n\n resized_img = sample_img.resize((w, h), Image.Resampling.BILINEAR)\n resized = np.array(resized_img)\n if resized.ndim == 2:\n resized = resized[:, :, np.newaxis]\n\n scaled = (resized / 127.5) - 1\n ordered = np.transpose(scaled, (2, 0, 1))\n \n return ordered.astype(dtype)\n\n\nif __name__ == \"__main__\":\n parser = argparse.ArgumentParser()\n parser.add_argument(\"--image_folder\", type=str, help=\"Input folder.\")\n FLAGS = parser.parse_args()\n\n client = InferenceClient.create_with(\"densenet_onnx\", \"0.0.0.0:8001\", input_dims=3, protocol=\"grpc\")\n client.output_kwargs = {\"class_count\": 1}\n\n image_data = []\n for filename in glob(os.path.join(FLAGS.image_folder, \"*\")):\n image_data.append(preprocess(Image.open(filename)))\n\n result = client(np.asarray(image_data))\n\n for output in result:\n max_value, arg_max, class_name = output[0].decode(\"utf-8\").split(\":\")\n print(f\"{max_value} ({arg_max}) = {class_name}\")\n```\n\n## Release Notes\n\n- 24.07.11 Upgrade minimum tritonclient version to 2.34.0\n- 23.08.30 Support `optional` with model input, `parameters` on config.pbtxt\n- 23.06.16 Support tritonclient>=2.34.0\n- Loosely modified the requirements related to tritonclient\n\n\n## Key Features\n\n- [x] Simple configuration. Only `$host:$port` and `$model_name` are required.\n- [x] Generating asynchronous requests with `asyncio.Queue`\n- [x] Simple Model switching\n- [ ] Support async tritonclient\n\n## Requirements\n\n $ pip install tritonclient[all]\n\n## Install\n\n $ pip install tritony\n\n## Test\n\n### With Triton\n\n```bash\n./bin/run_triton_tritony_sample.sh\n```\n\n```bash\npytest -s --cov-report term-missing --cov=tritony tests/\n```\n\n### Example with image_client.py\n\n- Follow steps\n in [the official triton server documentation](https://github.com/triton-inference-server/server#serve-a-model-in-3-easy-steps)\n\n```bash\n# Download Images from https://github.com/triton-inference-server/server.git\npython ./example/image_client.py --image_folder \"./server/qa/images\"\n```\n",
"bugtrack_url": null,
"license": "BSD",
"summary": "Tiny configuration for Triton Inference Server",
"version": "0.0.19",
"project_urls": {
"Download": "https://pypi.org/project/tritony/#files",
"Homepage": "https://github.com/rtzr/tritony",
"Source": "https://github.com/rtzr/tritony",
"Tracker": "https://github.com/rtzr/tritony/issues"
},
"split_keywords": [
"grpc",
" http",
" triton",
" tensorrt",
" inference",
" server",
" service",
" client",
" nvidia",
" rtzr"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "a91cb6f8203240578c33053c1f353ed300453c44b36aa215c15b0105d7457c3b",
"md5": "a4759bbf078146b3e70153b48656ddd3",
"sha256": "24cfc5029368347c94e1ed8fb8a8a34ecfd440ab9187699d4dcf10802249504b"
},
"downloads": -1,
"filename": "tritony-0.0.19-py3-none-any.whl",
"has_sig": false,
"md5_digest": "a4759bbf078146b3e70153b48656ddd3",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 11142,
"upload_time": "2024-12-04T06:52:19",
"upload_time_iso_8601": "2024-12-04T06:52:19.603317Z",
"url": "https://files.pythonhosted.org/packages/a9/1c/b6f8203240578c33053c1f353ed300453c44b36aa215c15b0105d7457c3b/tritony-0.0.19-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "73e2e87a66d934e9f7a48c4bf5967045eb1e5042e52504d589ac8ac3d0481fc6",
"md5": "0d9b3015c2bddf94d9ae508001ea2b4a",
"sha256": "4b1c2d24ea54879fa2bf71a7dc5c4af07ec260f53557e7fc88865ea2678e56bb"
},
"downloads": -1,
"filename": "tritony-0.0.19.tar.gz",
"has_sig": false,
"md5_digest": "0d9b3015c2bddf94d9ae508001ea2b4a",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 13896,
"upload_time": "2024-12-04T06:52:20",
"upload_time_iso_8601": "2024-12-04T06:52:20.804925Z",
"url": "https://files.pythonhosted.org/packages/73/e2/e87a66d934e9f7a48c4bf5967045eb1e5042e52504d589ac8ac3d0481fc6/tritony-0.0.19.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-12-04 06:52:20",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "rtzr",
"github_project": "tritony",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "tritony"
}
JSON
