Comments (4)
Hi @okyspace
I tested, when i first run my training but did not publish. When i start triton, this version is still avail in triton for inference. Is this correct?
You absolutely correct, here should be only_published=True
Let me make sure we fix that (or at least add a flag to control it)
The triton polling indicates no change, thus did not pull the models over.
How did you register the endpoint, with project/mode-name ? Any chance you can attach a log?
from clearml-serving.
Hi @bmartinn
- This is how I create my serving service and then register an endpoint.
clearml-serving triton --project "serving_engine_svs" --name "serving_engine_svs"
clearml-serving triton --project "serving_engine_svs" --name "serving_engine_svs" --endpoint "mnist_endpoint" --model-project "mnist_example" --model-name "mnist_task - mnist_model"
- Any idea where should I capture the useful logs? The triton logs basically shows this:
Skipping update, nothing changed
- Adding on more observations from related issue:
Once I add the model name to config file, the different versions (draft + published) are copied over to triton.
And this happens every polling iteration. But the published state did not take effect as well, i.e. last published version is still not available for inference.
from clearml-serving.
Once I add the model name to config file
What do you mean by that? Are you referring to the config.pbtxt ?
... last published version is still not available for inference.
I think I was able to confirm it. Can you test if the following works for you:
clearml-serving triton --model-tags deploy_me --project "serving_engine_svs" --name "serving_engine_svs" --endpoint "mnist_endpoint" --model-project "mnist_example" --model-name "mnist_task - mnist_model"
Then manually add the tag "deploy_me" to the selected models
(basically it should only get models with the selected project/name and the specified tag)
from clearml-serving.
Closing this issue as inactive, feel free to open a new issue and link to this one.
from clearml-serving.
Related Issues (20)
- clearml-serving not working with newer numpy 1.24 HOT 4
- Deploying Models from Azure Blob HOT 5
- Unable to canonicalize address from Kafka HOT 3
- AttributeError: module 'numpy' has no attribute 'int' HOT 2
- Incorrect shape size PyTorch HOT 2
- Configure options for gRPC (Triton server) HOT 3
- Unable to load onnx models into Triton HOT 7
- String not supported for Triotn HOT 1
- Error: Failed loading preprocess code for '<>': No module named 'transformers' HOT 1
- Where to find logging for preprocessing Custom model HOT 2
- triton model breaks serving instance HOT 4
- ValueError: dictionary update sequence element #0 has length <>; 2 is required HOT 2
- Error during creation endpoints with config.pbtxt HOT 4
- Docker image not up-to-date HOT 1
- torchserve support? HOT 1
- serving stuck because of deleted model
- Incorrect requests version specifier in requirements.txt HOT 3
- Shared functions across preprocess file HOT 2
- Removing model monitoring in endpoint HOT 1
- Multiple TensorRT handling(plan file per gpu)
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from clearml-serving.