Closed jesuino closed 1 year ago
Hi @jesuino, I went through all the steps described in the README "Serve a Model in 3 Easy Steps" section and was not able to reproduce this issue. From the JSON you got from wireshark, it seems like the model configuration is different from the one we provide here. Could you confirm that the config files for the models are correct? Could you also provide the full server log by running server with flag --log-verbose=1
?
Hi @krishung5 , I meet the same problem here. the step 3 fails for me with the following output: And Triton server seems working but receive no request : I wonder how to fix this problem. thanks!
Hi @rnwang04, what model are you using and how do you send the request? Are you using our own client with the same command /workspace/install/bin/image_client -m densenet_onnx -c 3 -s INCEPTION /workspace/images/mug.jpg
in the example?
Besides, could you confirm that the models are loaded successfully? If so, you should see something like this in the server log:
I0927 18:18:49.411099 1 server.cc:629]
+----------------------+---------+--------+
| Model | Version | Status |
+----------------------+---------+--------+
| densenet_onnx | 1 | READY |
| inception_graphdef | 1 | READY |
| simple | 1 | READY |
| simple_dyna_sequence | 1 | READY |
| simple_identity | 1 | READY |
| simple_int8 | 1 | READY |
| simple_sequence | 1 | READY |
| simple_string | 1 | READY |
+----------------------+---------+--------+
hi @krishung5 I just follow Serve a Model in 3 Easy Steps
and run following commands:
# Step 1: Create the example model repository
git clone -b r22.07 https://github.com/triton-inference-server/server.git
cd server/docs/examples
./fetch_models.sh
# Step 2: Launch triton from the NGC Triton container
sudo docker run --rm --net=host -v ${PWD}/model_repository:/models nvcr.io/nvidia/tritonserver:22.07-py3 tritonserver --model-repository=/models
# Step 3: Sending an Inference Request
# In a separate console, launch the image_client example from the NGC Triton SDK container
sudo docker run -it --rm --net=host nvcr.io/nvidia/tritonserver:22.07-py3-sdk
/workspace/install/bin/image_client -m densenet_onnx -c 3 -s INCEPTION /workspace/images/mug.jpg
and I confirm that the models are loaded successfully as I obtain following output: Besides, I have verified Triton Is Running Correctly But I still got the above error.
Thanks for confirming, @rnwang04. I followed the exact same commands you shared but still could not reproduce this issue. Filed a ticket for the team to investigate this further.
Closing this issue and related ticket due to inactivity. The README has also been updated a few times. Please let us know if you are still seeing this issue.
Description In README in "Serve a Model in 3 Easy Steps" section the step 3 fails for me with the following output:
From wireshark I can see that the request was not even sent, looks like it validates the request using a GET request to the model configuration.
I also tried with the other sample model, but looks like the image is not sent. This is the result I have:
Using wireshark I got the JSON sent, see [2]
Triton Information
The same from README
To Reproduce
Follow the steps from README
Expected behavior Have the inference result
[1]