Questions tagged [tensorflow-serving]
The tensorflow-serving tag has no usage guidance, but it has a tag wiki.
tensorflow-serving
1,259
questions
0
votes
0
answers
11
views
Compute specs required to build the tensorflow serving docker image
I'm trying to build the Tensorflow Serving docker image from scratch. Currently, I'm burning every CPU I can attempt to spin up. There is no official documentation regarding this.
Does anyone have any ...
0
votes
0
answers
22
views
Issue: StatusCode.FAILED_PRECONDITION
I try to serve model on TFserver through docker tensorflow/serving:2.16.1 and got this issue:
`<_InactiveRpcError of RPC that terminated with:
status = StatusCode.FAILED_PRECONDITION
details = &...
0
votes
0
answers
8
views
Tensorflow Serving: Adding warm start data at runtime
I'm trying to dynamically add warm-start data for our models via the SavedModel Warmup method (https://www.tensorflow.org/tfx/serving/saved_model_warmup). In our case, we need to actually have the ...
0
votes
0
answers
11
views
Multiple Model Configuration in tensorflow serving
i have created a file model.config for configuration detail and it is inside model_config and model_config is inside untitlled folder from where iam executing my script and getting the path not ...
0
votes
0
answers
12
views
Handle label feature in TFX in different environments
I'm new in MLOps and trying to figure out how to work with label feature in data. I read that for the uniformity of the data it is necessary to use the same schema for both the training and validation ...
0
votes
0
answers
32
views
Ragged Tensor as an output from Tensorflow serving
We use tensorflow serving to serve models in production. We have a use case where the output of the model is a ragged tensor.
To see if the tensorflow serving supports ragged tensor as output, we ...
0
votes
1
answer
15
views
Is there another loss that can replace seq2seq.sequence_loss in tensorflow
I am running a CVAE for text generation. I am using tensorflow > 2.0.
The problem is that for my loss I use the seq2seq.sequence_loss. I tried to update the tensorflow v1 to v2 since the code was ...
0
votes
0
answers
14
views
Why Tensorflow SavedModel is a directory
People who designed the SavedModelBundle definitely had a very good reason to make it export and load models to a directory on a disk. However doesn't this defeat the very important purpose of ...
0
votes
0
answers
42
views
Error with loading a .pb model for prediction: Op type not registered 'DecodeProtoSparseV4'
By running this command to load a trained model (saved_model.pb) for doing the prediction:
model_directory = 'C:/Users/.../predict/001/'
model = tf.saved_model.load(model_directory)
I get the ...
1
vote
0
answers
91
views
Serving keras model with tensorflow serving error
I have created a model using keras which is working locally. But after upgrade of tensorflow to the 2.17.0 I start receiving strange error on the tsserve side.
Model was serialized so:
import ...
0
votes
1
answer
77
views
TFLite converter not replacing dummy_function with TFLite_Detection_PostProcess
I want to implement a tf.Module for decoding box predictions and applying NonMaxSuppression that is convertible to tflite.
This implementation includes elements from here.
It also follows this guide ...
0
votes
0
answers
85
views
Tensorflow Serving prometheus metrics is unclear and latency is high
Background
We have deployed a service with Tensorflow Serving (TFServing) container to Kubernetes with server-side batching enabled. When the service receives a inference request, it invokes TFServing ...
0
votes
0
answers
22
views
Key and output issues with Tensorflow serving
My TF serving script is not predicting the correct result.
I am sending a request through grpc, here is my full code:
import grpc
import numpy as np
import tensorflow as tf
from tensorflow_serving....
0
votes
0
answers
66
views
How can I display logs for models served by TensorFlow Serving using GRPC?
I am using TensorFlow Serving and the GRPC pattern for serving models. However, I have encountered problems in displaying logs for the served models. While I've successfully used the env var ...
0
votes
1
answer
90
views
Error in HTTP POST request from Flask docker container to Tensorflow/Serving container
I am trying to deploy my micro projects using docker compose on AWS EC2. The images of major containers are Nginx, Flask, MySQL, and Tensorflow/Serving.
I have successfully made connection between ...