2,182 questions
Score of 0
0 answers
148 views
Running Inference on CPU and C++ using LiteRT
Problem:
I have two outputs, out_buffers.size() == 2 but when I read the data in the second tensor output_buffers[1] the data is rubbish. The first tensor, output_buffers[0] gives the same output as ...
- reputation score 1
Tooling
0
votes
1
replies
85
views
Offline-first face recognition in React Native using AWS Rekognition - how to cache face vectors locally and sync on reconnect?
I'm building a site access control kiosk app in React Native (targeting Android tablets) that uses face recognition for worker check in/check-out. The app must work fully offline - sites often have ...
- reputation score 1
Advice
0
votes
1
replies
52
views
How to integrate CNN model with Flutter app for plant disease detection - API vs TFLite?
I'm building a Flutter app for rice leaf disease detection using CNN as my thesis project.
I'm confused about the best architecture:
Option 1: Deploy CNN model as REST API (Python Flask/FastAPI), ...
- reputation score 1
Score of 0
0 answers
63 views
PTQ INT8 via TFLiteConverter — encoder-decoder seq2seq model loses encoder context entirely after conversion
I’m trying to deploy a seq2seq encoder-decoder model on an embedded target that only accepts INT8 TFLite models. The conversion via `TFLiteConverter` completes without errors, but the resulting model ...
- reputation score 1
Advice
0
votes
0
replies
114
views
How to use TFLite model signatures with InterpreterApi in LiteRT from Google Play Services ()?
I’m migrating my Android project from the bundled TensorFlow Lite to TensorFlow Lite via Google Play Services (LiteRT).
I use Movinet, which uses signatures and I’m feeding it individual video frames ...
- reputation score 43
Best practices
0
votes
2
replies
54
views
Optimization strategies for deploying multi-modal emotion recognition models on Nuvoton M55M1 (Cortex-M55) with Ethos-U55 NPU
I am currently working on an edge computing project.enter image description here The system analyzes both facial micro-expressions (via CMOS camera) and Electrodermal Activity (EDA) signals.
Hardware ...
- reputation score 1
Best practices
1
vote
3
replies
77
views
Normalization strategy for high-G (8.13g) sensors in TinyML and INT8 quantization?
Our team is building a specialized maritime logistics monitoring system designed to run on resource-constrained hardware, specifically an ARM Cortex-M4 microcontroller. We are using the TensorFlow ...
- reputation score 1
Score of 0
0 answers
55 views
Flutter TFLite: How to optimize dual-model inference (CNN + Res2Net) latency?
I'm working on a real-time detection app using Flutter. I need to run two TFLite models simultaneously: CNN for video and Res2Net for audio.
Current issue: The total inference time is around 500-600ms ...
- reputation score 11
Score of 1
1 answer
98 views
SIGSEGV in aidl-service-armnn-gpu when initializing TFLite NNAPI Delegate on specific Vendor Devices (Vivo/Mali)
I am developing a custom C++ Inference Engine for Android using the TensorFlow Lite C API. The engine is compiled as a shared library (.so) and loaded via JNI.
The engine works perfectly on Pixel and ...
- reputation score 3
Score of 1
0 answers
160 views
using Google LiteRT from a Tensorflow Lite model
data = tf.keras.utils.image_dataset_from_directory('snails', image_size=(256,256), shuffle=True)
class_names = data.class_names
num_classes = len(class_names)
print("Classes:", class_names)
...
- reputation score 11
Tooling
0
votes
1
replies
425
views
Looking for a working TensorFlow Lite (.tflite) NSFW detection model for Android
I specifically need:
-A real .tflite file (not .onnx, not .pb, not .pth, not .mlmodel)
-Image classifier model (NOT segmenter or detector)
-Preferably with 5 classes or at least 2: drawings, hentai, ...
- reputation score 1
Score of 2
0 answers
106 views
Issue Replicating TF-Lite Conv2D Quantized Inference Output
I am trying to reproduce the exact layer-wise output of a quantized EfficientNet model (TFLite model, TensorFlow 2.17) by re-implementing Conv2D, DepthwiseConv2D, FullyConnected, Add, Mul, Sub and ...
- reputation score 1
Score of 1
0 answers
55 views
Serial connection fails on very minor changes to tinyML voice keyword detection project (zephyr rtos, TFLM)
Below is the code snippet on dummy.cpp file, which is included in my main.c:
#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#include "dummy.h"
#include "...
- reputation score 11
Score of 2
1 answer
1065 views
Play Console warning: “Library that does not support 16 KB: base/lib/x86_64/libtensorflowlite_jni.so” when uploading AAB with TensorFlow Lite
I’m uploading my Android App Bundle (AAB) to Google Play Console and getting this warning:
Library that does not support 16 KB:
base/lib/x86_64/libtensorflowlite_jni.so after upgrading the lib
...
- reputation score 167
Score of 0
0 answers
133 views
Compiling Tensorflow/tflite C++ for ios
I want to build a framework for iOS so that I can run inference in C++ on both Android and iOS via Flutter.
So far I have managed to compile the libs for Android and succesfully loaded a model on my ...
- reputation score 1916