Skip to main content

Maixpy GO Mobilenet Transfer learning for Image Classfication

I have created a Colab notebook to perform transfer learning using Mobilenetv1 and then converts the model from h5 to tflite and then to kmodel. The kmodel file can be downloaded the Maixpy Go for realtime image classification.

The youtube video for the flower classification can be found here.

Labels:

Comments

  1. UnknownJuly 21, 2019 at 8:40 AM

    Hi,

    Thank you very much for posting such a good example. Your example was really helpful since I had been struggled to convert from a Keras model. I went through and could create a kmodel successfully.

    Thanks,
    Shigeru

    ReplyDelete
  2. gheesungJuly 21, 2019 at 9:55 PM

    You're welcome. I have created a thread in Maixpy bbs.

    ReplyDelete
  3. UnknownOctober 21, 2021 at 10:38 AM

    I tried to run, it runs but when I put it to identify the flower I get the following error:
    IndexError: list index out of range

    ReplyDelete
  4. UnknownOctober 21, 2021 at 10:42 AM

    help run mobilenet in maxduino
    hello, i'm doing some academic work and i saw your example of sorting flowers with mobilenet.

    I tried to run, it runs but when I put it to identify the flower I get the following error:
    IndexError: list index out of range.
    maybe I got it wrong in the txt file generation. Would you help me?

    ReplyDelete
  5. Angela BakerApril 10, 2025 at 9:04 PM

    Impressive work! Your mobile app showcases excellent functionality and a smooth user experience—truly top-notch development.

    visit: AI app development in Dubai

    ReplyDelete
  6. AnonymousMay 24, 2025 at 12:46 AM

    Modern Era Technology in 2025: Transforming the World We Live In


    Discover how modern era technology in 2025 is revolutionizing daily life, work, and industries through AI, IoT, and smart innovations.

    ReplyDelete
  7. jhonMay 28, 2025 at 5:24 PM

    "This smart video conferencing system by Skywire Broadcast looks like a robust solution for professional virtual meetings. With the integrated PTZ camera and speakerphone, it’s perfect for boardrooms, classrooms, or hybrid events. Impressive AV quality in a compact setup!"

    ReplyDelete
  8. Angela BakerJuly 2, 2025 at 8:53 PM

    Great insights! This blog provides valuable information that highlights the growing importance of mobile app development in Kuwait. Looking forward to more informative content like this!

    ReplyDelete
  9. AlmaSteeleJuly 12, 2025 at 2:37 PM

    MaixPy GO with MobileNet enables transfer learning flipper unleashed for efficient image classification on edge devices. It allows training custom models directly on low-power hardware like K210.










    ReplyDelete
  10. Mr TalhaAugust 8, 2025 at 11:01 AM

    That’s Iversær. A strange, slippery syllable tearing through TikTok like a digital wildfire. It’s not in the dictionary. It’s not in your textbook. Heck, it might not even be real. And that’s why it owns your feed.

    ReplyDelete
  11. Esferasoft Solutions Pvt. Ltd.August 25, 2025 at 7:07 PM

    Mobile App Development Company in Dubai UAE

    ReplyDelete
  12. hamza aliSeptember 11, 2025 at 8:42 PM

    During the last few years, technology’s impact on education has grown remarkably. A business that was once based only on books and chalkboards has transformed into a lively, tech-driven environment. Technology has changed how instructors and students learn, from computers in the classroom to the rise of e-learning platforms. The Useful Benefits of Technology in Education show that integration is becoming more than a trend; it’s a must that will increase communication, learning, and student preparation for the future.

    ReplyDelete
  13. HostpugSeptember 23, 2025 at 11:49 PM

    Great post — your walkthrough of using Mobilenet v1 + converting the model from .h5 → tflite → kmodel is super helpful, especially for people trying to do image classification on edge devices. I ran into some issues with quantization earlier, so seeing your process clears up a lot.

    Also, if anyone is looking for useful hosting solutions that support AI / ML workloads nicely, I recently came across https://hostpug.in Their plans are quite developer‐friendly and perform well for small to medium level ML projects.

    Looking forward to more posts like this — thanks again for sharing!

    ReplyDelete
  14. HostpugSeptember 23, 2025 at 11:50 PM

    Nice post <a href="https://hostpug.in>Hostpug</a>

    ReplyDelete
  15. HostpugSeptember 23, 2025 at 11:51 PM

    Nice post Hostpug

    ReplyDelete
  16. Teju Harpal October 17, 2025 at 3:12 AM

    Nice post

    ReplyDelete
  17. SEOOctober 23, 2025 at 4:26 PM

    Telecom Infrastructure Blogs delivers expert insights, industry trends, and in-depth analysis on the backbone of global communication networks - Telecom Infrastructure Blog

    ReplyDelete
  18. BellkahnNovember 12, 2025 at 1:03 PM

    This comment has been removed by the author.

    ReplyDelete

Post a Comment

Popular posts from this blog

Using ESP-Link transparent bridge (ESP-01 and Arduino Pro Mini)

Recently stumbled across an interesting open source project ESP-Link . Its main purpose is to network-enable a non-network microcontroller (MCU) such as Arduino Uno, Pro mini or Nano using ESP8266. The author termed it as "Transparent Bridge". The ESP and MCU  communicate via the serial link and there is a companion Arduino library EL-Client  for the MCU to connect up the network using MQTT, REST, TCP and UDP. Setup I have put together an ESP-01 and an Arduino Pro Mini for this experiment. I have chosen a 3.3 version Pro mini so that I do not need to do any voltage level shifting between the I/O pins. In order to have a stable voltage source, the ESP8266 is powered by Pro Mini and the Pro Mini "RAW" pin is connected to a 5v USB power source. The RAW pin can take voltage up to 12V. The reset pin of Pro Mini is connected to GPIO 0 of ESP-01. This enables the ESP-01 to reset the Pro Mini.   I have linked up an APDS 9960 sensor to it and periodically se...
5 comments
Read more

Using ESP-Link transparent bridge (ATmega2560+ESP8266 board)

I have found this interesting board selling on Aliexpress website. It is an Arduino Mega 2650 with a built-in ESP8266 which allows the Mega to connect to wifi. This board has cut down a lot of wiring and bring out interesting possibilities. One of this is to connect up both the microcontrollers using ESP-link firmware. I have an earlier blog post  that described how to wire up an ESP-01 and Arduino Pro Mini. One of the tricky parts in this setup is the sequence of flashing the ESP firmware. Flashing the onboard ESP8266 requires some setting changes to the dip switches. The following are the sequences which I get both the microcontrollers to talk to each other. Set the jumper 5,6,7 to ON and the rest OFF. This will set the ESP to update mode. I am using version 3.0.14  and flashed the ESP using the following parameters: esptool --port com26 --baud 230400 write_flash -fm dio -fs 4MB -ff 80m 0x0 boot_v1.6.bin 0x1000 user1.bin 0x3fc000 esp_init_data_defa...
3 comments
Read more

Xiaomi Mi Flora Chinese and International version comparison

Xiaomi Miflora sensor allows monitoring of the surrounding environment of the plant. It can monitor temperature, soil moisture, conductivity (acidity of the soil), ambience light. There are 2 versions of the sensor, international and Chinese version. Realistically, I cannot see the difference between the 2 versions. The left-hand side is the Chinese version and the right is the international version. Other than the packaging, the main difference is the price of the sensor. The International version cost 2x more. In other forums, there are discussions that the Chinese version cannot connect to the MiHome App. I have set my MiHome app to connect to China server and it is able to register the sensor correctly. For my usage, I will be connecting to Openhab to monitor the plants andusing  Thomas Dietrich MiFlora mqtt daemon  as the bridge between Openhab and the sensor. The International version is detected as Flower care and the firmware version is 3.1.9 wh...
1 comment
Read more