# Flower Help - Beginners **URL:** https://discuss.flower.ai/c/flower-help-beginners/5.md [Latest](https://discuss.flower.ai/latest.md) · [Categories](https://discuss.flower.ai/categories.md) · [Tags](https://discuss.flower.ai/tags.md) --- ## [About the Flower Help - Beginners category](https://discuss.flower.ai/t/about-the-flower-help-beginners-category/27)
**Author:** [@flower](https://discuss.flower.ai/u/flower)\ **Replies:** 0
Everybody has to start somewhere. Are you relatively new to Federated Learning and/or Flower? Or maybe you just have a basic question? Use this Category to look for answers! --- ## [\[Please help\] How to run the server\_app.py / clinet\_app.py on sprarate devices? ](https://discuss.flower.ai/t/please-help-how-to-run-the-server-app-py-clinet-app-py-on-sprarate-devices/1265)
**Author:** [@flwrforhw](https://discuss.flower.ai/u/flwrforhw)\ **Replies:** 0\ **Last updated:** [July 27, 2026, 9:34am UTC](https://discuss.flower.ai/t/please-help-how-to-run-the-server-app-py-clinet-app-py-on-sprarate-devices/1265 "2026-07-27T09:34:02Z")
Hello. I am using two raspberry pi 4B to build a simple FL. But i dont know how can i run the server script in one pi and run the client script in another pi. What command i can use? --- ## [New to flwr and having problem in running my first program. ](https://discuss.flower.ai/t/new-to-flwr-and-having-problem-in-running-my-first-program/1254)
**Author:** [@shubhamets](https://discuss.flower.ai/u/shubhamets)\ **Replies:** 1\ **Last updated:** [June 25, 2026, 8:36am UTC](https://discuss.flower.ai/t/new-to-flwr-and-having-problem-in-running-my-first-program/1254 "2026-06-25T08:36:23Z")
Hello, I am new to flwr and trying to follow the website instructions. But I am getting the following error. Can someone help me? :white\_check\_mark: Python 3.11.9 :white\_check\_mark: Ray 2.55.1 installed … --- ## [Unable to run the last part of the PyTorch tutorial, Communicating Arbitrary Objects](https://discuss.flower.ai/t/unable-to-run-the-last-part-of-the-pytorch-tutorial-communicating-arbitrary-objects/1250)
**Author:** [@mskbey](https://discuss.flower.ai/u/mskbey)\ **Replies:** 1\ **Last updated:** [June 14, 2026, 4:30pm UTC](https://discuss.flower.ai/t/unable-to-run-the-last-part-of-the-pytorch-tutorial-communicating-arbitrary-objects/1250 "2026-06-14T16:30:06Z")
Hey everyone. I’m trying to complete the PyTorch tutorial, and I’m stuck on the very last portion of part 4, Communicating Arbitrary Objects (Communicate custom Messages - Flower Framework) I had to fix quite a few undo… --- ## [Gossip Learning Simulation](https://discuss.flower.ai/t/gossip-learning-simulation/1249)
**Author:** [@khiro](https://discuss.flower.ai/u/khiro)\ **Replies:** 2\ **Last updated:** [June 12, 2026, 11:35pm UTC](https://discuss.flower.ai/t/gossip-learning-simulation/1249 "2026-06-12T23:35:42Z")
Hi, I wanted to know if there is a way to handle gossip learning simulation with Flower where I have a P2P network and the nodes of the network can interact with each other. Thank you --- ## [Finetuning Lizzy](https://discuss.flower.ai/t/finetuning-lizzy/1247)
**Author:** [@werner](https://discuss.flower.ai/u/werner)\ **Replies:** 1\ **Last updated:** [June 9, 2026, 10:20am UTC](https://discuss.flower.ai/t/finetuning-lizzy/1247 "2026-06-09T10:20:07Z")
Hi, probably not the right place to ask. But could not find Lizzy-7B related stuff here. I’d like to fine-tune Lizzy7B to German instruct examples. I have already a work flow that’s battle tested on Mistral 7B v03 and… --- ## [Understanding flower connection, communication, flow](https://discuss.flower.ai/t/understanding-flower-connection-communication-flow/1239)
**Author:** [@samoreno](https://discuss.flower.ai/u/samoreno)\ **Replies:** 1\ **Last updated:** [May 13, 2026, 1:30pm UTC](https://discuss.flower.ai/t/understanding-flower-connection-communication-flow/1239 "2026-05-13T13:30:11Z")
Hello you all! I’m running Deploy Flower on Multiple Machines with Docker Compose - Flower Framework . I did run the example, now I’m trying to extend it but I have some doubts about it. I’m trying to deploy two client… --- ## [Flower run toy examples error. ](https://discuss.flower.ai/t/flower-run-toy-examples-error/1231)
**Author:** [@samoreno](https://discuss.flower.ai/u/samoreno)\ **Replies:** 3\ **Last updated:** [May 11, 2026, 9:41am UTC](https://discuss.flower.ai/t/flower-run-toy-examples-error/1231 "2026-05-11T09:41:51Z")
Hi! I’m diving into flower and I just start with the initial toy examples. Specifically I’m running the pytorch example app. I follow the steps in the tutorial, i.e., Install flower, set up pytorch example with flwr n… --- ## [Learning to use Flower and ML with a concert case - My Fitbit data ](https://discuss.flower.ai/t/learning-to-use-flower-and-ml-with-a-concert-case-my-fitbit-data/1230)
**Author:** [@jmuilu](https://discuss.flower.ai/u/jmuilu)\ **Replies:** 2\ **Last updated:** [April 17, 2026, 5:25pm UTC](https://discuss.flower.ai/t/learning-to-use-flower-and-ml-with-a-concert-case-my-fitbit-data/1230 "2026-04-17T17:25:25Z")
How about implementing concrete study case using for example data from wearables like Fitbit… ? Would be nice to participate and learn to use Flower and ML techniques with a concrete case. --- ## [Federated learning using Supergrid - Different systems](https://discuss.flower.ai/t/federated-learning-using-supergrid-different-systems/1232)
**Author:** [@bmagoba](https://discuss.flower.ai/u/bmagoba)\ **Replies:** 1\ **Last updated:** [April 17, 2026, 3:01pm UTC](https://discuss.flower.ai/t/federated-learning-using-supergrid-different-systems/1232 "2026-04-17T15:01:46Z")
Hello everyone, My name is Bridget Magoba, a PhD fellow and very new to Flower. My research will involve federated learning across different systems and different formats, e.g., a hospital eHR, A lab system, mobility da… --- ## [Client selection in a single Flower federation with persistent client identities and metadata-based sampling](https://discuss.flower.ai/t/client-selection-in-a-single-flower-federation-with-persistent-client-identities-and-metadata-based-sampling/1161)
**Author:** [@flippchen](https://discuss.flower.ai/u/flippchen)\ **Replies:** 6\ **Last updated:** [January 12, 2026, 2:25pm UTC](https://discuss.flower.ai/t/client-selection-in-a-single-flower-federation-with-persistent-client-identities-and-metadata-based-sampling/1161 "2026-01-12T14:25:13Z")
Hi, First of all, thanks for the framework. I’m relatively new to Flower and currently using the Docker deployment on different devices. My current goal is to have a single federation and then decide, per training run,… --- ## [I was reviewing the train function in the quickstart-pytorch example and noticed a logical issue in how the average training loss is calculated.](https://discuss.flower.ai/t/i-was-reviewing-the-train-function-in-the-quickstart-pytorch-example-and-noticed-a-logical-issue-in-how-the-average-training-loss-is-calculated/1159)
**Author:** [@lefthook](https://discuss.flower.ai/u/lefthook)\ **Replies:** 3\ **Last updated:** [January 12, 2026, 2:03pm UTC](https://discuss.flower.ai/t/i-was-reviewing-the-train-function-in-the-quickstart-pytorch-example-and-noticed-a-logical-issue-in-how-the-average-training-loss-is-calculated/1159 "2026-01-12T14:03:13Z")
In the current implementation(task.py), running\_loss is initialized outside the epoch loop and accumulates the loss across all epochs. However, when calculating the average at the end, it is divided only by len(trainload… --- ## [Secure connections with certificates](https://discuss.flower.ai/t/secure-connections-with-certificates/1106)
**Author:** [@johannes](https://discuss.flower.ai/u/johannes)\ **Replies:** 5\ **Last updated:** [January 7, 2026, 9:02am UTC](https://discuss.flower.ai/t/secure-connections-with-certificates/1106 "2026-01-07T09:02:03Z")
Hi, I am trying to run flower in secure mode using server certificates from letsencrypt. (Using insecure mode without certifiactes on one single machine worked fine.) From letsencrypt I got the certificates privkey.pem,… --- ## [Using the Client and Server for different predicitions](https://discuss.flower.ai/t/using-the-client-and-server-for-different-predicitions/1148)
**Author:** [@neonduck](https://discuss.flower.ai/u/neonduck)\ **Replies:** 1\ **Last updated:** [December 30, 2025, 10:59am UTC](https://discuss.flower.ai/t/using-the-client-and-server-for-different-predicitions/1148 "2025-12-30T10:59:18Z")
I’m currently developing a flower application that runs the server on a smart home gateway (RPi 5 DiyHue) and is trained on different IoT devices (Ip camera, smart home control panel, etc.). I want to know if it’s able t… --- ## [the problem on docker when using flower](https://discuss.flower.ai/t/the-problem-on-docker-when-using-flower/1143)
**Author:** [@nielinhai](https://discuss.flower.ai/u/nielinhai)\ **Replies:** 1\ **Last updated:** [December 2, 2025, 12:07pm UTC](https://discuss.flower.ai/t/the-problem-on-docker-when-using-flower/1143 "2025-12-02T12:07:29Z")
When I was following the process in the ‘Quickstart with Docker’ section on https://flower.ai/docs/framework/docker/tutorial-quickstart-docker.html, I encountered the following issue when running the command flwr run . l… --- ## [Noob question network strugling](https://discuss.flower.ai/t/noob-question-network-strugling/1128)
**Author:** [@smartes](https://discuss.flower.ai/u/smartes)\ **Replies:** 1\ **Last updated:** [November 8, 2025, 4:59pm UTC](https://discuss.flower.ai/t/noob-question-network-strugling/1128 "2025-11-08T16:59:17Z")
Hi! I want to run the first deployment over net. I successfully pass quick start example for local approach Run Flower with the Deployment Engine - Flower Framework But now Im struggling with network setup. My primitiv… --- ## [Initial model distribution on deployed Flower architecture](https://discuss.flower.ai/t/initial-model-distribution-on-deployed-flower-architecture/1124)
**Author:** [@ddosantos](https://discuss.flower.ai/u/ddosantos)\ **Replies:** 3\ **Last updated:** [November 6, 2025, 8:28am UTC](https://discuss.flower.ai/t/initial-model-distribution-on-deployed-flower-architecture/1124 "2025-11-06T08:28:40Z")
I am currently exploring the Flower architecture deployment runtime for my thesis topic. I would like to ask whether this initial model distribution is bundled with the server and client code in the Flower Application Bu… --- ## [What is the difference between flwr.server and flwr.serverapp?](https://discuss.flower.ai/t/what-is-the-difference-between-flwr-server-and-flwr-serverapp/1117)
**Author:** [@mcarmo](https://discuss.flower.ai/u/mcarmo)\ **Replies:** 1\ **Last updated:** [October 23, 2025, 9:10pm UTC](https://discuss.flower.ai/t/what-is-the-difference-between-flwr-server-and-flwr-serverapp/1117 "2025-10-23T21:10:21Z")
Hello there! I am a student currently learning Flower for a research project, as I am planning on doing a Split Learning / SplitFed implementation using it. In this context, I am playing around with a few projects, learn… --- ## [PyTorch example clients into TinyML](https://discuss.flower.ai/t/pytorch-example-clients-into-tinyml/1066)
**Author:** [@neonduck](https://discuss.flower.ai/u/neonduck)\ **Replies:** 6\ **Last updated:** [September 15, 2025, 4:13pm UTC](https://discuss.flower.ai/t/pytorch-example-clients-into-tinyml/1066 "2025-09-15T16:13:51Z")
Hi, I’m using the pytorch example project and I want to use TinyML \[1\]. How can I do this? Is this something I can do with flower directly or do I have to go another way? \[1\] Federated learning for IoT devices: Enhanci… --- ## [issue in dependency](https://discuss.flower.ai/t/issue-in-dependency/1099)
**Author:** [@amine](https://discuss.flower.ai/u/amine)\ **Replies:** 1\ **Last updated:** [September 11, 2025, 10:11am UTC](https://discuss.flower.ai/t/issue-in-dependency/1099 "2025-09-11T10:11:02Z")
ERROR: Cannot install -r requirements\_client.txt (line 16), -r requirements\_client.txt (line 19), grpcio==1.43.0 and tensorflow-aarch64 because these package versions have conflicting dependencies. The conflict is cause… --- ## [Would you please provide an example for LLM fine-tuning and with MLX support?](https://discuss.flower.ai/t/would-you-please-provide-an-example-for-llm-fine-tuning-and-with-mlx-support/230)
**Author:** [@chrischaochn](https://discuss.flower.ai/u/chrischaochn)\ **Replies:** 2\ **Last updated:** [September 2, 2025, 9:38pm UTC](https://discuss.flower.ai/t/would-you-please-provide-an-example-for-llm-fine-tuning-and-with-mlx-support/230 "2025-09-02T21:38:32Z")
We read this, which is excellent to start with CUDU GPUs: Is it possible for you provide another version, which support MLX (Apple Silicon) and LLM? Thank you. --- ## [Gen AI 'Federated' policy-as-code Governance](https://discuss.flower.ai/t/gen-ai-federated-policy-as-code-governance/1095)
**Author:** [@avolyn](https://discuss.flower.ai/u/avolyn)\ **Replies:** 0\ **Last updated:** [August 30, 2025, 5:09pm UTC](https://discuss.flower.ai/t/gen-ai-federated-policy-as-code-governance/1095 "2025-08-30T17:09:12Z")
Hello, I am brand new to flower.ai, but I have a question that might lend itself to intermediate or advanced users - and it could very well be my limited understanding of how flower works, that I might be, ‘barking up t… --- ## [Clients not initializing in authexample](https://discuss.flower.ai/t/clients-not-initializing-in-authexample/1079)
**Author:** [@dj0nty](https://discuss.flower.ai/u/dj0nty)\ **Replies:** 8\ **Last updated:** [August 19, 2025, 7:45am UTC](https://discuss.flower.ai/t/clients-not-initializing-in-authexample/1079 "2025-08-19T07:45:38Z")
Im using the authexample, running local. And i have modified the project to use XGBoost instead of a neural net, I am also using a custom dataset imported from a CSV. It appears that the initial authentication is done co… --- ## [Not able to build CPP example](https://discuss.flower.ai/t/not-able-to-build-cpp-example/1074)
**Author:** [@neonduck](https://discuss.flower.ai/u/neonduck)\ **Replies:** 2\ **Last updated:** [August 7, 2025, 2:07pm UTC](https://discuss.flower.ai/t/not-able-to-build-cpp-example/1074 "2025-08-07T14:07:58Z")
Hi, I have a problem on building the C++ example -- Build files have been written to: /home/jan56580/Documents/dev/flower/examples/quickstart-cpp/build/\_deps/flwr\_repo-subbuild \[ 11%\] Performing update step for 'flwr\_r… --- ## [Return a \`List\` of metrics from client to server](https://discuss.flower.ai/t/return-a-list-of-metrics-from-client-to-server/1070)
**Author:** [@chongshenng](https://discuss.flower.ai/u/chongshenng)\ **Replies:** 0\ **Last updated:** [August 5, 2025, 3:55pm UTC](https://discuss.flower.ai/t/return-a-list-of-metrics-from-client-to-server/1070 "2025-08-05T15:55:23Z")
Reposted from Slack --- ## [How to use local datasets with Flower Datasets](https://discuss.flower.ai/t/how-to-use-local-datasets-with-flower-datasets/1069)
**Author:** [@chongshenng](https://discuss.flower.ai/u/chongshenng)\ **Replies:** 0\ **Last updated:** [August 5, 2025, 3:53pm UTC](https://discuss.flower.ai/t/how-to-use-local-datasets-with-flower-datasets/1069 "2025-08-05T15:53:09Z")
Reposted from Slack --- ## [Hyperparameter Selection with Optuna and Flower](https://discuss.flower.ai/t/hyperparameter-selection-with-optuna-and-flower/1068)
**Author:** [@chongshenng](https://discuss.flower.ai/u/chongshenng)\ **Replies:** 0\ **Last updated:** [August 5, 2025, 3:50pm UTC](https://discuss.flower.ai/t/hyperparameter-selection-with-optuna-and-flower/1068 "2025-08-05T15:50:26Z")
Reposted from Slack. --- ## [Unable to start federated training on CUDA11.8 when using YOLOv8 model on NVIDIA A10G GPU](https://discuss.flower.ai/t/unable-to-start-federated-training-on-cuda11-8-when-using-yolov8-model-on-nvidia-a10g-gpu/133)
**Author:** [@shubham.ecebtech14](https://discuss.flower.ai/u/shubham.ecebtech14)\ **Replies:** 2\ **Last updated:** [August 1, 2025, 3:29pm UTC](https://discuss.flower.ai/t/unable-to-start-federated-training-on-cuda11-8-when-using-yolov8-model-on-nvidia-a10g-gpu/133 "2025-08-01T15:29:30Z")
I wish to report one strange observation. I was able to run my federated code with available CPU. Today, I installed CUDA 11.8 and tried to run the code. It is getting stuck and not able to start training. To reproduce … --- ## [How to Manage Learning Rate Scheduling (and similar operations) in Flower?](https://discuss.flower.ai/t/how-to-manage-learning-rate-scheduling-and-similar-operations-in-flower/1025)
**Author:** [@mabduaguye](https://discuss.flower.ai/u/mabduaguye)\ **Replies:** 4\ **Last updated:** [July 17, 2025, 12:55pm UTC](https://discuss.flower.ai/t/how-to-manage-learning-rate-scheduling-and-similar-operations-in-flower/1025 "2025-07-17T12:55:48Z")
Hello, I recently started using Flower for a federated learning project I am working on. As a sanity check, I implemented a centralized training baseline on CIFAR-10 using a ResNet-18 model. The training recipe in this … --- ## [Strategy with multiple communication rounds per iteration](https://discuss.flower.ai/t/strategy-with-multiple-communication-rounds-per-iteration/313)
**Author:** [@rmazzier](https://discuss.flower.ai/u/rmazzier)\ **Replies:** 7\ **Last updated:** [July 16, 2025, 3:16pm UTC](https://discuss.flower.ai/t/strategy-with-multiple-communication-rounds-per-iteration/313 "2025-07-16T15:16:51Z")
Hi everyone, I’m working on implementing the GIANT algorithm (\[1709.03528\] GIANT: Globally Improved Approximate Newton Method for Distributed Optimization), a second-order distributed optimization algorithm that aims to… [Next page](https://discuss.flower.ai/c/flower-help-beginners/5.md?page=1)