PULP-Dronet
Robotics modelDeep-learning visual navigation engine for the Crazyflie 2.1 nano-drone: PyTorch CNNs, quantization and deployment flow for the AI-deck GAP8 processor, onboard flight application and a collected dataset; code Apache-2.0, v3 dataset CC BY-NC-SA 4.0. [1] [3] [4]
Setup & usage →CheckpointsModel hub / homepageGetting started
- For
- Embedded-AI and nano-drone researchers able to train PyTorch models and flash Crazyflie and GAP8 firmware [4]
- Embodiments
- Bitcraze Crazyflie 2.1 with AI-deck (GAP8 + Himax camera) for v2/v3, Bitcraze Crazyflie 2.0 with PULP-Shield for v1 [1] [4]
- Source rights
- Named licenses are documented by part. Includes noncommercial terms. Review licenses
Source researched · Not tested by us
Sources checked · Research details
Licenses & related hardware
- License
- Apache-2.0, CC BY-NC-SA 4.0 (noncommercial) · Details by part
Overview
PULP-Dronet, from the PULP Platform group at ETH Zurich and the University of Bologna, is a family of DroNet-derived 2D convolutional neural networks with training, NEMO INT8 quantization, and DORY deployment code for autonomous visual navigation aboard a pocket-size quadrotor. At pinned release pulp-dronet-v3 (b68ad38670ba150f7b502e5a848c30d0991beb42), the repository targets the Bitcraze Crazyflie 2.1 equipped with an AI-deck (GreenWaves GAP8 RISC-V SoC and Himax grayscale camera); v1 targeted the PULP-Shield on Crazyflie 2.0. [1] [2] [4]
Model card summary
- Architecture
DroNet-derived 2D CNN family for ultra-low-power microcontrollers: PULP-Dronet v3 (ResBlock with bypass, depth multiplier 1.0) and Tiny-PULP-Dronet v3 (depthwise/pointwise separable convolutions without bypass, depth multiplier 0.125), quantized via NEMO and compiled via DORY for the GAP8 SoC.
- Base model
DroNet (University of Zurich)
- Parameter count
320 kB quantized (~320k params; 1.30 MB FP32 .pth) for PULP-Dronet v3 ResBlock 1.0, and 2.9 kB quantized (~2.9k params; 24 KB FP32 .pth) for Tiny-PULP-Dronet v3 dw-pw 0.125
- Input modalities
- rgb
Monocular grayscale Himax camera images on the Bitcraze AI-deck (mapped to closest schema enum rgb).
[4] · Editorial assessment
- Output modalities
- waypoints
Outputs scalar collision probability (obstacle within 2 m) and continuous steering yaw-rate command over UART to the Crazyflie flight controller (mapped to closest schema enum waypoints).
[4] · Editorial assessment
- Action space
Two scalar visual-navigation outputs per grayscale camera frame: collision probability (obstacle-within-2m avoidance trigger) and steering yaw-rate command, passed over UART from GAP8 to the Crazyflie STM32 flight application.
- Control frequency
19 Hz (PULP-Dronet v3) and 139 Hz (Tiny-PULP-Dronet v3) upstream-reported onboard inference rates on GAP8
- Trained / supported embodiments
- Bitcraze Crazyflie 2.1 with AI-deck (GAP8 + Himax camera) for v2/v3
- Bitcraze Crazyflie 2.0 with PULP-Shield for v1
- Training datasets
- PULP-Dronet v3 Crazyflie dataset on Zenodo (66k–77,418 labelled grayscale flight images, CC BY-NC-SA 4.0); v1 Udacity and Zurich datasets (MIT) in submodules
- Interfaces
- UART (GAP8 AI-deck to Crazyflie STM32)
- PyTorch .pth / NEMO ONNX / DORY C codegen
- cfclient fly parameter
Checkpoints
| Checkpoint | Format | Size | Access | Pinned revision |
|---|---|---|---|---|
| pulp-dronet-v3-resblock-1.0.pth | pytorch | 0.0013 GB | open | b68ad38670ba |
| tiny-pulp-dronet-v3-dw-pw-0.125.pth | pytorch | 0.000024 GB | open | b68ad38670ba |
Pretrained FP32 .pth weights committed directly in the repository at tag pulp-dronet-v3 (1,298,645 bytes and 24,085 bytes).
Is this for you?
- Use cases
- Onboard collision avoidance and steering on nano-drones
- Dataset collection on a nano-drone
- Study of CNN quantization and deployment on microcontrollers
- Known limitations
- Deployment pinned to crazyflie-firmware 2022.01 and gap-sdk 3.9.1; upstream warns new firmware versions are incompatible
- New AI-decks may need specific NINA firmware
- v3 dataset is noncommercial (CC BY-NC-SA 4.0) and downloaded separately
- Submodules (NEMO, DORY, viewer, modified firmware) not included in the archive
Source-review boundaries; no training or flight performed.
- Safety and limitations
- Onboard deployment is tightly pinned to crazyflie-firmware tag 2022.01, gap-sdk 3.9.1, crazyflie-clients-python 2022.01, and crazyflie-lib-python 0.1.17.1; upstream warns newer firmware versions are incompatible.
- v3 training dataset on Zenodo is licensed CC BY-NC-SA 4.0 (noncommercial); whether noncommercial dataset terms affect weights trained on it is unresolved here.
- Submodules (NEMO, DORY, viewer, modified firmware) require recursive git clone and include GPL-3.0 external components.
- No training, quantization, firmware flashing, or flight testing was performed by Open Source Robotics.
Compute & fine-tuning
- Inference compute
- Minimum VRAM: 0 GB
- Documented hardware: GreenWaves GAP8 RISC-V SoC on Bitcraze AI-deck (quantized INT8 onboard; no GPU required at inference)
- Precision:
int8 - Edge / consumer capable: Yes
Quantized via NEMO and deployed via DORY to run fully aboard the GAP8 microcontroller.
- Fine-tuning
- Supported: Yes
- Methods: PyTorch training via training.py followed by NEMO INT8 quantization and DORY GAP8 deployment
- Recipe: Upstream fine-tuning documentation
- Host computer
GPU host for PyTorch training (--gpu flag); onboard inference runs quantized INT8 directly on the GreenWaves GAP8 microcontroller aboard the Bitcraze AI-deck (JTAG cable required to flash GAP8).
- Operating system
- Linux
Inferred from gap-sdk 3.9.1, conda_deps.yml, and Crazyflie build toolchain.
[4] · Editorial assessment
- Dependencies
- Conda environment (conda_deps.yml) with PyTorch
- NEMO quantization tool
- DORY deployment tool (dory_dronet)
- GreenWaves gap-sdk 3.9.1
- crazyflie-firmware tag 2022.01
- crazyflie-clients-python 2022.01 and crazyflie-lib-python 0.1.17.1
Setup & usage
Open the upstream getting-started guide [4]
- Clone recursively to fetch NEMO, DORY and other submodules
- Create the conda environment from conda_deps.yml
- Download the v3 dataset from Zenodo for training or testing
- Install gap-sdk 3.9.1 and crazyflie-firmware tag 2022.01
- Flash the Crazyflie application and the GAP8 application, then start via the cfclient fly parameter
Editorial source-navigation sequence, not a reproduced or safety-qualified procedure.
[4] · Editorial assessment
Upstream-reported benchmarks
Upstream-reported only: Open Source Robotics has not run training, inference, or hardware evaluations for these benchmarks.
| Benchmark | Upstream-reported score | Scope & caveats |
|---|---|---|
| PULP-Dronet v3 GAP8 onboard throughput and test-set evaluation (collision accuracy & yaw-rate RMSE) | PULP-Dronet v3: 19 fps (320 kB); Tiny-PULP-Dronet v3: 139 fps (2.9 kB) on GAP8 | Upstream-reported throughput figures in tiny-pulp-dronet-v3/README.md; OSR has not trained, quantized, flashed, or flown the models. |
Project notes
Models, checkpoints, and onboard action interface
The v3 release ships two pretrained PyTorch checkpoints directly in tiny-pulp-dronet-v3/model/: pulp-dronet-v3-resblock-1.0.pth (1,298,645 bytes / 1.30 MB FP32; ResBlock with bypass, depth multiplier 1.0, 320 kB quantized) and tiny-pulp-dronet-v3-dw-pw-0.125.pth (24,085 bytes / 24 KB FP32; depthwise-pointwise separable convolutions without bypass, depth multiplier 0.125, 2.9 kB quantized). Both take monocular grayscale Himax camera frames and output two scalar values—collision probability (obstacle within 2 m) and steering yaw-rate—sent over UART from the GAP8 to the Crazyflie STM32 flight application. [4] [6] [8]
Deployment is strictly pinned to 2022 toolchains: gap-sdk 3.9.1, crazyflie-firmware tag 2022.01, crazyflie-clients-python 2022.01, and crazyflie-lib-python 0.1.17.1. Upstream warns that newer Crazyflie firmware is incompatible, newly bought AI-decks may need specific NINA firmware, flashing GAP8 requires a JTAG cable, and recursive cloning is needed to populate empty submodule paths (NEMO, DORY, viewer, and modified firmware). [4] [6]
Upstream-reported throughput and unverified execution
The v3 README reports onboard GAP8 inference at 19 fps (320 kB) for PULP-Dronet v3 and 139 fps (2.9 kB) for Tiny-PULP-Dronet v3, with testing.py evaluating collision accuracy and yaw-rate RMSE. All performance figures are upstream-reported; Open Source Robotics has not run training, quantization, inference, flashing, or flight tests. [4]
What we checked
Source research only; no training, installation, flashing or flight test performed. No affiliate links added.
Updates
- Upstream last updated
- Jul 3, 2025 [2]
- Latest release
- Mar 10, 2025 [2]
- Profile updated
- Nov 21, 2026
- Sources last checked
- Nov 21, 2026
- Next review
- Feb 19, 2027
- Reviewed revision
- b68ad38670ba150f7b502e5a848c30d0991beb42 (tag pulp-dronet-v3, release published 2025-03-10) · combined hardware and software repository
Separate license matrix
Code and weights are Apache-2.0; the v3 dataset is noncommercial and external, and deployment depends on pinned external firmware and SDK versions.
| Layer | Availability / license |
|---|---|
| Code | available · Apache-2.0 Repository files Apache-2.0 per LICENSE and LICENSE_README.md. External crazyflie-firmware and viewer submodules are GPL-3.0 and not included in the archive. [3] [7] |
| Model weights | available · Apache-2.0 Pre-trained .pth files are in the repository and fall under its Apache-2.0 statement; training-data terms for the v3 dataset (CC BY-NC-SA 4.0) are separate and not resolved here. [3] [4] [6] |
| Training datasets | available · CC BY-NC-SA 4.0 v3 dataset hosted on Zenodo under CC BY-NC-SA 4.0 per upstream README; the Zenodo record was not preserved (HTTP 403). v1 Udacity/Zurich subsets are MIT and live in submodules. [1] [3] [4] |
| Documentation | available · Apache-2.0 READMEs in the repository under its Apache-2.0 statement. [1] [3] |
Split code, weight, and dataset licenses
Repository code, documentation, and bundled .pth weights carry an Apache-2.0 license. However, the v3 flight dataset (reported as 66k images in the main README and 77,418 tabulated in the v3 README) is hosted on Zenodo under CC BY-NC-SA 4.0 (noncommercial), external Crazyflie firmware and viewer submodules are GPL-3.0, and v1 Udacity/Zurich subsets are MIT. Whether noncommercial dataset terms affect downstream reuse of weights trained on that data is not resolved here. [1] [3] [4] [5] [7]
References
Each source links to the exact revision we read. “Used for” shows where this page relies on it.
- S001 PULP-Dronet README
Project purpose, hardware basis, version history and license summary.Revision, date & where cited
PULP Platform (ETH Zurich and University of Bologna) · github.com · revision
b68ad38· README.md at (tag pulp-dronet-v3) · Accessed Oct 5, 2026
Used for: Overview, Model card summary, Is this for you?, Separate license matrix - S002 pulp-platform/pulp-dronet: release pulp-dronet-v3
Selected release, dates, popularity signals, Apache-2.0 license hint.Revision, date & where cited
PULP Platform (ETH Zurich and University of Bologna) · github.com · Tag pulp-dronet-v3 -> ; GitHub release API fields · Accessed Oct 5, 2026
Used for: Overview, What we checked - S003 LICENSE_README.md (per-folder and external-module licenses)
Apache-2.0 for repository files; MIT for v1 Udacity/Zurich datasets; GPL-3.0 for external Crazyflie firmware and viewer; CC BY-NC-SA 4.0 for the v3 dataset.Revision, date & where cited
PULP Platform (ETH Zurich and University of Bologna) · github.com · revision
b68ad38· Accessed Oct 5, 2026
Used for: Model card summary, Is this for you?, Separate license matrix - S004 Tiny-PULP-Dronet v3 README
Structure, requirements, dataset description, training/testing/quantization/deployment commands, version pins and compatibility warnings.Revision, date & where cited
PULP Platform (ETH Zurich and University of Bologna) · github.com · revision
b68ad38· Accessed Oct 5, 2026
Used for: Overview, Model card summary, Checkpoints, Is this for you? and 6 more - S005 Dataset collection framework README
Hardware used for dataset collection: Crazyflie 2.1, AI-deck, Flow deck v2, optional Multi-ranger deck.Revision, date & where cited
PULP Platform (ETH Zurich and University of Bologna) · github.com · revision
b68ad38· dataset_collection_framework/README.md opening · Accessed Oct 5, 2026
Used for: Separate license matrix - S006 Submodule declarations and v3 file inventory
Submodule paths are empty in the archive; pre-trained .pth weights are present.Revision, date & where cited
PULP Platform (ETH Zurich and University of Bologna) · github.com · revision
b68ad38· .gitmodules and tiny-pulp-dronet-v3 archive members at · Accessed Oct 5, 2026
Used for: Checkpoints, Is this for you?, Models, checkpoints, and onboard action interface, Separate license matrix - S007 Repository LICENSE
Apache License, Version 2.0 text.Revision, date & where cited
PULP Platform (ETH Zurich and University of Bologna) · github.com · revision
b68ad38· Accessed Oct 5, 2026
Used for: Separate license matrix - S008 GitHub API file inventory of tiny-pulp-dronet-v3/model/ at tag pulp-dronet-v3
Verified pretrained weights in repo: pulp-dronet-v3-resblock-1.0.pth (1,298,645 bytes / 0.0013 GB) and tiny-pulp-dronet-v3-dw-pw-0.125.pth (24,085 bytes / 0.000024 GB).Revision, date & where cited
PULP Platform (GitHub) · github.com · revision
b68ad38· GitHub REST API /contents/tiny-pulp-dronet-v3/model?ref= · Accessed Nov 21, 2026
Used for: Model card summary, Checkpoints, Models, checkpoints, and onboard action interface
Correction guidance · Discussion is not enabled · Profile format 0.8.0-draft.1
Change history
2026-10-05 — Initial intake from tag pulp-dronet-v3 (Growbotics comparison lead verified against upstream).
2026-11-21 — Migrated from 0.7.0-draft.5 project to 0.8.0-draft.1 model; added verified .pth checkpoint entries (1.30 MB and 24 KB), GAP8 INT8 edge compute profile, and upstream-reported throughput benchmarks.