🧠📦 vbjax_dynamics
*A JAX-based library for numerical integration of dynamical systems.*
🚀 GitHub: github.com/Ziaeemehr/vbjax_dynamics
⚙️ Features
• ODE Integration — Ordinary Differential Equations
– Efficient loop-based integrators with JIT compilation
– Full support for
• SDE Integration — Stochastic Differential Equations
–
–
– Euler–Maruyama scheme
– Fully reproducible with seeds
• DDE Integration — Delay Differential Equations
– Fixed delays supported
– History interpolation
• SDDE Integration — Stochastic Delay Differential Equations
– Combine stochastic and delayed dynamics
• JAX-Native Design
– Fully JIT-compiled
– Auto-differentiable
– GPU/TPU compatible
– Pure functional API
💡 Installation
*A JAX-based library for numerical integration of dynamical systems.*
🚀 GitHub: github.com/Ziaeemehr/vbjax_dynamics
⚙️ Features
• ODE Integration — Ordinary Differential Equations
– Efficient loop-based integrators with JIT compilation
– Full support for
jax.vmap (parallel trajectories)• SDE Integration — Stochastic Differential Equations
–
make_sde(): Integration with pre-generated noise –
make_sde_auto(): Automatic noise generation from keys – Euler–Maruyama scheme
– Fully reproducible with seeds
• DDE Integration — Delay Differential Equations
– Fixed delays supported
– History interpolation
• SDDE Integration — Stochastic Delay Differential Equations
– Combine stochastic and delayed dynamics
• JAX-Native Design
– Fully JIT-compiled
– Auto-differentiable
– GPU/TPU compatible
– Pure functional API
💡 Installation
pip install vbjax_dynamics
GitHub
GitHub - Ziaeemehr/vbjax_dynamics: JAX-based integrators for ordinary, stochastic, and delay differential equations
JAX-based integrators for ordinary, stochastic, and delay differential equations - Ziaeemehr/vbjax_dynamics
👍2
2026PhDAd.pdf
141 KB
PhD Position in Computational & Systems Neuroscience (Marseille, France): Neuromodulatory control of predictive processing in visual cortical circuits
⏰️ Deadline: January 28, 2026
⏰️ Deadline: January 28, 2026
Dear colleagues,
Applications for non-EU students are now open for the Master 2 in Computational Neuroscineces (CNS) at the University of Lyon (France). See our website for more details: https://masterneuro.univ-lyon1.fr/m2-cns/
The Lyon neurograduate school also offers fully funded scholarships to enrol in the CNS program or in other neurosciences training paths. Deadline : 23/01/2026, more information here : https://neurograduate.univ-lyon1.fr/fellowships/
The CNS program in Lyon trains students in modern computational and analytical methods to study the brain: from the electrical activity of single neurons to neural networks, interacting brain regions, and animal behavior. Students learn the principles of brain function by applying statistics, signal processing, and computational modeling to real neuroscientific datasets, gaining extensive hands-on experience through practical Python-based work. The main objective of CNS is to equip students with the skills and critical perspective needed to design, conduct, and interpret analyses of electrophysiological and behavioral data from a mechanistic point of view.
Contact info: Matteo Di Volo (matteo.di-volo@univ-lyon1.fr) and Jeremie Mattout (jeremie.mattout@inserm.fr)
Applications for non-EU students are now open for the Master 2 in Computational Neuroscineces (CNS) at the University of Lyon (France). See our website for more details: https://masterneuro.univ-lyon1.fr/m2-cns/
The Lyon neurograduate school also offers fully funded scholarships to enrol in the CNS program or in other neurosciences training paths. Deadline : 23/01/2026, more information here : https://neurograduate.univ-lyon1.fr/fellowships/
The CNS program in Lyon trains students in modern computational and analytical methods to study the brain: from the electrical activity of single neurons to neural networks, interacting brain regions, and animal behavior. Students learn the principles of brain function by applying statistics, signal processing, and computational modeling to real neuroscientific datasets, gaining extensive hands-on experience through practical Python-based work. The main objective of CNS is to equip students with the skills and critical perspective needed to design, conduct, and interpret analyses of electrophysiological and behavioral data from a mechanistic point of view.
Contact info: Matteo Di Volo (matteo.di-volo@univ-lyon1.fr) and Jeremie Mattout (jeremie.mattout@inserm.fr)
NEUROSCIENCES
Fellowships - NEUROSCIENCES
EXCHANGE FELLOWSHIPS The Neurograduate School of the Claude Bernard Lyon1 University aims to develop international training for undergraduate and graduate students (undergraduate and master studentships, doctoral fellowships, international invited professors…).…
❤1
Conduit is a lightweight relay server used by Psiphon that helps users in censored networks access the open internet by securely forwarding their traffic through volunteer-run nodes outside the censorship zone.
(for LINUX)
(for LINUX)
docker run -d --name conduit \
--cpus="2.0" \
--memory="2g" \
--ulimit nofile=65536:65536 \
-v conduit-data:/home/conduit/data \
--restart unless-stopped \
--log-driver local --log-opt max-size=10m --log-opt max-file=5 \
ghcr.io/ssmirr/conduit/conduit:latest start \
-b 5 -m 40 2>&1
### Docker command flags explained (brief)
- `docker run -d`
Run the container in detached (background) mode.
- `--name conduit`
Assign a fixed name to the container.
- `--cpus="2.0"`
Limit the container to using at most 2 CPU cores.
- `--memory="2g"`
Limit the container’s RAM usage to 2 GB.
- `--ulimit nofile=65536:65536`
Increase the maximum number of open file descriptors (important for many connections).
- `-v conduit-data:/home/conduit/data`
Persist Conduit data (keys/config) in a Docker volume.
- `--restart unless-stopped`
Automatically restart the container unless manually stopped.
- `--log-driver local`
Use Docker’s local logging driver.
- `--log-opt max-size=10m`
Rotate logs when they reach 10 MB.
- `--log-opt max-file=5`
Keep up to 5 rotated log files.
- `ghcr.io/ssmirr/conduit/conduit:latest`
The Docker image containing the Conduit server.
- `start`
Start the Conduit service inside the container.
- `-b 5`
Set the bandwidth limit (in MB/s).
- `-m 40`
Set the maximum number of concurrent clients.
Scientific Programming pinned «Conduit is a lightweight relay server used by Psiphon that helps users in censored networks access the open internet by securely forwarding their traffic through volunteer-run nodes outside the censorship zone. (for LINUX) docker run -d --name conduit \ …»
To see how many clients is connected and other details:
docker logs conduit --tail 1
🧠 New release: ModelingNeuralDynamics v1.0.0 is now on PyPI!
This is an open-source Python port of the code behind An Introduction to Modeling Neuronal Dynamics by Christoph Borgers, a widely used text for computational neuroscience courses, originally accompanied by MATLAB programs.
What's new:
📦 pip install modelingneuraldynamics, the shared helper package is now a real PyPI release, not just a clone-and-hope setup.
📓 22 chapters converted to Jupyter notebooks, single-neuron models, bifurcations, synaptic dynamics, network rhythms (PING/ING), and STDP, each as one self-contained, tested notebook.
🎛️ Interactive widgets, several notebooks use ipywidgets so you can drag sliders and watch firing rates, bifurcations, and phase planes respond live, instead of re-running cells by hand.
▶️ One-click Colab, every converted notebook opens directly in Google Colab and installs its own dependencies, no local setup required.
✅ CI-tested, every chapter runs in automated tests on every change, so the examples you open are the examples that actually run.
Whether you're teaching, TA-ing, or just curious how a Hodgkin-Huxley neuron becomes a gamma-rhythm network, it's free to explore:
🔗 GitHub: github.com/ITNG/ModelingNeuralDynamics
🔗 PyPI: pypi.org/project/modelingneuraldynamics
#ComputationalNeuroscience #OpenSource #Python #Neuroscience #Jupyter #ScientificComputing
This is an open-source Python port of the code behind An Introduction to Modeling Neuronal Dynamics by Christoph Borgers, a widely used text for computational neuroscience courses, originally accompanied by MATLAB programs.
What's new:
📦 pip install modelingneuraldynamics, the shared helper package is now a real PyPI release, not just a clone-and-hope setup.
📓 22 chapters converted to Jupyter notebooks, single-neuron models, bifurcations, synaptic dynamics, network rhythms (PING/ING), and STDP, each as one self-contained, tested notebook.
🎛️ Interactive widgets, several notebooks use ipywidgets so you can drag sliders and watch firing rates, bifurcations, and phase planes respond live, instead of re-running cells by hand.
▶️ One-click Colab, every converted notebook opens directly in Google Colab and installs its own dependencies, no local setup required.
✅ CI-tested, every chapter runs in automated tests on every change, so the examples you open are the examples that actually run.
Whether you're teaching, TA-ing, or just curious how a Hodgkin-Huxley neuron becomes a gamma-rhythm network, it's free to explore:
🔗 GitHub: github.com/ITNG/ModelingNeuralDynamics
🔗 PyPI: pypi.org/project/modelingneuraldynamics
#ComputationalNeuroscience #OpenSource #Python #Neuroscience #Jupyter #ScientificComputing
GitHub
GitHub - ITNG/ModelingNeuralDynamics: An Introduction to Modeling Neuronal Dynamics - Borgers in python, Single Neuron Models,…
An Introduction to Modeling Neuronal Dynamics - Borgers in python, Single Neuron Models, Mathematical Modeling, Computational Neuroscience, Hodgkin-Huxley Equations, Differential Equations, Brain R...
🔥2❤1