Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Wire bonding and lead impedance calculation in Blender FastHenry! I've improved how the simulation results are visualised, the entire objects now gets overlaid with a colour rather than using bound boxes. Red is for showing the self-inductance of a selected objected, yellow for mutually coupled objects and blue for... show more

Sam M

9,039 subscribers

29,801 views • 1 year ago •via X (Twitter)

Science & Technology

Anya Rossi• Live Now

Private livecam show

9 Comments

outside five sigma1 year ago

Are we using the same blender

α1 year ago

"Applied Blender". Sublime job Sam ! Is there a Patreon somewhere to support you for what you are doing?

Moengke, Hungry Ghost1 year ago

Impressive and exciting! I'm inspired by how far Blender has been taken as a multi-physics engine

Guttural Prayer1 year ago

Looks awesome as always.

۟1 year ago

damn amazing

labo_electronico1 year ago

Impressive work...

Dr. Goldoval1 year ago

gadzooks!

Turbo1 year ago

What does this help you do better

Aiden⚡1 year ago

Interesting

Related Videos

The compositor in Blender is really fast now that it can use the GPU. I've made this cool x-ray effect for my upcoming FastHenry integration project, it makes it easier to visualise bond wires #b3d #blender3d #electronics #engineering #geometrynodes #science #physics #simulation #pcbdesign #kicad #altium #electromagnetics #educational #xray #3dmodeling

The compositor in Blender is really fast now that it can use the GPU. I've made this cool x-ray effect for my upcoming FastHenry integration project, it makes it easier to visualise bond wires #b3d #blender3d #electronics #engineering #geometrynodes #science #physics #simulation #pcbdesign #kicad #altium #electromagnetics #educational #xray #3dmodeling

Sam M

34,741 views • 1 year ago

Calculation of busbar impedance with Blender Fasthenry. You can quickly create busbar shapes with curvatures and turns and calculate their resistance and inductance at different frequencies. Also, you can now change the width and height of individual segments. Dimensions are also visualised! #b3d #geometrynodes #electronics #Engineering #Science #pcbdesign #hardware #physics #Simulation #kicad

Calculation of busbar impedance with Blender Fasthenry. You can quickly create busbar shapes with curvatures and turns and calculate their resistance and inductance at different frequencies. Also, you can now change the width and height of individual segments. Dimensions are also visualised! #b3d #geometrynodes #electronics #Engineering #Science #pcbdesign #hardware #physics #Simulation #kicad

Sam M

23,642 views • 1 year ago

You can now use Blender Fasthenry to do PCB layout studies, calculate impedances of loops and return paths and determine how many vias you need! This demo shows how the impedance of DC decoupling capacitor is impacted by its distance from the IC, and the number of vias under the IC and near the capacitor. This is for a two-layer 1.6mm PCB at 1MHz. Blender FastHenry is now a solid tool for layout simulations and impedance calculations, available on Github. #b3d #geometrynodes #engineering #science #electronics #Simulation #kicad #altium #hardware #technology #3dmodeling #pcbdesign hackaday Maurizio Di Paolo Emilio

You can now use Blender Fasthenry to do PCB layout studies, calculate impedances of loops and return paths and determine how many vias you need! This demo shows how the impedance of DC decoupling capacitor is impacted by its distance from the IC, and the number of vias under the IC and near the capacitor. This is for a two-layer 1.6mm PCB at 1MHz. Blender FastHenry is now a solid tool for layout simulations and impedance calculations, available on Github. #b3d #geometrynodes #engineering #science #electronics #Simulation #kicad #altium #hardware #technology #3dmodeling #pcbdesign hackaday Maurizio Di Paolo Emilio

Sam M

37,530 views • 1 year ago

Very excited to share the release of my addon Blender FastHenry! Use Blender's powerful modelling capability combined with geometry nodes to run FastHenry simulations! Blender is becoming a multi-purpose tool not just a tool for 3D graphics and animation, but also for engineering and scientific work too. This video shows how you can calculate the impedance of wire bonds and leads of a Mosfet in a TO-220 package, you can see what effect bending the leads and adding a heatsink have on the impedance! hackaday Dave Jones Sybren A. Stüvel 🔶 🐈‍⬛ (sybren@fosstodon.org) #b3d #blender3d #electronics #geometrynodes #3dart #pcbdesign #transistors #Engineering #science #physics #Technology #kicad #simulation #manufacturing

Very excited to share the release of my addon Blender FastHenry! Use Blender's powerful modelling capability combined with geometry nodes to run FastHenry simulations! Blender is becoming a multi-purpose tool not just a tool for 3D graphics and animation, but also for engineering and scientific work too. This video shows how you can calculate the impedance of wire bonds and leads of a Mosfet in a TO-220 package, you can see what effect bending the leads and adding a heatsink have on the impedance! hackaday Dave Jones Sybren A. Stüvel 🔶 🐈‍⬛ ([email protected]) #b3d #blender3d #electronics #geometrynodes #3dart #pcbdesign #transistors #Engineering #science #physics #Technology #kicad #simulation #manufacturing

Sam M

45,882 views • 1 year ago

very excited to share what I've been working on, I've created an addon which allows you to run FastHenry directly in Blender! you can use the power of Blender to create any curve object and then call FastHenry to simulate the resistance and inductance. This demo video shows how you can calculate the resistance and inductance of a bond wire in a power module. You can experiement with different shapes and curvatures and see how the inductance changes. Furthermore you can control the radius with geomotry nodes! Also, you can combine this with ElectroMag Nodes to then visualise the magnetic fields! This addon will be a free update to ElectroMag Nodes #b3d #blender3d #geometrynodes #inductance #electronics #hardware #science #engineering #pcbdesign #educational #magnet #kicad #freecad #altium Maurizio Di Paolo Emilio hackaday

very excited to share what I've been working on, I've created an addon which allows you to run FastHenry directly in Blender! you can use the power of Blender to create any curve object and then call FastHenry to simulate the resistance and inductance. This demo video shows how you can calculate the resistance and inductance of a bond wire in a power module. You can experiement with different shapes and curvatures and see how the inductance changes. Furthermore you can control the radius with geomotry nodes! Also, you can combine this with ElectroMag Nodes to then visualise the magnetic fields! This addon will be a free update to ElectroMag Nodes #b3d #blender3d #geometrynodes #inductance #electronics #hardware #science #engineering #pcbdesign #educational #magnet #kicad #freecad #altium Maurizio Di Paolo Emilio hackaday

Sam M

58,782 views • 2 years ago

New update on combining FastHenry and Blender 🔶 , I've added the ability to calculate the mutual inductance between objects! Also developed a basic UI inside Blender to visualise the results and to highlight the inductively coupled objects. This video shows some examples of what can be donw now. First demo shows how you can model the impedance of die bond wires and the coupling to adjacent wires, you make adjustments to their lengths and curvatures and see the results in near-real-time. Second demo shows calculating inductances and mutual inductance of air-core inductors with arbitrary positions and deformations. Third demo shows simulating PCB coils over a range of frequencies for applications such as wireless power and distance sensing. And you can combine this with Electromag Nodes to visualise the magnetic fields! #b3d #blender3d #geometrynodes #electronics #hardware #science #technology #pcbdesign #kicad #altium #ansys #3dart #programmer #python #engineering #semiconductors #circuits #inductors #components

New update on combining FastHenry and Blender 🔶 , I've added the ability to calculate the mutual inductance between objects! Also developed a basic UI inside Blender to visualise the results and to highlight the inductively coupled objects. This video shows some examples of what can be donw now. First demo shows how you can model the impedance of die bond wires and the coupling to adjacent wires, you make adjustments to their lengths and curvatures and see the results in near-real-time. Second demo shows calculating inductances and mutual inductance of air-core inductors with arbitrary positions and deformations. Third demo shows simulating PCB coils over a range of frequencies for applications such as wireless power and distance sensing. And you can combine this with Electromag Nodes to visualise the magnetic fields! #b3d #blender3d #geometrynodes #electronics #hardware #science #technology #pcbdesign #kicad #altium #ansys #3dart #programmer #python #engineering #semiconductors #circuits #inductors #components

Sam M

35,035 views • 2 years ago

It has finally happened - my first video guide on how to create electromagnetic openEMS simulations with Blender 🔶 is now available! This video guide will show you to first create geometries and specify ports in #blender3d , export them to openEMS, write a few python scripts, and then using geometry nodes to visualise and animate the results back in Blender. The video is based on the Rotman lens animation I did previously as the workflow for it covers a wide range of topics. I hope you find this video informative. Everything is done using open-source tools including #inkscape, #paraview and Octave! #b3d #electronics #hardware #science #educational #physics #engineering #geometrynodes hackaday #openEMS #electromagnetics #tutorial #mathematics #antenna #rf #kicad #altium

It has finally happened - my first video guide on how to create electromagnetic openEMS simulations with Blender 🔶 is now available! This video guide will show you to first create geometries and specify ports in #blender3d , export them to openEMS, write a few python scripts, and then using geometry nodes to visualise and animate the results back in Blender. The video is based on the Rotman lens animation I did previously as the workflow for it covers a wide range of topics. I hope you find this video informative. Everything is done using open-source tools including #inkscape, #paraview and Octave! #b3d #electronics #hardware #science #educational #physics #engineering #geometrynodes hackaday #openEMS #electromagnetics #tutorial #mathematics #antenna #rf #kicad #altium

Sam M

24,221 views • 2 years ago

continuing with my PEEC solver in Blender for inductors. Now I have a basic Python/Numba code that can calculate skin and proximity. You can see as the frequency increases the current shifts towards to the outer surface and also gets affected by adjacent turns #b3d #geometrynodes #electronics #Technology #science

continuing with my PEEC solver in Blender for inductors. Now I have a basic Python/Numba code that can calculate skin and proximity. You can see as the frequency increases the current shifts towards to the outer surface and also gets affected by adjacent turns #b3d #geometrynodes #electronics #Technology #science

Sam M

14,696 views • 6 months ago

Very excited to share this #openEMS simulation of a 60GHz #mmwave radar PCB! You can see how the e-field component is generated and then propagates in the transmission lines towards the antenna array. The radiated electric field shows one main lobe at 0 deg, and smaller side lobes #b3d #radar #kicad #electronics #educational hackaday #blender3d #antenna #rf Altium KiCad PCB FreeCAD (CAD+CAM+CAE) Jacques Lucke Chris Gammell #Physics #Engineering #Science

Very excited to share this #openEMS simulation of a 60GHz #mmwave radar PCB! You can see how the e-field component is generated and then propagates in the transmission lines towards the antenna array. The radiated electric field shows one main lobe at 0 deg, and smaller side lobes #b3d #radar #kicad #electronics #educational hackaday #blender3d #antenna #rf Altium KiCad PCB FreeCAD (CAD+CAM+CAE) Jacques Lucke Chris Gammell #Physics #Engineering #Science

Sam M

83,718 views • 3 years ago

i don't know how many people have noticed this but beomgyu is very sensitive to the textures of food or objects and many times he tends to look for something that he likes to touch and plays with it for a while, i find it very cute and very relatable

i don't know how many people have noticed this but beomgyu is very sensitive to the textures of food or objects and many times he tends to look for something that he likes to touch and plays with it for a while, i find it very cute and very relatable

♡

87,393 views • 3 months ago

📢 Announcing one of the most exciting works from us this year on **scalable robot policy evaluation through real-to-sim transfer**, moving toward a scalable evaluation engine with structured world models that capture the appearance, geometry, and dynamics of environments involving deformable objects. 🤖 Evaluation remains one of the biggest bottlenecks in building general-purpose robots. Today, robots are still evaluated only in the real world, which is **orders of magnitude slower** than the development of language agents. We propose a new framework where simulation performance **strongly correlates** with the real world (r > 0.9), even for deformable objects. The key difference from existing work lies in the correlation between simulation and reality: if a robot model performs better in the digital world, does it also perform better in the real world? This question has long made people hesitant about simulation-based evaluation — especially for deformable objects. We are changing that. Our pipeline achieves effective real-to-sim transfer, establishing **state-of-the-art correlation** between simulation and reality for deformable object manipulation. It provides a **scalable and reproducible evaluation engine** for robot learning. 🌐

📢 Announcing one of the most exciting works from us this year on scalable robot policy evaluation through real-to-sim transfer, moving toward a scalable evaluation engine with structured world models that capture the appearance, geometry, and dynamics of environments involving deformable objects. 🤖 Evaluation remains one of the biggest bottlenecks in building general-purpose robots. Today, robots are still evaluated only in the real world, which is orders of magnitude slower than the development of language agents. We propose a new framework where simulation performance strongly correlates with the real world (r > 0.9), even for deformable objects. The key difference from existing work lies in the correlation between simulation and reality: if a robot model performs better in the digital world, does it also perform better in the real world? This question has long made people hesitant about simulation-based evaluation — especially for deformable objects. We are changing that. Our pipeline achieves effective real-to-sim transfer, establishing state-of-the-art correlation between simulation and reality for deformable object manipulation. It provides a scalable and reproducible evaluation engine for robot learning. 🌐

Yunzhu Li

39,960 views • 9 months ago

Did you know that in Blender you can easily animate oscillating objects with a single driver expression? ✨ 1. Select your object 2. In its rotation write: "# sin(frame / 300 * pi * 2 * 3) * 0.5" where 3 is the number of swings it does, and 0.5 is the amplitude multiplier (use a bigger number for bigger swings), and 300 is your end frame 3. Enjoy the perfect loop! #b3d

Did you know that in Blender you can easily animate oscillating objects with a single driver expression? ✨ 1. Select your object 2. In its rotation write: "# sin(frame / 300 * pi * 2 * 3) * 0.5" where 3 is the number of swings it does, and 0.5 is the amplitude multiplier (use a bigger number for bigger swings), and 300 is your end frame 3. Enjoy the perfect loop! #b3d

passivestar

296,083 views • 3 years ago

🔥Sakana Fugu-ultra 🟩OpenAI GPT 5.5 🟩GLM 5.2 🟩Opus 4.8 TASK Create a high-quality single HTML file simulation of a Rube Goldberg / chain-reaction machine. The simulation should run automatically from start to finish without user input.The goal is to demonstrate whether the coding model understands real-world physics well enough to create believable cause-and-effect behavior - Rube Goldberg / chain-reaction machine. The scene should show a small physical world where objects interact through realistic mechanics: gravity, collisions, momentum transfer, ramps, pulleys, levers, springs, falling objects, rolling balls, dominoes, and water or particles if possible. The simulation should include: * A ball rolling down a ramp * Dominoes falling one after another * A lever or seesaw transferring force * A spring launcher * A pendulum or swinging weight * A pulley or elevator mechanism * Falling objects affected by gravity * Objects with different mass, friction, and bounce * Clear visual labels explaining what physical law is being demonstrated * A progress timeline showing the current stage of the chain reaction * Automatic reset/replay after the sequence ends The simulation should visually demonstrate: * Gravity * Momentum transfer * Conservation of energy * Friction * Torque * Elastic potential energy * Collision response * Cause and effect The model should not just animate objects on fixed paths. Objects should appear to interact physically. For example, the ball should knock over dominoes, dominoes should push a lever, the lever should launch another object, and so on. Evaluation criteria 1. Does the simulation run automatically without user input? 2. Do objects interact through believable physical cause and effect? 3. Are physical laws represented correctly? 4. Are labels and UI helpful? 5. Is the animation smooth and stable? 6. Does the code avoid fake-looking hardcoded movement? 7. Does the scene reset cleanly?

🔥Sakana Fugu-ultra 🟩OpenAI GPT 5.5 🟩GLM 5.2 🟩Opus 4.8 TASK Create a high-quality single HTML file simulation of a Rube Goldberg / chain-reaction machine. The simulation should run automatically from start to finish without user input.The goal is to demonstrate whether the coding model understands real-world physics well enough to create believable cause-and-effect behavior - Rube Goldberg / chain-reaction machine. The scene should show a small physical world where objects interact through realistic mechanics: gravity, collisions, momentum transfer, ramps, pulleys, levers, springs, falling objects, rolling balls, dominoes, and water or particles if possible. The simulation should include: * A ball rolling down a ramp * Dominoes falling one after another * A lever or seesaw transferring force * A spring launcher * A pendulum or swinging weight * A pulley or elevator mechanism * Falling objects affected by gravity * Objects with different mass, friction, and bounce * Clear visual labels explaining what physical law is being demonstrated * A progress timeline showing the current stage of the chain reaction * Automatic reset/replay after the sequence ends The simulation should visually demonstrate: * Gravity * Momentum transfer * Conservation of energy * Friction * Torque * Elastic potential energy * Collision response * Cause and effect The model should not just animate objects on fixed paths. Objects should appear to interact physically. For example, the ball should knock over dominoes, dominoes should push a lever, the lever should launch another object, and so on. Evaluation criteria 1. Does the simulation run automatically without user input? 2. Do objects interact through believable physical cause and effect? 3. Are physical laws represented correctly? 4. Are labels and UI helpful? 5. Is the animation smooth and stable? 6. Does the code avoid fake-looking hardcoded movement? 7. Does the scene reset cleanly?

Remek Kinas

14,146 views • 1 month ago

I think I've written the world's fastest Blender exporter. It exports full levels in 9 milliseconds. Check out this video where I show live editing of levels from blender, and it updates instantly in the game. For some context: Okay, since the beginning of development on this game project, I knew I didn't want to write a custom level editor for it and rather spend my time programming the game. I decided to use blender as the level editor, and export directly from blender to my game engine's file format. Blender only lets you write add-ons in Python, which is not just incredibly slow but also a terrible platform for this kind of systems programming work. So, after having a Python exporter for the longest time, I decided enough was enough, and I programmed this exporter in C. Blender doesn't expose any C API or interface. So my only option was to download Blender's source code and insert my code directly into Blender. And that's what I did :) So this new C exporter not only is orders of magnitude faster than the python version, but also is way more robust. And because I'm in C land, my hands are much freer to do things. There are two elements of secret sauce that make this exporter run so fast: - A custom memory arena implementation. - Lockless multithreading.

I think I've written the world's fastest Blender exporter. It exports full levels in 9 milliseconds. Check out this video where I show live editing of levels from blender, and it updates instantly in the game. For some context: Okay, since the beginning of development on this game project, I knew I didn't want to write a custom level editor for it and rather spend my time programming the game. I decided to use blender as the level editor, and export directly from blender to my game engine's file format. Blender only lets you write add-ons in Python, which is not just incredibly slow but also a terrible platform for this kind of systems programming work. So, after having a Python exporter for the longest time, I decided enough was enough, and I programmed this exporter in C. Blender doesn't expose any C API or interface. So my only option was to download Blender's source code and insert my code directly into Blender. And that's what I did :) So this new C exporter not only is orders of magnitude faster than the python version, but also is way more robust. And because I'm in C land, my hands are much freer to do things. There are two elements of secret sauce that make this exporter run so fast: - A custom memory arena implementation. - Lockless multithreading.

Taha Torabpour

205,717 views • 11 months ago

Blender tutorials are about to age very badly. In this demo, Kimi K3 gets two character references and a plain-English request. Minutes later, it builds the first version inside Blender. The user asks for a better horn, Kimi edits the same scene, and the character starts moving. This is not another AI tool generating one lucky image. It is working with an actual editable 3D project, then changing the geometry after feedback instead of restarting from zero. It still needs human direction, because the first result is rarely the good one. But blockouts, repetitive edits, Blender Python, and basic iteration are becoming a conversation rather than hours of clicking through menus. Kimi K3 + Blender MCP is basically turning 3D modeling into: describe it, inspect it, complain about it, and let the AI fix it.

Blender tutorials are about to age very badly. In this demo, Kimi K3 gets two character references and a plain-English request. Minutes later, it builds the first version inside Blender. The user asks for a better horn, Kimi edits the same scene, and the character starts moving. This is not another AI tool generating one lucky image. It is working with an actual editable 3D project, then changing the geometry after feedback instead of restarting from zero. It still needs human direction, because the first result is rarely the good one. But blockouts, repetitive edits, Blender Python, and basic iteration are becoming a conversation rather than hours of clicking through menus. Kimi K3 + Blender MCP is basically turning 3D modeling into: describe it, inspect it, complain about it, and let the AI fix it.

Rina

18,920 views • 16 days ago

Benchmarking, evaluating, and developing robotics code is difficult, and part of this is because no simulator really reflects the diversity and scale of real embodiments. Enter MolmoSpaces from AI2: a massive open ecosystem with a range of 230,000 handcrafted and procedurally-generated home environments, including 48,000 manipulable objects. Crucially, MolmoSpaces provides simulation environments which work for both navigation and manipulation. We talked to the team: Yejin Kim @ RSS2026, Omar Rayyan, and Max Argus, to tell us more. Watch Episode 69 of RoboPapers, with Michael Cho - Rbt/Acc and Jiafei Duan, now!

Benchmarking, evaluating, and developing robotics code is difficult, and part of this is because no simulator really reflects the diversity and scale of real embodiments. Enter MolmoSpaces from AI2: a massive open ecosystem with a range of 230,000 handcrafted and procedurally-generated home environments, including 48,000 manipulable objects. Crucially, MolmoSpaces provides simulation environments which work for both navigation and manipulation. We talked to the team: Yejin Kim @ RSS2026, Omar Rayyan, and Max Argus, to tell us more. Watch Episode 69 of RoboPapers, with Michael Cho - Rbt/Acc and Jiafei Duan, now!

RoboPapers

21,116 views • 4 months ago

I'm not a 3d modeler. I barely know how to open blender let alone do anything functional in it. But with the siddharth ahuja blender MCP and Claude I was able to vibe code an entire environment for my upcoming grow a garden game in Portals Claude used the MCP to assemble and arrange every single asset in this space (i literally don't even know how to import an asset into blender lol) within the spawn of a few hours and with access to a folder of low poly assets. When I had it start doing its own modeling, it even matched the style of the environment to keep things consistent If you are building UGC games and you don't know anything about modeling or design like me, you gotta try it out...

I'm not a 3d modeler. I barely know how to open blender let alone do anything functional in it. But with the siddharth ahuja blender MCP and Claude I was able to vibe code an entire environment for my upcoming grow a garden game in Portals Claude used the MCP to assemble and arrange every single asset in this space (i literally don't even know how to import an asset into blender lol) within the spawn of a few hours and with access to a folder of low poly assets. When I had it start doing its own modeling, it even matched the style of the environment to keep things consistent If you are building UGC games and you don't know anything about modeling or design like me, you gotta try it out...

Hugh MFER (PORTALS Guy)

180,520 views • 6 months ago

When we run out of good training data in reality, simulation is the next gold mine. Enters Infinigen: an open-source, procedurally generated, photorealistic dataset for 3D vision. The quality is stunning! No two worlds are the same. ▸ Every little detail is randomized and customizable, even the wrinkles on a flower petal! ▸ Diverse objects & scenes in the natural world: plants, animals, terrains; fire, cloud, rain, and snow. ▸ Groundtruth automatic annotations: optical flow, 3D scene flow, depth, surface normals, panoptic segmentation, occlusion boundaries. ▸ Written in Blender. Made by Jia Deng's lab at Princeton: ▸ Website: ▸ Code: ▸ Paper:

When we run out of good training data in reality, simulation is the next gold mine. Enters Infinigen: an open-source, procedurally generated, photorealistic dataset for 3D vision. The quality is stunning! No two worlds are the same. ▸ Every little detail is randomized and customizable, even the wrinkles on a flower petal! ▸ Diverse objects & scenes in the natural world: plants, animals, terrains; fire, cloud, rain, and snow. ▸ Groundtruth automatic annotations: optical flow, 3D scene flow, depth, surface normals, panoptic segmentation, occlusion boundaries. ▸ Written in Blender. Made by Jia Deng's lab at Princeton: ▸ Website: ▸ Code: ▸ Paper:

Jim Fan

363,656 views • 3 years ago

Open science is how we continue to push technology forward and today at Meta FAIR we’re sharing eight new AI research artifacts including new models, datasets and code to inspire innovation in the community. More in the video from Joelle Pineau. This work is another important step towards our goal of achieving Advanced Machine Intelligence (AMI). What we’re releasing: • Meta Spirit LM: An open source language model for seamless speech and text integration. • Meta Segment Anything Model 2.1: An updated checkpoint with improved results on visually similar objects, small objects and occlusion handling. Plus a new developer suite to make it easier for developers to build with SAM 2. • Layer Skip: Inference code and fine-tuned checkpoints demonstrating a new method for enhancing LLM performance. • SALSA: New code to enable researchers to benchmark AI-based attacks in support of validating security for post-quantum cryptography. • Meta Lingua: A lightweight and self-contained codebase designed to train language models at scale. • Meta Open Materials: New open source models and the largest dataset of its kind to accelerate AI-driven discovery of new inorganic materials. • MEXMA: A new research paper and code for our novel pre-trained cross-lingual sentence encoder with coverage across 80 languages. • Self-Taught Evaluator: a new method for generating synthetic preference data to train reward models without relying on human annotations. Access to state-of-the-art AI creates opportunities for everyone. We’re excited to share this work and look forward to seeing the community innovation that results from it. Details and access to everything released by FAIR today ➡️

Open science is how we continue to push technology forward and today at Meta FAIR we’re sharing eight new AI research artifacts including new models, datasets and code to inspire innovation in the community. More in the video from Joelle Pineau. This work is another important step towards our goal of achieving Advanced Machine Intelligence (AMI). What we’re releasing: • Meta Spirit LM: An open source language model for seamless speech and text integration. • Meta Segment Anything Model 2.1: An updated checkpoint with improved results on visually similar objects, small objects and occlusion handling. Plus a new developer suite to make it easier for developers to build with SAM 2. • Layer Skip: Inference code and fine-tuned checkpoints demonstrating a new method for enhancing LLM performance. • SALSA: New code to enable researchers to benchmark AI-based attacks in support of validating security for post-quantum cryptography. • Meta Lingua: A lightweight and self-contained codebase designed to train language models at scale. • Meta Open Materials: New open source models and the largest dataset of its kind to accelerate AI-driven discovery of new inorganic materials. • MEXMA: A new research paper and code for our novel pre-trained cross-lingual sentence encoder with coverage across 80 languages. • Self-Taught Evaluator: a new method for generating synthetic preference data to train reward models without relying on human annotations. Access to state-of-the-art AI creates opportunities for everyone. We’re excited to share this work and look forward to seeing the community innovation that results from it. Details and access to everything released by FAIR today ➡️

AI at Meta

150,222 views • 1 year ago