[HN Gopher] GPUDrive: Data-driven, multi-agent driving simulatio...
       ___________________________________________________________________
        
       GPUDrive: Data-driven, multi-agent driving simulation at 1M FPS
        
       Author : jonbaer
       Score  : 90 points
       Date   : 2024-08-08 20:40 UTC (1 days ago)
        
 (HTM) web link (arxiv.org)
 (TXT) w3m dump (arxiv.org)
        
       | cs702 wrote:
       | Unless I'm missing something big, this looks like a significant
       | deal for independent developers of self-driving AI software:
       | GPUDrive enables them to run driving simulations with hundreds of
       | AI agents on consumer-grade GPUs at 1M FPS, and it comes with
       | Python bindings, wrappers for Pytorch and Jax, and a friendly
       | standard MIT license. Thank you for sharing this on HN!
        
         | ipsum2 wrote:
         | It's not particularly useful since the simulation is too high
         | level. Good for games or homework projects, but not anything
         | real world.
        
         | nivertech wrote:
         | Unfortunately, 1M FPS is not for the simulated front camera
         | view of the road, it's just an overhead view of the map of the
         | roads/tracks.
        
           | jaquers wrote:
           | I am not an expert, but the way that I understand self-
           | driving systems is that there are multiple models running,
           | and then those outputs are fused into yet another model which
           | outputs the raw controls/actuations. In other words, I see
           | this model/trainer as the "conductor", telling the car how it
           | should approach an intersection, enter a highway, deal with
           | merging traffic or construction zones, etc.
           | 
           | There is another model which interprets visual data to assist
           | with lane-keeping, slow down or stop for pedestrians, inform
           | the conductor of road signs... The final model combines all
           | these inputs and incorporates the user preferences and then
           | decides whether to brake or accelerate, how much to rotate
           | the steering wheel.
           | 
           | Idk heh. The point of the high performance training is you
           | can train the "conductor" role faster, and run inference
           | faster. Assuming the car has limited compute/gpu resources,
           | if you have a very high performance conductor function, you
           | can dedicate that much more budget to visual/sensor inference
           | and or any other models like the Trolley Problem decider
           | (jk).
           | 
           | edit: grammar/details
        
       | BetterWhisper wrote:
       | https://github.com/Emerge-Lab/gpudrive - the repo
        
       | foota wrote:
       | Is this just the location data being trained on, or is there
       | image and sensor input data too? It looks like it's just
       | location, which seems like it limits the applicability, but I'm
       | not sur
       | 
       | Edit: reading a bit more it's somewhere in between. Afaict no raw
       | sensor data etc.,. but different "parsed" sensor inputs are
       | supported. I'm not sure whether this is synthetic or not? E.g.,
       | is the LIDAR view real LIDAR data from some system or a processed
       | result of what the system thinks LIDAR would be able to see? I
       | can't tell.
        
       | toppy wrote:
       | I don't know this field of research thus my question: why such a
       | high framerate is consider as a feature at all? Does it help with
       | learning rate?
        
         | Wingy wrote:
         | If you have a simulation where realtime is 60 fps, you could
         | simulate a little over 4.5 hours per second if you could run it
         | at 1M fps. That would definitely help with learning rate.
        
           | ramon156 wrote:
           | Feels like less [?] more. If you want to fill in these gaps
           | you might aswell interpolate
        
             | mlyle wrote:
             | He's not saying "break realtime into microsecond chunks."
             | 
             | He's saying: run through 4.5 hours of 16 millisecond chunks
             | of time in a second. This is good for regression testing or
             | producing training data quickly.
        
       ___________________________________________________________________
       (page generated 2024-08-09 23:02 UTC)