[HN Gopher] Qwen3-VL can scan two-hour videos and pinpoint nearl...
       ___________________________________________________________________
        
       Qwen3-VL can scan two-hour videos and pinpoint nearly every detail
        
       Author : thm
       Score  : 54 points
       Date   : 2025-11-30 07:27 UTC (2 days ago)
        
 (HTM) web link (the-decoder.com)
 (TXT) w3m dump (the-decoder.com)
        
       | thot_experiment wrote:
       | anyone have a tl;dr for me on what the best way to get the video
       | comprehension stuff going is? i use qwen-30b-vl all the time
       | locally as my goto model because it's just so insanely fast,
       | curious to mess with the video stuff, the vision comprehension
       | works great and i use it for OCR and classification all the time
        
         | xrd wrote:
         | How much VRAM do you need for local usage may I ask?
        
       | moralestapia wrote:
       | To me, this qualifies as some sort ASI already.
        
       | visioninmyblood wrote:
       | I was using this for video understanding with inference form
       | vlm.run infra. It definitely has outperformed Gemini which
       | generally is much better than openai or Claude on videos. The
       | detailed extraction is pretty good. With agents you can also crop
       | into a segment and do more operations on it. have to see how the
       | multi modal space progresses:
       | 
       | link to results:
       | https://chat.vlm.run/c/82a33ebb-65f9-40f3-9691-bc674ef28b52
       | 
       | Quick demo: https://www.youtube.com/watch?v=78ErDBuqBEo
        
       | djmips wrote:
       | Does anyone else worry about this technology used for Big Brother
       | type surveillance?
        
         | reactordev wrote:
         | Where have you been the last decade? It's already in use, or
         | models like it, by companies selling access to _The State_
         | 
         | https://deflock.me
         | 
         | Not to mention cloud platforms that collect evidence and
         | process it with _all the models_ and store that information for
         | searching...
         | 
         | https://www.revir.ai
        
           | mptest wrote:
           | or if you prefer your depression in book format: surveillance
           | capitalism by zuboff pegasus: a spy in your pocket laurent
           | richard
        
           | eurekin wrote:
           | No mention of palantir?
        
       | spwa4 wrote:
       | It's so weird how that works with transformers.
       | 
       | Finetuning an LLM "backbone" (if I understand correctly: a fully
       | trained but not instruction tuned LLM, usually small because
       | students) with OCR tokens bests just about every OCR network out
       | there.
       | 
       | And it's not just OCR. Describing images. Bounding boxes. Audio,
       | both ASR and TTS, all works better that way. Now many research
       | papers are only really about how to encode image/audio/video to
       | feed it into a Llama or Qwen model.
        
         | zmmmmm wrote:
         | It is fascinating. Vision language models are unreasonably good
         | compared to dedicated OCR and even the language tasks to some
         | extent.
         | 
         | My take is it fits into the general concept that generalist
         | models have significant advantages because so much more latent
         | structure maps across domains than we expect. People still talk
         | about fine tuning dedicated models being effective but my
         | personal experience is it's still always better to use a larger
         | generalist model than a smaller fine tuned one.
        
       | eurekin wrote:
       | Insane if true... now I wonder, if I use it to go through some
       | old dance routing video catalogue to recognize and write
       | individual move lists
        
       ___________________________________________________________________
       (page generated 2025-12-02 23:00 UTC)