Having redundancy isn’t some unusual thing it’s the standard thing. No reason to have a huge proportion of a network reliant on a single chassis let alone a single line card. Most of the rise that does exist could also be explained by customers of those networks being unable to contact those on Three’s service.ĭecemat 1:57 No ISP worth their salt will have any single point of failure outside of the very edge where the customer connects so a failure there will impact a limited number of people. Some posters have also reported that O2, Vodafone and EE were impacted by the same issue, but their complaint volumes are still relatively low (measured in the hundreds, rather than the tens of thousands) – not even remotely in the same ballpark as Three experienced. But we won’t know if there’s still a major problem this morning until everybody starts waking up to use their mobile phones. Three appears to have restored data connectivity last night via various unspecified network mitigation measures, while some of those experiencing calling issues have informed us that they seemed to have been resolved in the small hours of this morning. They are working to replace this device with delivery to site anticipated by 19:00 and are reviewing what additional mitigation they can put in place to minimise disruption.” According to that page: “ 3UK have identified a faulty device in their network causing intermittent disruption to service. Meanwhile, Cloudflare has released a chart that shows how the outage has been recorded on their internet traffic trends for the operator’s network, which highlights a clear drop around 15:00 UTC and it hasn’t returned, yet.Īt the same time Gamma’s service status page may give us our first real indication of the cause. If you're into this field or just curious, I'd love to get your feedback.We’re seeing some reports of connections returning. While it's still a work in progress and not exactly the Vox style I envisioned, I genuinely hope this approach of storytelling simplifies Deep Learning for many. The result? Not only better image analysis but also a more efficient use of computing resources.Ĭoming back to the video. This ensures tailored analysis for different object sizes and details, veering away from the "one size fits all" concept. Think of it as using different magnifying glasses for various parts of a picture. In essence, DiT introduces a fascinating approach that tailors how each section of an image is processed, based on its content. The video dives into a recent paper I read - "DiT: Efficient Vision Transformers with Dynamic Token Routing." And yes, this idea had been living rent-free in my mind for way too long! □ So, I finally took the plunge and tried my hand at it. You know how Vox-style videos simplify complex topics so brilliantly? I've always dreamt of having something similar for Deep Learning topics. I have never seen the ML Community of Kolkata so vibrant, punctual, and eager.Įastern India’s largest ML Conference and rightfully so!Īlso if anyone is interested, here are the slides which me and Aritra Roy Gosthipaty presented: #keras #kerascommunitydays #gde #googledevelopergroups #tensorflowusergroup #kolkata #machinelearning Rishiraj Acharya has outdone himself, along with his co-organizer Anubhav Singh, and a team of excellent and adept volunteers who have all done a fantastic job. Through the day and the weeks that preceded one person's dedication and drive to bring this fantastic event to fruition has been clearly visible! The event not only boasts a fantastic lineup of speakers but also is one of the most thoughtfully organized and well-managed events I have been to. Keras Community Days Kolkata organised by TensorFlow User Group (TFUG) Kolkata probably ticks all these boxes and does more. What makes a tech community successful? The number of schwags at their events? The quality of food served? The quality of the talks and the technical prowess of the speakers? The curiosity and the eagerness of the audience to attend and soak the sessions?
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |