Raman R(@vramanrs) 's Twitter Profileg
Raman R

@vramanrs

SE

ID:1295250940316942336

calendar_today17-08-2020 06:47:52

232 Tweets

40 Followers

785 Following

Aran Komatsuzaki(@arankomatsuzaki) 's Twitter Profile Photo

Looking Backward: Streaming Video-to-Video Translation with Feature Banks

Runs 20 FPS on one A100 GPU, being 15×, 46×, 108×, and 158× faster than FlowVid, CoDeF, Rerender, and TokenFlow, respectively

proj: jeff-liangf.github.io/projects/strea…
abs: arxiv.org/abs/2405.15757

account_circle
Akshay Narisetti(@AkshayNarisetti) 's Twitter Profile Photo

My biggest wins in college:

- Do an internship every semester, all of them off-campus

- Lead two top deep technical clubs on campus

account_circle
Akshat Shrivastava(@Akshat_World) 's Twitter Profile Photo

In a few days, we will have India's NEXT Prime Minister.

After grand celebrations, we will go back to our usual life:-

- Life struggles would go on.
- Office timings won't change.
- EMIs won't be waived off (unless of course you are well connected)

Needless to say that

account_circle
Deedy(@deedydas) 's Twitter Profile Photo

And if you believe this still isn't perfect or looks terrible, always remember Midjourney's progression in <2yrs

And if you believe this still isn't perfect or looks terrible, always remember Midjourney's progression in <2yrs
account_circle
Farza 🇵🇰🇺🇸(@FarzaTV) 's Twitter Profile Photo

I feel like building a great product is less about brilliance and feels 100X more like you're playing Age of Empires 2 where you just gotta survive in the fog of war long enough to make it to your village's next evolution upgrade.

account_circle
Aman Goel(@amangoeliitb) 's Twitter Profile Photo

Because:

1. Not everyone wants to be an entrepreneur.
2. Not everyone wants to only focus on making money. Some people want a job that they enjoy doing.
3. Not everyone likes managing other people. Some people enjoy individual deep work.
4. Not everyone wants to take risks.

account_circle
Morbid Knowledge(@Morbidful) 's Twitter Profile Photo

Philip Gale was an American Internet software developer and a sophomore at MIT.

He gained recognition for developing Total Access while at MIT and was hired by EarthLink at 16 to further its development.

By 17, Gale had earned approximately a million dollars.

In 1995, Gale's

Philip Gale was an American Internet software developer and a sophomore at MIT. He gained recognition for developing Total Access while at MIT and was hired by EarthLink at 16 to further its development. By 17, Gale had earned approximately a million dollars. In 1995, Gale's
account_circle
Adam Singer(@AdamSinger) 's Twitter Profile Photo

Remote work firm NVIDIA now 3rd largest company globally by market cap (worth >$2.3 trillion) powering the future of computing, AI, gaming, crypto; while your boss is still living in the 90s & demands you commute 2 hours/day to sign TPS reports. At this point it's just comical

account_circle
Kalpit Veerwal(@kalpitveerwal) 's Twitter Profile Photo

Remote work is a blessing, specially for those just starting their career. Working from home can save a huge portion of your income which would otherwise be spent on rent, commute, groceries, etc. Not to mention, you also get family support. What are your thoughts on remote work?

account_circle
Alberto Hojel(@AlbyHojel) 's Twitter Profile Photo

Tim Darcet was looking through DiNOv2 (a self supervised vision model used for deep feature extraction) and saw something strange happening...

A dive into one of the recent ICLR outstanding paper award recipients: 'Vision Transformers Need Registers' 🧵

account_circle
Bindu Reddy(@bindureddy) 's Twitter Profile Photo

Big Models Consume Tons Of Power - ChatGPT Requires 10x That Of Google Search

These giant LLMs require massive amounts of power to train and infer. We don't discuss this as much as we should.

This is one of the reasons why small models are so important.

h/t zerohedge

Big Models Consume Tons Of Power - ChatGPT Requires 10x That Of Google Search These giant LLMs require massive amounts of power to train and infer. We don't discuss this as much as we should. This is one of the reasons why small models are so important. h/t zerohedge
account_circle
Gowthami Somepalli(@gowthami_s) 's Twitter Profile Photo

Why is self-correction not a big thing in LLM research? This is one of the biggest criticisms of the auto-regressive modeling paradigm!

When I googled, I see only 2 papers somewhat related to this! Please drop in a comment with any relevant papers/blogs!

Paper 1:

account_circle
Akshat Shrivastava(@Akshat_World) 's Twitter Profile Photo

This is Aswini and Anish, two IT professionals who were returning home in Pune on May 19 when an unregistered Porsche, driving at 150 km/h in Kalyani Nagar, hit them. One of them died on the spot, and the other died during medical treatment.

The driver of the car was a

This is Aswini and Anish, two IT professionals who were returning home in Pune on May 19 when an unregistered Porsche, driving at 150 km/h in Kalyani Nagar, hit them. One of them died on the spot, and the other died during medical treatment. The driver of the car was a
account_circle
Akshat Shrivastava(@Akshat_World) 's Twitter Profile Photo

Aditya Vikram If you are young. And, live in a la-la land, reality will hit you hard.

At that point: you will have no options.

It is far better to be pessimistic. And, build your options.

Rather than believing in a Fairy-Tail story that is being carefully crafted (and is false)

account_circle
Logan Kilpatrick(@OfficialLoganK) 's Twitter Profile Photo

People are still underestimating the value of Gemini 1.5 Flash.

For $0.35, you can get 1 million tokens and start building natively multi-modal projects.

The cost + latency + context window size + intelligence of Flash is going to create so many new startups.

account_circle