memossy

15 ms·

1.
by memossy 3y ago
It was interesting Aurora used GPU Max & definitely looking forward to Falcon Shores. I think Gaudi2 was bad timed & they had to build stack, Gaudi3 is where I think we will see mass adoption given availability, way cheaper price&#x
2.
by memossy 3y ago
It's a great company & will do well, plenty of demand & B100s/BH200s etc coming The Hopper stuff is particulalry interesting
3.
by memossy 3y ago
Falcon shores next year will be crazy with 300gb VRAM & new lith
4.
by memossy 3y ago
Gaudi2s started coming out in 2022 ( https://huggingface.co/blog/habana-gaudi-2-benchmark ) but didn't hit mass scale. I think Gaudi3 will & others have seen similar performance for Gaudi2 eg Databricks: https:
5.
by memossy 3y ago
We use v4s, v5es & v5ps. Mostly v5ps, very stable int8 training (versus the horror that is fp8 stability)
6.
by memossy 3y ago
I mean they work well, here is another blog by Databricks: https://www.databricks.com/blog/llm-training-and-inference-i...
7.
by memossy 3y ago
I think as we go to enterprise workloads the total cost of ownership becomes important. NVIDIA is still the best for research given ecosystem but once the models are standardised as with transformers/LLaMA and likely multimodal diffusi
8.
by memossy 3y ago
It took less than a day to port our code over, we do custom CUDA across modalities. Gaudi2 was actually announced 2 years ago and is 7nm like the A100 80Gb it was meant to be competitive with, Gaudi3 later this year is probably going to be
9.
by memossy 3y ago
The v5es and v5ps are pretty amazing at running SD, giving code for SD3 now to optimise it on those. v5es are particularly interesting given the millions that will land and the large pod sizes, particularly well constructed for million toke
10.
by memossy 3y ago
Think it'll probably crack on with Gaudi3 at 4x performance, twice VRAM etc later this year. We found cuda sycl conversion surprisingly good https://www.intel.com/content/www/us/en/developer/art
11.
Intel Gaudi2 chips outperform Nvidia H100 on diffusion transformers(stability.ai)
146 points by memossy 3y ago|60 comments
12.
by memossy 3y ago
"For Stable Diffusion 3, we measured the training throughput for the 2B Multimodal Diffusion Transformer (MMDiT) architecture model. Gaudi 2 trained images 1.5x faster than the H100-80GB, and 3x faster than A100-80GB GPU’s when scaled
13.
by memossy 3y ago
I could but I won't as legal stuff :)
14.
by memossy 3y ago
I mean open models yo
15.
by memossy 3y ago
If we trained it with videos yes but need more GPUs for that.
16.
by memossy 3y ago
It'll be out soon, doing benchmark tests etc
17.
by memossy 3y ago
We have highly efficient models for inference and a quantization team. Need moar GPUs to do a video version of this model similar to Sora now they have proved that Diffusion Transformers can scale with latent patches (see stablevideo.com an
18.
by memossy 3y ago
As the leader in open image models it is incumbent upon us as the models get to this level of quality to take seriously how we can release open and safe models from a legal, societal and other considerations. Not engaging in this will indee
19.
by memossy 3y ago
800m is good for mobile, 8b for graphics cards. Bigger than that is also possible, not saturated yet but need more GPUs.
20.
by memossy 3y ago
We did this for every stable diffusion release, you get the feedback data to improve it continuously ahead of open release.
21.
by memossy 3y ago
Training on 4096 v5es how did you handle crazy batch size :o
22.
by memossy 5y ago
You should go web3. There are incredibly few lawyers who know anything about crypto and huge demand. You can also join DAOs etc and earn significant amounts of money for your effort versus who you are. If you want to do web3 code go to htt
23.
by memossy 5y ago
"Payment can be accepted in any currency, including cryptocurrencies as we outrun the end of civilisation."
24.
by memossy 6y ago
I was helped with my sinus pressure and dizziness by 2 x 1g N Acetyl Cysteine per day. Very short half life, very quick response (if any), available on Amazon.
25.
by memossy 6y ago
You may wish to look into Clemastine/Tavegil antihistamine, easy to get over the counter.
26.
by memossy 6y ago
I would trust Tether more if it ever had meaningful outflows in dollar terms, had $1bn or so in the last crash ( https://coinmarketcap.com/currencies/tether/ ) but has been generally on an upward march, not what you
27.
by memossy 6y ago
It's important to see if you are before or after the cliff. If before then depending on your employment agreement and other docs there could be a scenario where you are fired/let go and get 0% shares. Your last round valuation was
28.
An AI Girlfriend Seducing China’s Lonely Men(sixthtone.com)
133 points by memossy 6y ago|195 comments
29.
by memossy 6y ago
Lots of promise and a nice front end, but I'm struggling to see where after three years an example of high value knowledge in this platform is. For example on the top topic AI most of the sections are unfilled / or lightweight asi
30.
by memossy 6y ago
I will try to get back to this thread with some resources but I would recommend as you are in the UK to look into clemastine/tavegil which you can get without prescription (noted in the article as about to be trialled with metformin) d

More ›