HackerTrans
TopNewTrendsCommentsPastAskShowJobs

ioedward

no profile record

comments

ioedward
·hace 3 años·discuss
AMD didn't submit results, despite geohot/Lisa Su's plan to get AMD on MLPerf: https://twitter.com/realGeorgeHotz/status/166980346408248934...
ioedward
·hace 3 años·discuss
8 H100s would have enough VRAM to finetune a 70B model.
ioedward
·hace 3 años·discuss
Not at all, PyTorch uses sub-classing (nn.Module) and other Pythonic features heavily that are not supported with Mojo.
ioedward
·hace 3 años·discuss
Trying the fuse feature, seemed the most interesting:

> spaghetti ≈ trumpet

> Both spaghetti and a trumpet can be difficult to eat without making a mess—spaghetti with its long, slippery noodles, and a trumpet with its wide, flared bell.

ok google
ioedward
·hace 3 años·discuss
Related: 1 year represented in a photo. https://eirikso.com/2011/01/04/one-year-in-one-image/
ioedward
·hace 3 años·discuss
The technical term is GPGPU, general-purpose computing on GPUs, but I like to call them GPUs for short.
ioedward
·hace 3 años·discuss
Nvidia's enterprise GPUs are surprisingly unreliable. Working on a 128 GPU A100 cluster on AWS, 1 would fail every few days. I didn't have any insight on whether it was a hardware or software failure.
ioedward
·hace 3 años·discuss
You also need the optimizer (e.g. Adam)'s state, which is usually double the parameter's size. So if using fp16, one parameter takes up 6 bytes in memory.
ioedward
·hace 3 años·discuss
TPUs are mostly hoarded by Google Research (including Deepmind) and Ads. Very few are being used by external people.
ioedward
·hace 3 años·discuss
The twitter thread is worth reading, Yann responds to some of the questions raised here.
ioedward
·hace 3 años·discuss
No it won't. Large language models are trained on 1,000 - 50,000 GPUs. No one's going to buy hundreds of Mac pros to mount them in a datacenter for training ML models.
ioedward
·hace 3 años·discuss
Normally people split up the model across multiple GPUs, i.e. model/tensor parallelism.
ioedward
·hace 3 años·discuss
Doesn't seem to be affiliated with OpenAI. Domain was registered in 2017, 2 years after OpenAI was created. Trolling/domain squatting?
ioedward
·hace 3 años·discuss
> I understand that PyTorch is an awesome tool for researchers, but it doesn't necessarily fit into a prod environment.

Some of the largest deployments of ML are using PyTorch models, e.g. OpenAI, Meta, Microsoft.
ioedward
·hace 3 años·discuss
Where did you buy the Instinct hardware and what's a good price?