파이토치 한국 사용자 모임 (@PyTorchKR) header
파이토치 한국 사용자 모임 (@PyTorchKR) profile picture

파이토치 한국 사용자 모임

@PyTorchKR

PyTorch Korea User Group, 한국어 커뮤니티의 새로운 글을 공유합니다. 📰 #PyTorchKR #파이토치 #파이토치한국사용자모임 --- 파이토치 한국 사용자 모임은 인공지능에 관심있는 분들께 PyTorch를 소개하고 함께 배우며 성장하는 것을 목표로 하고 있습니다🔥

 Seoul, Korea
discuss.pytorch.kr
Joined June 2022
  • Tweets2.7K
  • Following51
  • Followers147
3Media
Photo by 파이토치 한국 사용자 모임 (@PyTorchKR)Photo by 파이토치 한국 사용자 모임 (@PyTorchKR)Photo by 파이토치 한국 사용자 모임 (@PyTorchKR)
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch31 Oct 2023
Researchers and developers working with ML models and algorithms using PyTorch can now use AMD ROCm 5.7 ⚡🙌 Learn more on our blog: hubs.la/Q027fXch0
Image from PyTorch (@PyTorch)
4
49
2
261
59,600
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch4 Oct 2023
Announcing PyTorch 2.1 Stable Release ⚡✨ PyTorch 2.1 offers automatic dynamic shape and NumPy compilation support in torch.compile, as well as torch.distributed.checkpoint for saving/loading distributed training jobs on multiple ranks. Learn more 👇 hubs.ly/Q024nsHY0
Image from PyTorch (@PyTorch)
5
127
8
568
86,309
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch1 Dec 2022
Interested in learning more about the #PyTorchFoundation and what it means for PyTorch? @b0noi shares reflections from his perspective, exploring technical and business governance. Learn more: bit.ly/3XLrhs6
Image from PyTorch (@PyTorch)
9
24
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch4 Oct 2022
If your training jobs are using DataParallel, consider using DDP! Watch the video to learn what DDP is doing under all those function calls, and how it can help more than using DP for multi-GPU training bit.ly/3SSqDG3
Image from PyTorch (@PyTorch)
11
1
37
파이토치 한국 사용자 모임 retweeted
AI at Meta (@AIatMeta) profile picture
AI at Meta
@AIatMeta23 Sep 2022
In the latest episode of "The Untold Stories of Open Source" podcast, @PyTorch co-creator @SoumithChintala talks about the creation of PyTorch. Listen now: bit.ly/3SIakvx #PyTorchFoundation @linuxfoundation

Easily listen to The Untold Stories of Open Source in your podcast app of choice

Quickly and easily listen and subscribe to The Untold Stories of Open Source for free in your podcast app of choice.

untold-stories-of-open-source.captivate.fm
The Linux Foundation (@linuxfoundation) profile pictureThe Linux Foundation
@linuxfoundation22 Sep 2022
Hear the story behind Soumith Chintala and the creation of @PyTorch in the latest episode of "The Untold Stories of Open Source" podcast. Listen now on Spotify: hubs.la/Q01n09Pl0 #PyTorch #PyTorchFoundation #AI @linuxfoundation
3
6
33
파이토치 한국 사용자 모임 retweeted
NVIDIA AI Developer (@NVIDIAAIDev) profile picture
NVIDIA AI Developer
@NVIDIAAIDev23 Sep 2022
NVIDIA GET3D is a generative #AI model that can be used to populate virtual worlds with 3D objects and characters. #GTC22 Learn more: nvda.ws/3UzNS9B
8
109
29
491
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch23 Sep 2022
(3/3) Find out why we save models and optimizer states separately in the 5th FSDP video. And if you’re wondering what happens when the model size is >20B, sit tight 🪑 for the following tutorial on local_state_dict!
2
8
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch23 Sep 2022
(2/3) Aside from handling the model on rank 0 CPU, loading the model is very similar to regular PyTorch loading. With the fullstate_save_policy, you can adjust offload to CPU for CPU memory usage and configure rank 0 to prevent loading redundancy.
1
2
4
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch23 Sep 2022
(1/3) Discover how to load and save models with FSDP full_state_dict, a process that fully assembles models up to 20B and optimizers on the rank 0 CPU memory. We’ll clearly explain each step. Learn how to maximize checkpoints: bit.ly/3R6mvAT
Image from PyTorch (@PyTorch)
2
14
58
파이토치 한국 사용자 모임 retweeted
Soumith Chintala (@soumithchintala) profile picture
Soumith Chintala
@soumithchintala22 Sep 2022
Here's a 30-minute podcast digging into my personal journey -- how I got into AI, Open Source, eventually rabbit-holing to my time in Hyderabad, Vellore, New York. (it incorrectly says I went to IIIT for Undergrad, but I went to VIT, and spent the last 6 months at IIIT)
Mark Miller, Take Responsibility. Give Credit. (@EUSP) profile pictureMark Miller, Take Responsibility. Give Credit.@EUSP22 Sep 2022
This episode of The Untold Stories of Open Source is available on all major audio streaming platforms. untold-stories-of-open-sourc… @Meta @soumithchintala @linuxfoundation #ai #ml #opensource
5
15
4
248
파이토치 한국 사용자 모임 retweeted
AK (@_akhaliq) profile picture
AK
@_akhaliq21 Sep 2022
NeRF-Factory: an awesome PyTorch NeRF collection(NeRF, NeRF++, Mip-NeRF, Mip-NeRF360, Ref-NeRF, Plenoxels, DVGO) Project Page: kakaobrain.github.io/NeRF-Fa… GitHub: github.com/kakaobrain/NeRF-F… @huggingface models: huggingface.co/nrtf/nerf_fac…
3
89
5
447
파이토치 한국 사용자 모임 retweeted
Joshua Starmer (@joshuastarmer) profile picture
Joshua Starmer@joshuastarmer19 Sep 2022
Coding Neural Networks doesn't have to be frightening...all you need is PyTorch + Lightning!!! BAM! youtu.be/khMzi6xPbuM #BAM #STATQUEST
Introduction to Coding Neural Networks with PyTorch and Lightning

Introduction to Coding Neural Networks with PyTorch and Lightning

Although we've seen how to code a simple neural network with PyTorc...

youtube.com
4
76
3
452
파이토치 한국 사용자 모임 retweeted
PyTorch Best Practices (@PyTorchPractice) profile picture
PyTorch Best Practices@PyTorchPractice19 Sep 2022
MultiGrain: a unified image embedding for classes and instances github.com/zjykzj/multigrain… #deeplearning #machinelearning #ml #ai #neuralnetworks #datascience #PyTorch
GitHub - zjykzj/multigrain-pytorch: MultiGrain: a unified image embedding for classes and instances

GitHub - zjykzj/multigrain-pytorch: MultiGrain: a unified image embedding for classes and instances

MultiGrain: a unified image embedding for classes and instances - zjykzj/multigrain-pytorch

github.com
17
1
23
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch19 Sep 2022
(3/3) A transformer wrapper is not only faster but also improves overall performance. Whereas the default wrapper makes sharding choices based on parameter count, the transformer wrapper knows how your model works and detects appropriate breaks to shard.
2
11
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch19 Sep 2022
(2/3) To start, we cover how to expedite transformer model training speed by up to 2x+ with its new transformer wrapper and go over layer classes. ⏩ bit.ly/3qNLL4b
1
3
8
파이토치 한국 사용자 모임 retweeted
PyTorch (@PyTorch) profile picture
PyTorch
@PyTorch19 Sep 2022
(1/3) This week, we’re kicking off a new series of 10 video tutorials from @lessw2020 to walk you through Fully Sharded Data Parallel (FSDP). Here’s an intro to what you’ll be learning: bit.ly/3qNTnUk
Image from PyTorch (@PyTorch)
1
16
111
파이토치 한국 사용자 모임 retweeted
ptrblck (@ptrblck_de) profile picture
ptrblck@ptrblck_de11 Sep 2022
If you are using @StableDiffusion and want to make it faster, check Simo's post and notebook showing how nvFuser will speed it up: reddit.com/r/MachineLearning…
4
37
1
272
파이토치 한국 사용자 모임 retweeted
Soumith Chintala (@soumithchintala) profile picture
Soumith Chintala
@soumithchintala9 Sep 2022
A magically easy way to get accurate RGBD-SLAM for your robot: use an iPhone + @record3d ! The baseline: RealSense + 2D Lidar + bad open-source SLAM Record3D streams RGBD + Odometry from iPhone to robot (USB or WiFi). Nice Python API. iPhone Pro is sensor-rich: LiDAR, IR, RGB
5
32
2
200
파이토치 한국 사용자 모임 retweeted
PyTorch Best Practices (@PyTorchPractice) profile picture
PyTorch Best Practices@PyTorchPractice4 Aug 2022
Develop like a Pro with NVIDIA + Docker + VS Code + PyTorch blog.roboflow.com/nvidia-doc… #deeplearning #machinelearning #ml #ai #neuralnetworks #datascience #PyTorch
Develop like a Pro with NVIDIA + Docker + VS Code + PyTorch

Develop like a Pro with NVIDIA + Docker + VS Code + PyTorch

SUMMARY Managing NVIDIA drivers and CUDA versions directly on a host machine leads to brittle environments that break when dependencies drift. This guide walks through installing NVIDIA drivers on...

roboflow.com
23
61
파이토치 한국 사용자 모임 retweeted
Soumith Chintala (@soumithchintala) profile picture
Soumith Chintala
@soumithchintala1 Sep 2022
PyTorch M1 GPU support (still in alpha) already being put to good use: replicate.com/blog/run-stabl…
Run Stable Diffusion on your M1 Mac’s GPU – Replicate blog

Run Stable Diffusion on your M1 Mac’s GPU – Replicate blog

How to run Stable Diffusion locally so you can hack on it

replicate.com
13
104
7
721
Load more