Notices by Akionux (akionux@status.akionux.net)

Akionux (akionux@status.akionux.net)'s status on Sunday, 04-Sep-2022 08:23:28 JST Akionux

フランスで「ONE PIECE」上映中に観客が大暴れ…　お国柄？日本アニメの変わった“観られ方”（デイリー新潮） - Yahoo!ニュース - https://news.yahoo.co.jp/articles/553fc863c85121c591aa3bf4c8751d30e29bae55?page=1
In conversation Sunday, 04-Sep-2022 08:23:28 JST from status.akionux.net permalink
Attachments
1. Untitled attachment
Akionux (akionux@status.akionux.net)'s status on Sunday, 04-Sep-2022 08:11:24 JST Akionux

『耳をすませば』テレビ放映でコンプラツッコミ殺到　自転車二人乗りや個人情報流出｜ニフティニュース - https://news.nifty.com/article/entame/etc/12184-1847083/
In conversation Sunday, 04-Sep-2022 08:11:24 JST from status.akionux.net permalink
Attachments
1. Untitled attachment
Akionux (akionux@status.akionux.net)'s status on Saturday, 03-Sep-2022 22:32:53 JST Akionux

画像生成AI「Stable Diffusion」をKritaで使えるようにするオープンソースのプラグイン登場 - GIGAZINE - https://gigazine.net/news/20220902-stable-diffusion-krita-plugin/
In conversation Saturday, 03-Sep-2022 22:32:53 JST from status.akionux.net permalink
Attachments
1. Untitled attachment
Akionux (akionux@status.akionux.net)'s status on Saturday, 03-Sep-2022 12:09:51 JST Akionux

「Stable Diffusionであの構図や雰囲気を再現するには何と書けば良いのか？」が目で見て一発でわかる「The Ai Art」 - GIGAZINE - https://gigazine.net/news/20220902-stable-diffusion-the-ai-art/
In conversation Saturday, 03-Sep-2022 12:09:51 JST from status.akionux.net permalink
Attachments
1. Untitled attachment
Akionux (akionux@status.akionux.net)'s status on Friday, 02-Sep-2022 21:21:55 JST Akionux
in reply to
- Akionux
@akionux お家で stable diffusion 動いた!やったね

In conversation Friday, 02-Sep-2022 21:21:55 JST from status.akionux.net permalink
Akionux (akionux@status.akionux.net)'s status on Friday, 02-Sep-2022 07:17:36 JST Akionux

仮面ライダーブラック？漆黒のバッタ、草刈り中に発見（南日本新聞） - Yahoo!ニュース - https://news.yahoo.co.jp/articles/863726400b12f96f0106a1bd131d6e028551bdc6
In conversation Friday, 02-Sep-2022 07:17:36 JST from status.akionux.net permalink
Attachments
1. Untitled attachment
Akionux (akionux@status.akionux.net)'s status on Thursday, 01-Sep-2022 15:35:16 JST Akionux

Introducing nvFuser, a deep learning compiler for PyTorch | PyTorch - https://pytorch.org/blog/introducing-nvfuser-a-deep-learning-compiler-for-pytorch/
In conversation Thursday, 01-Sep-2022 15:35:16 JST from status.akionux.net permalink
Attachments
1. Untitled attachment
Akionux (akionux@status.akionux.net)'s status on Wednesday, 31-Aug-2022 23:17:09 JST Akionux

"How Crash Bandicoot Hacked The Original Playstation | War Stories | Ars Technica" を YouTube で見る - https://youtu.be/izxXGuVL21o
クラッシュ・バンディクーがいかに技術、デザイン、考え方で抜きん出ていたかがよくわかった。
In conversation Wednesday, 31-Aug-2022 23:17:09 JST from status.akionux.net permalink
Attachments
1. Untitled attachment
Akionux (akionux@status.akionux.net)'s status on Wednesday, 31-Aug-2022 18:38:01 JST Akionux

Intel製CPUでも画像生成AI「Stable Diffusion」を動かせる「stable_diffusion.openvino」が登場、誰でもダウンロード可能に - GIGAZINE - https://gigazine.net/news/20220831-stable-diffusion-openvino/
In conversation Wednesday, 31-Aug-2022 18:38:01 JST from status.akionux.net permalink
Attachments
1. Untitled attachment
Akionux (akionux@status.akionux.net)'s status on Wednesday, 24-Aug-2022 14:54:23 JST Akionux

Canonical enables Ubuntu on Allwinner’s Nezha RISC-V boards | Canonical - https://canonical.com/blog/canonical-enables-ubuntu-on-allwinners-nezha-risc-v-boards
In conversation Wednesday, 24-Aug-2022 14:54:23 JST from status.akionux.net permalink
Attachments
1. Untitled attachment
Akionux (akionux@status.akionux.net)'s status on Friday, 05-Aug-2022 12:23:27 JST Akionux

Gradio - https://gradio.app/
In conversation Friday, 05-Aug-2022 12:23:27 JST from status.akionux.net permalink
Attachments
1. Domain not in remote thumbnail source whitelist: gradio.app
  
  Gradio
  
  from @Gradio
  
  Build & Share Delightful Machine Learning Apps
Akionux (akionux@status.akionux.net)'s status on Friday, 05-Aug-2022 08:09:02 JST Akionux

半導体洗浄時におけるナノ構造物の倒壊の様子、北大などが詳細観察に成功 (1) | TECH+ - https://news.mynavi.jp/techplus/article/20220801-2413594/
半導体と表面張力にこういう関わりがあるのか
In conversation Friday, 05-Aug-2022 08:09:02 JST from status.akionux.net permalink
Attachments
1. Domain not in remote thumbnail source whitelist: news.mynavi.jp
  
  半導体洗浄時におけるナノ構造物の倒壊の様子、北大などが詳細観察に成功 (1)
  
  北海道大学(北大)とSCREENホールディングスは7月29日、半導体洗浄時のナノ構造物の倒壊挙動を明らかにしたことを発表した。
Akionux (akionux@status.akionux.net)'s status on Thursday, 04-Aug-2022 18:47:49 JST Akionux

Rooting instructions | Valetudo - https://valetudo.cloud/pages/general/rooting-instructions.html
In conversation Thursday, 04-Aug-2022 18:47:49 JST from status.akionux.net permalink
Attachments
1. Domain not in remote thumbnail source whitelist: valetudo.cloud
  
  Rooting instructions
  
  Rooting instructions
Akionux (akionux@status.akionux.net)'s status on Saturday, 30-Jul-2022 22:28:26 JST Akionux

ドウシシャの照明&サーキュレーターが想像を超えてQOL爆上がり【家電レビュー】- 家電 Watch - https://kaden.watch.impress.co.jp/docs/column_review/kaden/1426620.html
In conversation Saturday, 30-Jul-2022 22:28:26 JST from status.akionux.net permalink
Attachments
1. Domain not in remote thumbnail source whitelist: kaden.watch.impress.co.jp
  
  ドウシシャの照明&サーキュレーターが想像を超えてQOL爆上がり【家電レビュー】
  
  from 株式会社インプレス
  
  仕事部屋であり、トレーニング部屋でもある筆者宅の4畳半。デスクとパソコン2台とモニター類3台で部屋の半分近くを占め、さらに撮影ブース、トレーニング用の室内自転車があって、一時的にメーカーなどからお借りしているアイテムやら空き箱やらが転がっているので、時には立錐の余地もない状態になることもある。
Akionux (akionux@status.akionux.net)'s status on Saturday, 16-Jul-2022 16:59:08 JST Akionux

北海道道　大泉洋ＳＰ～スターの素顔に“ミスター”鈴井貴之が迫る！
https://plus.nhk.jp/watch/st/010_g1_2022071555752?playlist_id=baa8cbd5-e614-419e-a46e-15576189efc0
In conversation Saturday, 16-Jul-2022 16:59:08 JST from status.akionux.net permalink
Attachments
1. No result found on File_thumbnail lookup.
  
  https://plus.nhk.jp/watch/st/010_g1_2022071555752?playlist_id=baa8cbd5-e614-419e-a46e-15576189efc0
Akionux (akionux@status.akionux.net)'s status on Saturday, 16-Jul-2022 10:25:17 JST Akionux
in reply to
- Akionux
SBCでRK3588Sのメモリ32GBはかなり強い

In conversation Saturday, 16-Jul-2022 10:25:17 JST from status.akionux.net permalink
Akionux (akionux@status.akionux.net)'s status on Saturday, 16-Jul-2022 10:22:15 JST Akionux
in reply to
- Akionux
これ読んだけど、EfficientFormerよりパラメータのレンジが高いとこにあるみたい

In conversation Saturday, 16-Jul-2022 10:22:15 JST from status.akionux.net permalink
Akionux (akionux@status.akionux.net)'s status on Friday, 15-Jul-2022 19:19:34 JST Akionux

Orange Pi 5 could be the most affordable Rockchip RK3588S SBC - CNX Software - https://www.cnx-software.com/2022/07/15/orange-pi-5-most-affordable-rockchip-rk3588s-sbc/
In conversation Friday, 15-Jul-2022 19:19:34 JST from status.akionux.net permalink
Attachments
1. Domain not in remote thumbnail source whitelist: www.cnx-software.com
  
  Orange Pi 5 could be the most affordable Rockchip RK3588S SBC - CNX Software
  
  from @cnxsoft
  
  Orange Pi 5 (LTS) is an upcoming single board computer powered by Rockchip RK3588S cost-down octa-core Cortex-A76/A55 processor that should offer one of
Akionux (akionux@status.akionux.net)'s status on Wednesday, 13-Jul-2022 12:43:29 JST Akionux

[2207.05221] Language Models (Mostly) Know What They Know - https://arxiv.org/abs/2207.05221
In conversation Wednesday, 13-Jul-2022 12:43:29 JST from status.akionux.net permalink
Attachments
1. Domain not in remote thumbnail source whitelist: static.arxiv.org
  
  Language Models (Mostly) Know What They Know
  
  We study whether language models can evaluate the validity of their own claims and predict which questions they will be able to answer correctly. We first show that larger models are well-calibrated on diverse multiple choice and true/false questions when they are provided in the right format. Thus we can approach self-evaluation on open-ended sampling tasks by asking models to first propose answers, and then to evaluate the probability "P(True)" that their answers are correct. We find encouraging performance, calibration, and scaling for P(True) on a diverse array of tasks. Performance at self-evaluation further improves when we allow models to consider many of their own samples before predicting the validity of one specific possibility. Next, we investigate whether models can be trained to predict "P(IK)", the probability that "I know" the answer to a question, without reference to any particular proposed answer. Models perform well at predicting P(IK) and partially generalize across tasks, though they struggle with calibration of P(IK) on new tasks. The predicted P(IK) probabilities also increase appropriately in the presence of relevant source materials in the context, and to the presence of hints towards the solution of mathematical word problems. We hope these observations lay the groundwork for training more honest models, and for investigating how honesty generalizes to cases where models are trained on objectives other than the imitation of human writing.
Akionux (akionux@status.akionux.net)'s status on Wednesday, 13-Jul-2022 12:42:57 JST Akionux

[2207.05501] Next-ViT: Next Generation Vision Transformer for Efficient Deployment in Realistic Industrial Scenarios - https://arxiv.org/abs/2207.05501
コスパViTの競争が最近かなり激しい
つい最近出たEfficientFormer、超えられた？
In conversation Wednesday, 13-Jul-2022 12:42:57 JST from status.akionux.net permalink
Attachments
1. Domain not in remote thumbnail source whitelist: static.arxiv.org
  
  Next-ViT: Next Generation Vision Transformer for Efficient Deployment in Realistic Industrial Scenarios
  
  Due to the complex attention mechanisms and model design, most existing vision Transformers (ViTs) can not perform as efficiently as convolutional neural networks (CNNs) in realistic industrial deployment scenarios, e.g. TensorRT and CoreML. This poses a distinct challenge: Can a visual neural network be designed to infer as fast as CNNs and perform as powerful as ViTs? Recent works have tried to design CNN-Transformer hybrid architectures to address this issue, yet the overall performance of these works is far away from satisfactory. To end these, we propose a next generation vision Transformer for efficient deployment in realistic industrial scenarios, namely Next-ViT, which dominates both CNNs and ViTs from the perspective of latency/accuracy trade-off. In this work, the Next Convolution Block (NCB) and Next Transformer Block (NTB) are respectively developed to capture local and global information with deployment-friendly mechanisms. Then, Next Hybrid Strategy (NHS) is designed to stack NCB and NTB in an efficient hybrid paradigm, which boosts performance in various downstream tasks. Extensive experiments show that Next-ViT significantly outperforms existing CNNs, ViTs and CNN-Transformer hybrid architectures with respect to the latency/accuracy trade-off across various vision tasks. On TensorRT, Next-ViT surpasses ResNet by 5.4 mAP (from 40.4 to 45.8) on COCO detection and 8.2% mIoU (from 38.8% to 47.0%) on ADE20K segmentation under similar latency. Meanwhile, it achieves comparable performance with CSWin, while the inference speed is accelerated by 3.6x. On CoreML, Next-ViT surpasses EfficientFormer by 4.6 mAP (from 42.6 to 47.2) on COCO detection and 3.5% mIoU (from 45.2% to 48.7%) on ADE20K segmentation under similar latency. Code will be released recently.

Before

Public

Notices by Akionux (akionux@status.akionux.net)

User actions

Following 1

Followers 0

Groups 0

Statistics

Feeds