Naza

NLP Engineer | AI Player

[アル猫🐱｜AI × 投資 @alpha_cat0]

AIエージェント用スキル「テスタ株判断スキル（tesuta-stock-judge）」をOSS公開した🚀

普段から株の売買判断を自動化するためにAIスキルを自作しているけど、テスタさんの公開情報・取引原則を参考にしたこれはかなり実用的。

特に「高エッジのみ参加、弱い場面は見送り」が効く。不要なエントリーと損失をちゃんと減らせる。

特徴
・短期は需給読み／中長期は材料・業績で判断
・リスクファースト（損切りを先に定義）
・資金量×流動性でサイジングを考慮
・ハルシネーション対策（数値を捏造しない）

※ 非公式ファンメイド
※ 教育・参考目的であり、投資助言ではありません

#OSS #AI #株式投資

[Pratham @Prathkum]

10 AI websites you can't live without:

1. https://t.co/AOEU60iMQt → AI noise cancellation

2. https://t.co/XeYeMz1x5R → Build websites with AI

3. https://t.co/YeVY1NH7lf → Easy writing as Friday

4. https://t.co/OBnTpEFl5k → Design with AI

5. https://t.co/3lngP3x3eM → Google search + ChatGPT

6. https://t.co/DoUAIekKTG → Discover AI apps 
for every usecase

7. https://t.co/pTTBH0ppCU → Slides in minutes with AI

8. https://t.co/28kiIXyT7t → Summarize Slack messages

9. https://t.co/TaPorLiBEV → Automatic meeting notes

10. https://t.co/pxXPyJhwPB → Text to video

[Chris Ociepa @ChrisOciepa]

@marktenenholtz Great example! 👏
As a side note, I would like to add that for production deployments, more sophisticated message queuing and handling are needed. Here is a nice tool from 🤗that address this: https://t.co/BuAy4IlLaS

[Mark Tenenholtz @marktenenholtz]

If you thought this was interesting, you can check out all of the code here: https://t.co/8ID6KDVwES (code is in the backends folder)

And the website is live here: https://t.co/VDjZWL8L43 (disclaimer: I'm an awful front-end dev!)

Let me know what you think of it!

[Mark Tenenholtz @marktenenholtz]

(shoutout to this issues thread: https://t.co/sPcJWP0iZu)

That callback is passed to langchain's ChatGPT wrapper (although you can certainly implement this yourself).

[Sully @SullyOmarr]

OpenAI's discord is filled with amazing prompts.

Unfortunately it's closed (full).

Here are some of the more interesting ones i've found (GPT4 - prompts in alt text)

[Matei Zaharia @matei_zaharia]

MLflow just added first-class support for LLMs, including integrations with @huggingface transformers/pipelines, @OpenAI and @langchain! Open source #LLMOps is here. https://t.co/l870bYyPdE

[Sanyam Bhutani @bhutanisanyam1]

The best tutorials on building LLM powered applications 📚

@GregKamradt is an incredible teacher of @langchain:

✅ Top down & applied series
✅ Amazing teaching style
✅ Very practical examples

https://t.co/5mMrkswmSf

[Madisen Taylor @madisenxtaylor]

Behind the Scenes of #WoodstockAI: How me and the @huggingface team pulled off the largest-ever AI meetup (5k attendees) and lived to tell the tale!

[Hugging Face @huggingface]

The amazing @lvminzhang has open-sourced the next version of ControlNet 🎨

- Canny v1.1
- Open-pose v1.1
- & 12 others new checkpoints ♥️

You can find the🧨Diffusers' checkpoints here:
👉https://t.co/hE8fUh7nAl

[elvis @omarsar0]

Building LLM applications for production

Great list of topics to consider when building LLM applications for production. Great experience sharing by @chipro. Huge overlap with what I have experienced in the space as well.

https://t.co/AwdAeGllbR

[Andrej Karpathy @karpathy]

Love it 👏 - much fertile soil for indie games populated with AutoGPTs, puts "Open World" to shame. Simulates a society with agents, emergent social dynamics.
Paper: https://t.co/I07IJwweHE
Demo: https://t.co/pYNF4BBveG
Authors: @joon_s_pk @msbernst @percyliang @merrierm et al.

[Naza @CatKnight0]

紹介ありがとうございます！
結構時間がかかって書いた記事だから、役に立つのは嬉しいです！

[Naza @CatKnight0]

My new article about some useful tips to speed up Python efficiency.

[Towards Data Science @TDataScience]

10 Techniques to Speed Up Python Runtime by Xu LIANG https://t.co/leRRYYy5gZ

[Naza @CatKnight0]

グッズをもらえたい！
#VanguardJapan20

[Naza @CatKnight0]

I picked up ACL2020 papers that are related to these NLP tasks, Named Entity Recognition, Relation Extraction, Event Detection, Coreference Resolution, Knowledge Graph Related
#acl2020nlp #acl2020
https://t.co/vYnEDOfI7v

[Naza @CatKnight0]

I knew it! The TACRED data set has so many annotation errors. This paper correct the errors and boost the SOTA of relation extraction.

TACRED Revisited: A Thorough Evaluation of the TACRED Relation Extraction Task
#acl2020 #acl2020nlp

[Naza @CatKnight0]

@Fatal1tyV 创立早没什么，问题是3M公司一直在创新，这点是真了不起

[Naza @CatKnight0]

@sudharsan2020 These papers might be helpful.
https://t.co/J64SuSCFGb

[Naza @CatKnight0]

@sudharsan2020 I am glad you like it. There are two common approaches for imbalanced problems. One is the model approach, another is the data approach. 10:90 is pretty skewed. In such a case, I recommend the data approach, which increases the data size with data augmentation method.