LogoThread Easy
  • Explorar
  • Criar thread
LogoThread Easy

Seu parceiro completo para threads do Twitter

© 2025 Thread Easy All Rights Reserved.

Explorar

Newest first — browse tweet threads

Keep on to blur preview images; turn off to show them clearly

RT @robzolkos: Wrote some notes on Fizzy's webhook system.  I'm sure the first of many deep dives I will do on the codebase.  A blessing fo…

RT @robzolkos: Wrote some notes on Fizzy's webhook system. I'm sure the first of many deep dives I will do on the codebase. A blessing fo…

Father of three, Creator of Ruby on Rails + Omarchy, Co-owner & CTO of 37signals, Shopify director, NYT best-selling author, and Le Mans 24h class-winner.

avatar for DHH
DHH
Tue Dec 02 16:54:58
Fizzy is deployed in three different data centers: US Mid-West (Chicago), US East (Ashburn), and EU (Amsterdam). Our writer is in US East, and the two other DCs are for reading. It's fast! P95 response time is 50 for readers, 80ms for writer!

Fizzy is deployed in three different data centers: US Mid-West (Chicago), US East (Ashburn), and EU (Amsterdam). Our writer is in US East, and the two other DCs are for reading. It's fast! P95 response time is 50 for readers, 80ms for writer!

Father of three, Creator of Ruby on Rails + Omarchy, Co-owner & CTO of 37signals, Shopify director, NYT best-selling author, and Le Mans 24h class-winner.

avatar for DHH
DHH
Tue Dec 02 16:52:24
RT @EnricoShippole: One of the most impactful areas of open-source software and artificial intelligence is the total democratization of leg…

RT @EnricoShippole: One of the most impactful areas of open-source software and artificial intelligence is the total democratization of leg…

Cofounder and Head of Post Training @NousResearch, prev @StabilityAI Github: https://t.co/LZwHTUFwPq HuggingFace: https://t.co/sN2FFU8PVE

avatar for Teknium (e/λ)
Teknium (e/λ)
Tue Dec 02 16:50:03
500K上下文微调只需要80G显存?

来看 Unsloth 的新教程,他们使用了损失重构和自动化块大小调优, 甚至他们去年搞的激活卸载算法已经成为业界标准了,这次更新的新版本进一步优化了,让微调模型的开销降低到了原来的1/10。

教程地址:https://t.co/ZIByZyt9on

500K上下文微调只需要80G显存? 来看 Unsloth 的新教程,他们使用了损失重构和自动化块大小调优, 甚至他们去年搞的激活卸载算法已经成为业界标准了,这次更新的新版本进一步优化了,让微调模型的开销降低到了原来的1/10。 教程地址:https://t.co/ZIByZyt9on

技术细节/1

avatar for karminski-牙医
karminski-牙医
Tue Dec 02 16:49:16
If you're at #NeurIPS, check out Erich Elsen's talk at 5:15 today on a new AI benchmark competition we created at @Databricks for economically valuable tasks, where frontier model still struggle significantly. https://t.co/DJBcfgtdqw

If you're at #NeurIPS, check out Erich Elsen's talk at 5:15 today on a new AI benchmark competition we created at @Databricks for economically valuable tasks, where frontier model still struggle significantly. https://t.co/DJBcfgtdqw

CTO at @Databricks and CS prof at @UCBerkeley. Working on data+AI, including @ApacheSpark, @DeltaLakeOSS, @MLflow, @DSPyOSS. https://t.co/nmRYAKFsWr

avatar for Matei Zaharia
Matei Zaharia
Tue Dec 02 16:48:03
RT @DbrxMosaicAI: We’re headed to #NeurIPS2025 as a Platinum sponsor! Catch us at booth 1619 to talk research, demos, and opportunities. Al…

RT @DbrxMosaicAI: We’re headed to #NeurIPS2025 as a Platinum sponsor! Catch us at booth 1619 to talk research, demos, and opportunities. Al…

CTO at @Databricks and CS prof at @UCBerkeley. Working on data+AI, including @ApacheSpark, @DeltaLakeOSS, @MLflow, @DSPyOSS. https://t.co/nmRYAKFsWr

avatar for Matei Zaharia
Matei Zaharia
Tue Dec 02 16:43:41
  • Previous
  • 1
  • More pages
  • 1794
  • 1795
  • 1796
  • More pages
  • 5634
  • Next