🚀 Join the JinaAI x #GSoC 2023 webinar & unlock your full potential!
🧑💻 Dive into cutting-edge projects
🌟 Gain invaluable skills
🌍 Collaborate with global talents
Don’t miss out, #opensource heroes!
🗓️ Save the date: Thursday Mar 23rd, 2 PM CET.
https://t.co/a0oB3mBT7H
In this @DataTalksClub Open Source Spotlight series, our @JinaAI_ NLP Engineer, @isabelle_mohr demonstrates the capabilities of Jina NOW and what's coming next!
https://t.co/nhG6I2c4Sk
We created AgentChain, a multimodal LLM agent amplified with foundation models and a set of tools 💪 to take action:
- Fed it a data frame 📊
- Replies to my question 🔎
- Calls me ☎️ to say I won more Oscars than @ABelaweid
Checkout what else it can do: https://t.co/Jmkig0eqbw
Multimodal Vision-Language BLIP-2 (from @SalesforceOrg ) as a service using @JinaAI_ ! 🚀
https://t.co/N9olA2T92k
👉 Implement a Visual Question Answering service with BLIP-2, a model that's also useful for image captioning and image-text retrieval. 📸💬
On #InternationalWomensDay, we're recognizing trailblazing #WomenInTech breaking barriers for the next generation of female leaders.
At Jina, we're committed to championing diversity & inclusivity in tech for a more equitable future!
#IWD2023
🪦 Expired: Deploy a deep learning model
🥱 Tired: Deploy a deep learning model with cute kittens
😻 Wired: Deploy a deep learning model with Jina and rainbow butterfly unicorn kittens 🌈
https://t.co/c10W57UC6b
#Python has changed a lot over the years. @samsja19 is going to walk us through how using it has changed over the years 🐍 #PWC2023 begins March 13! Check out this talk and more 👉 https://t.co/EjMyBSNqZl
Register today: https://t.co/OB4PS30tgb
"Confabulation" is a better word than "hallucination" for AI models getting things wrong. Confabulation is a sign of brain damage.
There is another technical term for what ChatGPT says: bullshit.
Read more in our latest blog by @ScottMartens8: https://t.co/m3pfkCZ1Ua
ChatGPT may tell you it knows languages that it doesn't know.
Large language models say things without regard for any relationship between what they say and any external or objective truth.
Read more in our latest blog by @ScottMartens8: https://t.co/m3pfkCZ1Ua
Ask Generative AI for “a stop sign, with the word 'STOP' printed on it in all capital letters.”
AI models are not trained to understand written text.
Read more in our latest blog by @ScottMartens8: https://t.co/m3pfkCZ1Ua
Ask generative AI to draw "six apples."
Large language models are some of the most massive and computationally demanding software in the world, and they cannot count to six!
Read more in our latest blog by @ScottMartens8: https://t.co/m3pfkCZ1Ua
Confabulation affects AI art, leading to odd flaws like rendering Boris Johnson as a woman.
Read more in our latest blog by @ScottMartens8: https://t.co/m3pfkCZ1Ua
Is Generative AI a dancing bear? Or CocaineBear?
The marvel is not that the bear dances well, but that the bear dances at all. — old proverb
DON'T PANIC! @JinaAI_ will show you what's really going on beneath the hype.
https://t.co/v7USQznFmZ
#chatgpt#aiartwork#CocaineBear
Imagine the event starts at 6:30pm and you are still prepping some slides for your talk at 6pm.
Btw. there were 160 people on the attendee list @PyBerlinPython
I would have been nervous. Sami was cool AF. 🤣 and then DESTROYED it.
Next stop keynote @samsja19 😎
The house was packed at @PyBerlinPython last night, with all 👀👂 on @samsja19 as he presented his talk on the power of type hinting and it's impact on ML.
Grateful to meet and reconnect with so many awesome folks in the #python community including @tiangolo!
If you have been working on content-based image retrieval tasks such as Google Landmark Retrieval Challenge, you know that ArcFaceLoss+GeM pooling+CNN is the answer. In the latest Finetuner 0.7.2, it is as easy as: https://t.co/l7qx6vfkEz