Summary of Gemini's 60-page technical report. 1. Written in Jax and trained using TPUs. The architecture, while not explained in details, seems similar to Flamigo's. 2. Gemini Pro's performance is similar to GPT-3.5 and Gemini Ultra is reported to be better than GPT-4. Nano-1
BIG TECH
-
Meta Launches Standalone Image Generator for EMU
By
–
Meta just launched a standalone image generator for EMU too!
-
Google’s AI Model Competes with GPT-4, Cuts Infrastructure Costs
By
–
Seems to solidly compete with GPT-4 on benchmarks.
Google has existing customers and surfaces to start the feedback loop, without worrying about adoption.
And Google will use TPUs for inference, so doesn't have to pay NVIDIA their 70% margins (like @OpenAI and @Microsoft has to -
Gemini Pro Fails to Address Current Events Properly
By
–
So I just asked Gemini Pro in Bard for the updates regarding the hottest news story these days, the war in Israel and Gaza, and its answer essentially was, I kid you not, "just Google it."
-
Google Delays High-End AI Model Over Safety Assessment Concerns
By
–
Google claims that this delay is primarily due to their ongoing efforts to assess and refine safety for this high end model. 2/4
-
Gemini Pro Launched: Performance Compared to GPT-3.5
By
–
Gemini Pro is available right now and you can check it out in Bard. That version, according to Google, is slightly better than GPT-3.5, so in other words Gemini in production is about a year behind OpenAI's capabilities. 3/4
-
Google Gemini Closing Gap Against OpenAI Next Generation Models
By
–
It remains to be seen if they can close that gap in the upcoming months, as I'm sure @OpenAI has also been working diligently on their next generation model. Read the full blog post announcement here: https://
blog.google/products/bard/
google-bard-try-gemini-ai/
… #AI #ArtificialIntelligence #GenerativeAI #GenAI 4/4 -
Google Gemini Ultra Benchmarks Beat GPT-4 Across Categories
By
–
The benchmarks that Google shared for Gemini Ultra look really impressive, and seem to be generally beating GPT-4 in almost all categories. However, as of right now, Gemini Ultra is still not available, and will not be available until some time early next year. 1/4
-
Evaluating Google’s Titan and PaLM Embeddings with MTEB Benchmarks
By
–
Here I am trying to figure out how I can get MTEB scores for Titan and PaLM embeddings and doing GPU hour math and I'm not cool enough to get a technical report for a Big Boy Model from Google.
-
Gemini Model Card Analysis Delayed by Jury Duty
By
–
Still reading through the model card/report on Gemini (which I did not make the pre-brief cut for this one, alas) so issue will be coming later due to getting called in for jury duty yesterday!