Ram Ananth

Machine learning Engineer, former Founder-in-Residence @join_ef, former Grad Researcher @uoft. Interested in the world of startups

[Waymo @Waymo]

We’re bringing the magic of Waymo to Toronto! 🇨🇦

In the coming days, you’ll begin seeing Waymo’s all-electric vehicles driving through Toronto’s streets as we lay the groundwork for our ride-hailing service.

Our goal is to begin serving riders in 2028.

Sign up for future updates: https://t.co/ndrwYQlbZJ

[Pratyush Kumar @pratykumar]

📢 Open-sourcing the Sarvam 30B and 105B models! Trained from scratch with all data, model research and inference optimisation done in-house, these models punch above their weight in most global benchmarks plus excel in Indian languages.

Get the weights at Hugging Face and AIKosh. Thanks to the good folks at SGLang for day 0 support, vLLM support coming soon. Links, benchmark scores, examples, and more in our blog - https://t.co/DcCG3zlN8p

[Ram Ananth @RamAnanth29]

Built a fun mini project using these stunning 🍌 powered city views and real-time weather. “Navigating these temperature swings is an adventure

Compare cities, teleport randomly

https://t.co/bEfOS6HeCu (be gentle pls)

Github: https://t.co/6IN6kVDrVK

[Pavol Rusnak @PavolRusnak]

Nice Nano Banana Pro prompt for weather app:

CITY=Prague,Czechia

Present a clear, 45° top-down isometric miniature 3D cartoon scene of [CITY], featuring its most iconic landmarks and architectural elements. Use soft, refined textures with realistic PBR materials and gentle, lifelike lighting and shadows. Integrate the current weather conditions directly into the city environment to create an immersive atmospheric mood.
Use a clean, minimalistic composition with a soft, solid-colored background.

At the top-center, place the title “[CITY]” in large bold text, a prominent weather icon beneath it, then the date (small text) and temperature (medium text).
All text must be centered with consistent spacing, and may subtly overlap the tops of the buildings.
Square 1080x1080 dimension.

[contexto 🦞 @getcontexto]

Our manifesto 👇

Every LLM conversation you have makes someone else richer with your intelligence.
We are living through the largest transfer of intelligence in history, one prompt at a time.

[Ram Ananth @RamAnanth29]

This is going to enable a lot of use cases

[Gradio @Gradio]

Everyone's rushing to implement MCP but nobody's talking about the easiest way to do it.

import gradio as gr
demo = gr.Interface(fn=your_function, ...)
demo.launch(mcp_server=True) # ← that's it 🤯

Your Gradio app is now an MCP server for any LLM.

Learn how:👇

[Soumye Singhal @soumyesinghal]

⚡⚡ Llama-Nemotron-Ultra-253B just dropped: our most advanced open reasoning model
🧵👇

[Ram Ananth @RamAnanth29]

Glad to be a small part of this🤗 Please do give some of these amazing people on this Space a follow

[Matt Valoatto @mvaloatto]

Want more “good machine learning” in your X feed? Here is a new Space for you: 🔔 Top HF Users To Follow On X - https://t.co/82aFwsNIgt

Ever since I fell down the AI rabbit hole, it hasn’t been super easy to spot and follow the most impactful @HuggingFace contributors on X. So, inspired by @Weyaxi leaderboards, I decided to create a list just for this purpose.

Why, you ask? More details here: https://t.co/eLkeb9pxn4

[MMitchell @mmitchell_ai]

With the rise of AI-generated "fake" human content--"deepfake" imagery, voice cloning scams & chatbot babble plagiarism--those of us working on social impact @huggingface put together a collection of some of the state-of-the-art technology that can help:
https://t.co/nFS7GW8dtk

[Ram Ananth @RamAnanth29]

@1littlecoder @lukestanley @Gradio @clipdropapp You are correct about the model not being released yet. According to the Stability AI website, the XL beta is only available for its API customers and dreambooth users, so the API being used in the space is not clipdrop but the platform API provided by https://t.co/4cCIzjvLhZ

[Ram Ananth @RamAnanth29]

Stable Diffusion XL beta is definitely something worth trying @Gradio demo at https://t.co/cRVgZm2KO8

[Ram Ananth @RamAnanth29]

@Gradio The ability to draw text is also very interesting

[Ram Ananth @RamAnanth29]

For anyone else wanting to try the latest open-source model Dolly 2.0, I made a simple @Gradio demo at https://t.co/zuY9eu4NVO thanks to @_akhaliq

[Ram Ananth @RamAnanth29]

It would also be interesting to see if any open-source LLM on @huggingface could replace ChatGPT as the controller

[Jim Fan @DrJimFan]

HuggingGPT is the most interesting paper I read this week. It gets very close to the "Everything App" vision that I described a while ago.

ChatGPT acts as a controller over the *AI model space*, picks the right model (app) given the human specification, and assembles them correctly to solve the task. It's multimodal in a "low-bandwidth" way - all modalities need to be compressed and connected through text strings.

HuggingGPT is also related to Prismer's idea: leverage pre-trained domain expert models as much as possible. Sometimes training less is doing more!

[Ram Ananth @RamAnanth29]

@akarsh_ghale According to their blog https://t.co/QA6H0KtN05 it has a lower average WER and better robustness to noise compared to Whisper

[Ram Ananth @RamAnanth29]

AssemblyAI recently announced API access to their state-of-the-art Speech Recognition Model, Conformer-1. Try out the unofficial @Gradio demo at https://t.co/sYbxYUXbVl, which includes a suite of audio intelligence options such as transcription, summarization, topic detection

[Ram Ananth @RamAnanth29]

Happy to have contributed to the Tasks documentation for Reinforcement Learning during Hacktoberfest

[merve @mervenoyann]

Few months ago we have opened Tasks (https://t.co/eadCF2ZjGj) documentation project for crowdsourcing and I'm thrilled to announce that most of the below pages are now complete thanks to community! 😍

Couple of things community has accomplished:
❤️ It all started with this Call for Contributions on Hacktoberfest 🎃 https://t.co/KJx4TzdEo0
♥️ Thirteen of below tasks were contributed completely by community 🌟
♥️ It was the first ever contribution of some folks! 🙌

If you want, you can still contribute to this project if you think there are things to improve in any of the pages 😊

[Ram Ananth @RamAnanth29]

Tried out the newly released visual ChatGPT and decided to move a part of the demo to @huggingface Spaces (with a fraction of the tools that could fit in memory) with the help of GPU grant @_akhaliq. Try it out with your OpenAI API key at https://t.co/hAJDPkZ4Rh

[Ram Ananth @RamAnanth29]

The demo is based on the paper https://t.co/5LEyPUSZTv and its associated open-source implementation. Once again thanks to @_akhaliq for the prompt response in granting a GPU

[Ram Ananth @RamAnanth29]

Made an unofficial @huggingface demo for ZoeDepth, a technique that tackles the problem of depth estimation from a single image by combining relative and metric depth. Try it out at https://t.co/G0Wkr8Iy5w

[Ram Ananth @RamAnanth29]

Appreciate all the help from @_akhaliq for all the compute resources and also the authors for open-sourcing the code at https://t.co/lgP3ydVAuy

[Ram Ananth @RamAnanth29]

Synthetic data has several possible use cases in healthcare, finance,etc. RealTabFormer is one such technique from World Bank researchers to generate synthetic tabular data using GPT2. Created an unofficial demo on @huggingface for this technique https://t.co/dquasW7PAf