Gabriel Martín Blázquez (@gabrielmbmb_) 's Twitter Profile
Gabriel Martín Blázquez

@gabrielmbmb_

ML Engineer

ID: 1374788418606088195

linkhttps://www.gabrielmb.com/ calendar_today24-03-2021 18:21:40

564 Tweet

665 Takipçi

599 Takip Edilen

Dani Carmona (@danipolymath) 's Twitter Profile Photo

just did a live demo with the CRO and GMs for a fortune 500, in person, live with their product - nobody can move faster than we do

Dani Carmona (@danipolymath) 's Twitter Profile Photo

24h after the launch our agent demoed Supersonik around 2k times. To give perspective, if an average AE is doing 5 demos per day, we would need a team of 400 to achieve that, and even then the demos would never be immediate

Gabriel Martín Blázquez (@gabrielmbmb_) 's Twitter Profile Photo

We launched Supersonik a few weeks ago and we're hiring engineers. If you love browser automation, LLMs, voice models, ultra-low latency, and serious infra, and you want to live in Barcelona, DM me.

elie (@eliebakouch) 's Twitter Profile Photo

We’re presenting our work on SmolLM2 tomorrow at COLM, and I guess it’s a good time to say that I’m really proud of what our science team at Hugging Face has accomplished with the smollm ecosystem 🤗 While going through the poster sessions this week and talking with researchers,

We’re presenting our work on SmolLM2 tomorrow at COLM, and I guess it’s a good time to say that I’m really proud of what our science team at <a href="/huggingface/">Hugging Face</a> has accomplished with the smollm ecosystem 🤗

While going through the poster sessions this week and talking with researchers,
Loubna Ben Allal (@loubnabenallal1) 's Twitter Profile Photo

Last week, I gave an talk at COLM about the SmolLM2 paper, covering the curation of FineMath, Stack-Edu, SmolTalk and our multi-stage training approach. Slides here: docs.google.com/presentation/d…

Andrej Karpathy (@karpathy) 's Twitter Profile Photo

Excited to release new repo: nanochat! (it's among the most unhinged I've written). Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single,

Excited to release new repo: nanochat!
(it's among the most unhinged I've written).

Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single,
Loubna Ben Allal (@loubnabenallal1) 's Twitter Profile Photo

After ~4 years building SOTA models & datasets, we're sharing everything we learned in ⚡The Smol Training Playbook We cover the full LLM cycle: designing ablations, choosing an architecture, curating data, post-training, and building solid infrastructure. We'll help you

After ~4 years building SOTA models &amp; datasets, we're sharing everything we learned in ⚡The Smol Training Playbook

We cover the full LLM cycle: designing ablations, choosing an architecture, curating data, post-training, and building solid infrastructure.

We'll help you
Supersonik (@supersonikai) 's Twitter Profile Photo

One of our customers has done +3.4k demos with Supersonik in <2 months. If a typical AE is doing 5 demo/day, we've done almost 3 years of work of an AE

One of our customers has done +3.4k demos with Supersonik in &lt;2 months. If a typical AE is doing 5 demo/day, we've done almost 3 years of work of an AE