Lewis Tunstall's banner
Lewis Tunstall's profile picture

Lewis Tunstall

@_lewtun • 20,205 subscribers

🤠 post-training @huggingface

Shorts

If you're using GRPO in TRL, you should really switch to the new async trainer. In our benchmarks, it's ~2-4x faster 🔥

If you're using GRPO in TRL, you should really switch to the new async trainer. In our benchmarks, it's ~2-4x faster 🔥

29,624 Aufrufe

The ML Intern can now ping you on Slack when it's finished training models, generating datasets, or compiling an analysis of the 1000 ablations you launched 💸

The ML Intern can now ping you on Slack when it's finished training models, generating datasets, or compiling an analysis of the 1000 ablations you launched 💸

39,605 Aufrufe

Videos

LIVE
1.2k

Anya Rossi

sweetdream.ai

SweetDream.aiSponsored•Livecam
Streaming Now

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

HD live stream
Exclusive private shows
1.2k viewers online

Free preview available • Premium content

The recording from my CERN lecture on training LLMs is now online! You can follow along with the slides here: I recommend watching it with 2x speed because my Aussie drawl is a little slow 😃
1:33:30
_lewtun's profile picture

The recording from my CERN lecture on training LLMs is now online! You can follow along with the slides here: I recommend watching it with 2x speed because my Aussie drawl is a little slow 😃

Lewis Tunstall

21,982 Aufrufe • vor 4 Tagen

You can now have an AI researcher running on your laptop 24/7 for free! Running Qwen3-35B-A3B with llama.cpp and a 4-bit quant from Unsloth
0:30
_lewtun's profile picture

You can now have an AI researcher running on your laptop 24/7 for free! Running Qwen3-35B-A3B with llama.cpp and a 4-bit quant from Unsloth

Lewis Tunstall

118,729 Aufrufe • vor 3 Monaten

We just added speculative decoding for Carbon, so you can now generate DNA sequences 2-3x faster with llama.cpp This enables anyone to process huge genomes locally! GGUF weights and quants:
0:20
_lewtun's profile picture

We just added speculative decoding for Carbon, so you can now generate DNA sequences 2-3x faster with llama.cpp This enables anyone to process huge genomes locally! GGUF weights and quants:

Lewis Tunstall

16,007 Aufrufe • vor 3 Monaten

Chatting to CNN about why DeepSeek-R1 and the open source community is awesome was not on my bingo card this morning (I’ve never been on TV before 😂)
5:10
_lewtun's profile picture

Chatting to CNN about why DeepSeek-R1 and the open source community is awesome was not on my bingo card this morning (I’ve never been on TV before 😂)

Lewis Tunstall

33,925 Aufrufe • vor 1 Jahr

Keine weiteren Inhalte verfügbar