❄️ If you lock in you can do anything.

Christina Emilie Sørensen PRO

cafkafk

·

https://point.free

AI & ML interests

I'm not an ML person! I just like to learn...

Recent Activity

repliedto their post about 18 hours ago

I'm working on a local model that can be run on stuff like an an RTX 3080 at 60 t/s. I'm focused on making it good at Rust and Nix. It's... not done yet, but doing great already. Had a 28/100 on livebench v6 earlier, which obviously isn't amazing. But for my first real model, very exciting. Can't wait to share more as I get closer to the finish line in the coming weeks.

reacted to pollix's post with 🔥 about 21 hours ago

First stuntd model is on the Hub :) pollix/stuntd-support-triage is three small heads on the Laya encoder that triage a support ticket in one request: category, urgency and needs_human. About 50 MB each, all three answers come back at a p50 of 71ms through the daemon. On 1,000 tickets they never saw, each head answers on its own when it's sure: category 99.9%, needs_human 92%, urgency 76%. A ticket only skips the big model when all three are sure, that's 72.7% of them, and all three are right on 97.1% of those. It's the support demo from the repo, so the tickets are generated and the teacher is a rule. The point is to show what a head looks like and how fast it is, then you train the same thing on your own traffic with your own LLM as the teacher. ``` hf download pollix/stuntd-support-triage --local-dir support-heads ``` Model: https://huggingface.co/pollix/stuntd-support-triage Everything in one place: https://huggingface.co/collections/pollix/stuntd-6abe0a33303828e10c72ab41 Code: https://github.com/bladedevoff/stuntd

posted an update 2 days ago

I'm working on a local model that can be run on stuff like an an RTX 3080 at 60 t/s. I'm focused on making it good at Rust and Nix. It's... not done yet, but doing great already. Had a 28/100 on livebench v6 earlier, which obviously isn't amazing. But for my first real model, very exciting. Can't wait to share more as I get closer to the finish line in the coming weeks.

View all activity

Organizations

None yet

liked a model 13 days ago

Alibaba-NLP/Tongyi-DeepResearch-30B-A3B

Text Generation • 31B • Updated Oct 10, 2025 • 6.51k • 823

liked 2 models about 1 month ago

Eriskii/LFM2.5-8B-A1B-Multichannel

Text Generation • 8B • Updated Jun 14 • 132 • 13

QuantTrio/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-v2-AWQ

Image-Text-to-Text • 28B • Updated Mar 30 • 108k • 14

liked a model 4 months ago

google/gemma-4-12B

Any-to-Any • 12B • Updated Jul 15 • 117k • 759

liked a Space 5 months ago

Open ASR Leaderboard

Explore ASR model performance across languages and datasets

liked 15 models 5 months ago

ibm-granite/granite-speech-4.1-2b-nar

Feature Extraction • 2B • Updated Jun 18 • 89.6k • 61

cafkafk/gemma-4-31B-it-assistant-GGUF-noimatrix

0.5B • Updated May 8 • 590 • 8

cafkafk/gemma-4-26B-A4B-it-assistant-GGUF-noimatrix

0.4B • Updated May 8 • 955 • 2

cafkafk/gemma-4-E4B-it-assistant-GGUF-noimatrix

78.8M • Updated May 8 • 507 • 3

cafkafk/gemma-4-E2B-it-assistant-GGUF-noimatrix

78M • Updated May 8 • 519 • 2

google/alphagenome-all-folds

Updated Jan 27 • 129

google/alphagenome-fold-3

Updated Jan 27 • 23

google/alphagenome-fold-2

Updated Jan 27 • 13

google/alphagenome-fold-1

Updated Jan 27 • 12

google/alphagenome-fold-0

Updated Jan 27 • 13

google/medasr

Automatic Speech Recognition • 0.1B • Updated May 26 • 20.9k • 369

google/gemma-4-E2B-it

Any-to-Any • 5B • Updated Jul 20 • 3.03M • 996

google/gemma-4-E4B-it

Any-to-Any • 8B • Updated Jul 20 • 4.38M • 1.64k

google/gemma-4-26B-A4B-it

Image-Text-to-Text • 26B • Updated Jul 20 • 13.1M • • 1.58k

google/gemma-4-26B-A4B-it-assistant

Any-to-Any • 0.4B • Updated Jul 15 • 128k • 186