martin r

kulminaator

13 5

AI & ML interests

None yet

Recent Activity

liked a model 6 days ago

poolside/Laguna-S-2.1

liked a model 7 days ago

poolside/Laguna-S-2.1-GGUF

liked a model 7 days ago

microsoft/Fara1.5-27B

View all activity

Organizations

None yet

liked a model 6 days ago

poolside/Laguna-S-2.1

Text Generation • 118B • Updated 10 days ago • 86.5k • • 931

liked 2 models 7 days ago

#4 opened about 1 month ago by

pedroecfernandes

liked a model 25 days ago

mistralai/Leanstral-1.5-119B-A6B

Updated 22 days ago • 381 • 211

New activity in bartowski/deepreinforce-ai_Ornith-1.0-35B-GGUF about 1 month ago

IQ4_XS is broken

#1 opened about 1 month ago by

QBANIN

Anyone got EAGLE3 GGUF's for speculation for this one?

#3 opened about 1 month ago by

kulminaator

New activity in LiquidAI/LFM2.5-8B-A1B 2 months ago

Can't follow instructions

#8 opened 2 months ago by

pcomte

This Model Has Virtually No Broad Knowledge

👍 2

#2 opened 2 months ago by

phil111

New activity in unsloth/LFM2.5-8B-A1B-GGUF 2 months ago

The current unsloth gguf's have broken chat template

👍 2

#1 opened 2 months ago by

kulminaator

New activity in Qwen/Qwen3.6-27B-FP8 3 months ago

Performance report on RTX 4090D (48GB VRAM): 40 t/s

👍 1

#11 opened 3 months ago by

SlavikF

reacted to danielhanchen's post with ❤️ 3 months ago

Post

6007

We’re excited to announce that Unsloth has joined the PyTorch Ecosystem! 🔥🦥

Unsloth is an open-source project that makes training & running models more accurate and faster with less compute. Our mission is to make local AI accessible to everyone. Thanks to all of you for making this possible! 💕

Blog: https://unsloth.ai/blog/pytorch
GitHub: https://github.com/unslothai/unsloth

2 replies

replied to DavidAU's post 3 months ago

I will add to the list; may wait for specific Heretic and/or tuned version.

I already have a 43B-A3B version running in the lab ; however tuning these sparse moe models take a lot more work/time and ahh... detail. AND a lot more VRAM!!! [can't compress these atm, so BF16 required => 100 GB+ ]

What hardware do you even use for it and how long does the retraining + quants generation take?