In a Training Loop 🔄

417 167 1141

Maxime Labonne PRO

mlabonne

https://mlabonne.github.io/blog

AI & ML interests

Post-training, model editing, quantization

Recent Activity

new activity about 16 hours ago

LiquidAI/LFM2.5-1.2B-Instruct:Reproduction of evaluation scores

liked a model about 24 hours ago

lmstudio-community/LFM2-1.2B-MLX-bf16

liked a model about 24 hours ago

lmstudio-community/LFM2-1.2B-MLX-8bit

View all activity

Organizations

New activity in LiquidAI/LFM2.5-1.2B-Instruct about 16 hours ago

Reproduction of evaluation scores

🔥 1

#6 opened about 16 hours ago by

lino-levan

liked 2 models about 24 hours ago

lmstudio-community/LFM2-1.2B-MLX-bf16

Text Generation • 1B • Updated Jul 22, 2025 • 25.3k • 6

lmstudio-community/LFM2-1.2B-MLX-8bit

Text Generation • 0.3B • Updated Jul 22, 2025 • 25.7k • 4

liked a Space 1 day ago

LiquidAI/LFM2.5-1.2B-Instruct

⚡

LFM2.5 (1.2B) runs on Ollama using only a dual-core CPU

New activity in LiquidAI/LFM2.5-1.2B-Instruct 1 day ago

Liquid AI, You NEED to Make a 16B MoE Next!

❤️ 6

#5 opened 2 days ago by

tanyiades

commented on Uncensor any LLM with abliteration 1 day ago

It's a good question. Experimentally, it generalizes very well to other languages, but adding prompts in your target languages will always be more efficient and fine-grained.

upvoted an article 2 days ago

Article

Open Responses: What you need to know

3 days ago

•

commented on SmolLM-Smashed: Tiny Giants, Optimized for Speed 4 days ago

Great article, thanks!

upvoted an article 4 days ago

Article

SmolLM-Smashed: Tiny Giants, Optimized for Speed

5 days ago

•

liked a dataset 4 days ago

MadlabOSS/synpaca

Viewer • Updated 6 days ago • 64.8k • 17 • 1

upvoted a collection 5 days ago

Jan 5 Releases

Collection

35 items • Updated 5 days ago • 5

updated a Space 5 days ago

LFM2.5-VL-1.6B WebGPU

🧠

In-browser vision-language inference with LFM2.5-VL-1.6B

liked 3 models 6 days ago

updated a model 7 days ago

LiquidAI/LFM2.5-VL-1.6B

Image-Text-to-Text • 2B • Updated 1 day ago • 7.49k • 178

liked a dataset 8 days ago

openmed-community/MedReason-Stenographic

Viewer • Updated 9 days ago • 31.5k • 50 • 50

updated a collection 9 days ago

💧 LFM2

Collection

LFM2 is a new generation of hybrid models, designed for on-device deployment. • 27 items • Updated 6 days ago • 136

liked a model 9 days ago

LiquidAI/LFM2-8B-A1B-ONNX

Text Generation • Updated 9 days ago • 8 • 4

published a model 9 days ago