Pruned Falcon-11B variant optimized for 13 European languages through layer selection.
S. Smits
ssmits
AI & ML interests
Synthetic data, on-premise / local language-specific LLMs, optimizing embeddings models retrieval
Recent Activity
liked
a model
about 4 hours ago
nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
liked
a Space
5 days ago
Tevatron/BrowseComp-Plus
liked
a model
6 days ago
mistralai/Devstral-2-123B-Instruct-2512
Organizations
Strategic merging of language models through layer-level architecture optimization.
-
ssmits/Llama-3.1-Nemotron-92B-Instruct-HF-early
Text Generation • 92B • Updated • 57 • 2 -
ssmits/Llama-3.1-Nemotron-92B-Instruct-HF-late
Text Generation • 92B • Updated • 36 • 2 -
ssmits/Qwen2.5-95B-Instruct
Text Generation • 95B • Updated • 28 • 4 -
ssmits/Qwen2.5-125B-Instruct
Text Generation • 125B • Updated • 12
Pruned Falcon-11B variant optimized for 13 European languages through layer selection.
Tokenized Dutch datasets based on BramVanroy/occiglot-fineweb-v0.5-nl
Strategic merging of language models through layer-level architecture optimization.
-
ssmits/Llama-3.1-Nemotron-92B-Instruct-HF-early
Text Generation • 92B • Updated • 57 • 2 -
ssmits/Llama-3.1-Nemotron-92B-Instruct-HF-late
Text Generation • 92B • Updated • 36 • 2 -
ssmits/Qwen2.5-95B-Instruct
Text Generation • 95B • Updated • 28 • 4 -
ssmits/Qwen2.5-125B-Instruct
Text Generation • 125B • Updated • 12