SmolLM-Smashed is a collection of optimized language models. Each model is quantized and compiled for maximum efficiency while preserving performance.
Parag Ekbote
AINovice2005
AI & ML interests
ML Engineer passionate about taking models from research to production. 1 year supporting tech startups. Active OSS contributor.
Recent Activity
reacted
to
danielhanchen's
post
with 🔥
about 21 hours ago
We collaborated with Hugging Face to enable you to train MoE models 12× faster with 35% less VRAM via our new Triton kernels (no accuracy loss). 🤗
Train gpt-oss locally on 12.8GB VRAM with our free notebooks: https://unsloth.ai/docs/new/faster-moe
upvoted
an
article
8 days ago
Training Design for Text-to-Image Models: Lessons from Ablations
upvoted
an
article
15 days ago
Unlocking Agentic RL Training for GPT-OSS: A Practical Retrospective
Organizations
models
15
AINovice2005/SmolLM3-3B-smashed
Text Generation
•
Updated
•
2
AINovice2005/SmolLM2-1.7B-smashed
Text Generation
•
Updated
AINovice2005/SmolLM2-360M-smashed
Text Generation
•
Updated
•
2
AINovice2005/SmolLM-360M-smashed
Text Generation
•
Updated
AINovice2005/SmolLM-135M-smashed
Text Generation
•
Updated
•
1
AINovice2005/Voxtral-Mini-3B-2507-smashed
Audio-Text-to-Text
•
Updated
AINovice2005/medgemma-4b-it-smashed
Image-Text-to-Text
•
4B
•
Updated
•
2
AINovice2005/FLUX.1-schnell-smashed
Text-to-Image
•
Updated
AINovice2005/Sana_600M_ControlNet_HED-smashed
Text-to-Image
•
Updated
AINovice2005/quantized-GLM-4.1V-9B-Thinking
Image-Text-to-Text
•
Updated
•
17
datasets
5
AINovice2005/cicflow-ids-binary
Viewer
•
Updated
•
1.05M
•
10
AINovice2005/cicflow-ids-multiclass
Viewer
•
Updated
•
1.05M
•
71
AINovice2005/toy-optuna-hpo-dataset
Viewer
•
Updated
•
5
•
3
AINovice2005/EldenRing_Small
Viewer
•
Updated
•
49.8k
•
4
•
2
AINovice2005/Notus_Aegis-v1
Viewer
•
Updated
•
590
•
26