Official UniQL models
AI & ML interests
Energy-aware Computing, Low Power Design, EDA, Dark Silicon, Efficient Deep Learning
Papers
UniQL: Unified Quantization and Low-rank Compression for Adaptive Edge LLMs
Quamba2: A Robust and Scalable Post-training Quantization Framework for Selective State Space Models
models
43
ut-enyac/mamba2-8b-converted-uniql-1.0-masked-lora-rft-w4a16
0.1B
•
Updated
•
4
ut-enyac/Qwen2.5-7B-uniql-1.0-masked-lora-rft-w4a16
1B
•
Updated
ut-enyac/Bamba-9B-v2-uniql-1.0-masked-lora-rft-w4a16
67.4M
•
Updated
ut-enyac/Nemotron-H-8B-Base-8K-uniql-1.0-masked-lora-rft-w4a16
68.8M
•
Updated
ut-enyac/Llama-3.1-8B-uniql-1.0-masked-lora-rft-w4a16
65.9M
•
Updated
ut-enyac/Llama-2-7b-hf-uniql-1.0-masked-lora-rft-w4a16
0.9B
•
Updated
ut-enyac/quamba2-8b-converted-w4aX
Text Generation
•
Updated
ut-enyac/quamba-chat-w4a8
Text Generation
•
Updated
•
4
ut-enyac/quamba2-2.7b-w4a8
Text Generation
•
Updated
•
5
•
1
ut-enyac/quamba2-8b-converted-w4a8
Text Generation
•
Updated
•
11
•
1
datasets
0
None public yet