Shears: Unstructured Sparsity with Neural Low-rank Adapter Search
Paper • 2404.10934 • Published
Shears
Shears Models (Shears: Unstructured Sparsity with Neural Low-rank Adapter Search)
-
-
Low-Rank Adapters Meet Neural Architecture Search for LLM Compression
Paper • 2501.16372 • Published • 11 -
IntelLabs/shears-llama-7b-50-math-super-adapter
Updated • 6 • 3
IntelLabs/shears-llama-7b-50-math-heuristic-adapter
Updated • 12 • 3Note Shears fine-tuned models for Llama-7B on Math Instruction Tuning (50% sparsity)
-
IntelLabs/shears-llama-7b-50-cs-super-adapter
Updated • 10 • 3
IntelLabs/shears-llama-7b-50-cs-heuristic-adapter
Updated • 12 • 3Note Shears fine-tuned models for Llama-7B on CS Instruction Tuning (50% sparsity)
-
IntelLabs/shears-llama-13b-50-math-super-adapter
Updated • 6 • 4
IntelLabs/shears-llama-13b-50-math-heuristic-adapter
Updated • 14 • 3Note Shears fine-tuned models for Llama-13B on Math Instruction Tuning (50% sparsity)
IntelLabs/shears-mpt-7b-50-base
Text Generation • 7B • Updated • 25 • 2Note Shears base models (sparse via Wanda) for MPT-7B
-
IntelLabs/shears-mpt-7b-50-gsm8k-super-adapter
Updated • 7 • 3
IntelLabs/shears-mpt-7b-50-gsm8k-heuristic-adapter
Updated • 10 • 3Note Shears fine-tuned models for MPT-7B on GSM8K (50% sparsity)