baa.ai

company

black-sheep-ai-baa

AI & ML interests

Model Quantization

Recent Activity

trevk updated a Space 2 days ago

baa-ai/sakd-knowledge-distillation

trevk published a Space 2 days ago

baa-ai/sakd-knowledge-distillation

trevk updated a Space 2 days ago

baa-ai/sensitivity-aware-training

View all activity

trevk

updated a Space 2 days ago

SAKD: SWAN-Guided Knowledge Distillation

Generate quantization‑ready student models via guided distillation

trevk

published a Space 2 days ago

SAKD: SWAN-Guided Knowledge Distillation

Generate quantization‑ready student models via guided distillation

trevk

updated a Space 2 days ago

Sensitivity-Aware Training (SAT)

Train LLMs to be quantization‑ready with sensitivity‑aware methods

trevk

published a Space 2 days ago

Sensitivity-Aware Training (SAT)

Train LLMs to be quantization‑ready with sensitivity‑aware methods

trevk

updated 2 Spaces 3 days ago

README

Swan Paper

Quantize LLMs without data using per‑tensor mixed precision

trevk

updated a model 6 days ago

baa-ai/Qwen3.5-397B-A17B-SWAN-3bit

Text Generation • 396B • Updated 6 days ago • 656

trevk

published a model 6 days ago

baa-ai/Qwen3.5-397B-A17B-SWAN-3bit

Text Generation • 396B • Updated 6 days ago • 656

trevk

in baa-ai/Qwen3.5-397B-A17B-SWAN-4bit 6 days ago

Update MMLU-Pro: 77.1% with thinking enabled (+5.0 vs non-thinking)

#1 opened 6 days ago by

trevk

published a Space 7 days ago

Swan Paper

Quantize LLMs without data using per‑tensor mixed precision

trevk

updated a model 7 days ago

baa-ai/Qwen3.5-397B-A17B-SWAN-4bit

Text Generation • 396B • Updated 6 days ago • 1.07k • 3

trevk

published a model 7 days ago

baa-ai/Qwen3.5-397B-A17B-SWAN-4bit

Text Generation • 396B • Updated 6 days ago • 1.07k • 3

trevk

published a Space 9 days ago

README