Hannu Varjoranta
varjoranta
·
AI & ML interests
Weight and KV cache compression for production LLM serving. Building turboquant-plus-vllm.
Recent Activity
updated a model 9 days ago
varjosoft/DeepSeek-V4-Flash-TQ3-native published a model 9 days ago
varjosoft/DeepSeek-V4-Flash-TQ3-native updated a model 9 days ago
varjosoft/Qwen3.6-27B-TQ3-native