secfa/
DeepSeek-R1-UD-Q2_K_XL

2,029 7 months ago

Unsloth's DeepSeek-R1 , I just merged the thing and uploaded it here. This is the full 671b model. MoE Bits:2.51bit Type:UD-Q2_K_XL Disk Size:212GB Accuracy:Best Details:MoE all 2.5bit. down_proj in MoE mixture of 3.5/2.5bit

Models

View all →

2 models

DeepSeek-R1-UD-Q2_K_XL:latest

227GB · 4K context window · Text · 7 months ago

DeepSeek-R1-UD-Q2_K_XL:24g

227GB · 4K context window · Text · 7 months ago

Readme

tag latest ：The graphics card is not supported for both ollama and screen display at the same time

tag 24g for 24g vRAM: 21g model + 1.8g fp16 kv cache + windows system reserve