secfa/
DeepSeek-R1-UD-IQ1_M:24g

4,296 7 months ago

Unsloth's DeepSeek-R1 , I just merged the thing and uploaded it here. This is the full 671b model. MoE Bits:1.73bit Type:UD-IQ1_M Disk Size:158GB Accuracy:Good Details:MoE all 1.56bit. down_proj in MoE left at 2.06bit

... /
params
75e2da1529b1 · 161B
{
"num_gpu": 14,
"stop": [
"<｜begin▁of▁sentence｜>",
"<｜end▁of▁sentence｜>",
"<｜User｜>",
"<｜Assistant｜>"
]
}