Nf4 targetted layer approach (ovedrive) Not tested yet.

Text encoder was untouched but is has gone from 126GB model repo to less than 30GB of which the TE is 11GB (CPU offload recommended)

Downloads last month
40
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for ovedrive/Wan2.2-I2V-A14B-4bit

Quantized
(49)
this model