QuantStack/SkyReels-V2-T2V-14B-720P-VACE-GGUF

Same weights, converted vs Skywork/SkyReels-V2-T2V-14B-720P. Same weights stored in another dtype (F32 → F16/F32), plus 236 new tensors.

Tensors changed: Attention 0/640, Norms 0/240, MLP 0/160, other 0/45, Embeddings 0/10

On the Hub · Downloads on Model Pulse