AlicanKiraz0/Kizagan-TTS-v1.0

LoRA-style vs openbmb/VoxCPM2. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, o_proj, q_proj, v_proj changed, in 24 of 28 layers; norms, biases and embeddings are untouched.

Tensors changed: Embeddings 0/1, Norms 0/124, MLP 0/188, Attention 192/240, other 0/24

1 derivatives checked

On the Hub · Downloads on Model Pulse