JJarvinen/chatterbox-finnish-nano
Partial fine-tune vs ResembleAI/chatterbox-nano. Partial change: 43% of the weights changed (attention, embeddings, mlp, norms, other), in 2 of 6 layers.
Tensors changed: other 382/1106, Norms 266/814, MLP 72/108, Attention 390/450, Embeddings 11/11
1 derivatives checked
- JJarvinen/chatterbox-finnish-nano-GGUF: Not comparable yet