tensopolis/virtuoso-small-v2-tensopolis-v1
LoRA-style vs arcee-ai/Virtuoso-Small-v2. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj changed, in 8 of 48 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/97, MLP 0/144, Attention 8/336, lm_head 0/1
1 derivatives checked
- mradermacher/virtuoso-small-v2-tensopolis-v1-GGUF: Full fine-tune