Delta-Vector/Austral-Xgen-9B-Base
LoRA-style vs Salesforce/xgen-small-9B-instruct-r. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 45 of 45 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/91, MLP 135/135, Attention 180/180, lm_head 0/1
2 derivatives checked
- mradermacher/Austral-Xgen-9B-Base-i1-GGUF: Quantization
- mradermacher/Austral-Xgen-9B-Base-GGUF: Same weights, converted