mradermacher/Shako-4B-it-GGUF
LoRA-style (fewer layers) vs anaspro/Shako-iraqi-4B-it. mradermacher/Shako-4B-it-GGUF matches part of anaspro/Shako-iraqi-4B-it (809 of its 1676 tensors; the rest of anaspro/Shako-iraqi-4B-it has no counterpart). On the shared part: mradermacher/Shako-4B-it-GGUF is a GGUF (F16) of different weights. 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 35 of 35 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, other 0/246, Norms 0/317, MLP 105/105, Attention 140/140