blascotobasco/GigaChat3.1-Lightning-Uncensored
LoRA-style vs ai-sage/GigaChat3.1-10B-A1.8B-bf16. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, o_proj, q_proj, up_proj changed, in 26 of 26 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/79, MLP 2609/2696, Attention 52/104
3 derivatives checked
- mradermacher/GigaChat3.1-Lightning-Uncensored-i1-GGUF: Quantization
- mradermacher/GigaChat3.1-Lightning-Uncensored-GGUF: Same weights, converted
- blascotobasco/GigaChat3.1-Lightning-Uncensored-GGUF: Same weights, converted