Delta-Vector/Francois-PE-V2-Huali-12B
LoRA-style vs Delta-Vector/Francois-PE-V2. 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 40 of 40 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/81, MLP 120/120, Attention 160/160
10 derivatives checked
- mradermacher/Francois-PE-V2-Huali-12B-i1-GGUF: Quantization
- DevQuasar/Delta-Vector.Francois-PE-V2-Huali-12B-GGUF: Same weights, converted
- mradermacher/Francois-PE-V2-Huali-12B-GGUF: Quantization
- Delta-Vector/Archaeo-12B-V2: Full fine-tune
- yamatazen/FusionEngine-12B: LoRA-style
- DoppelReflEx/LilithCore-v1-12B: Full fine-tune
- yamatazen/NeonMaid-12B: Full fine-tune
- DoppelReflEx/LilithCore-v0.9-12B: Full fine-tune
- DreadPoor/Ward-12B-Model_Stock: Full fine-tune
- NewEden-Forge/archaeo-12b-v2-rei-base-arcee-fusion-merge: Full fine-tune