google/gemma-3-27b-it-qat-q4_0-unquantized
Full fine-tune vs google/gemma-3-27b-it. Fine-tune: 98% of the weights changed, across 62 of 62 layers.
Tensors changed: Embeddings 1/4, Norms 373/484, MLP 186/294, Attention 248/464, other 1/1
9 derivatives checked
- unsloth/gemma-3-27b-it-qat-GGUF: Same weights, converted (fewer layers)
- bartowski/google_gemma-3-27b-it-qat-GGUF: Same weights, converted (fewer layers)
- mlabonne/gemma-3-27b-it-qat-abliterated-GGUF: Quantization (fewer layers)
- DevQuasar-12/google.gemma-3-27b-it-qat-q4_0-unquantized-GGUF: Same weights, converted (fewer layers)
- dahara1/gemma-3-27b-it-qat-japanese-imatrix: Quantization (fewer layers)
- leon-se/gemma-3-27b-it-qat-W4A16-G128: Quantization
- soob3123/amoral-gemma3-27B-v2-qat: LoRA-style
- mlabonne/gemma-3-27b-it-qat-abliterated: Abliterated (fewer layers)
- ConicCat/Gemma-3-Fornax-V4-27B-QAT: LoRA-style
1 other repos hold byte-for-byte identical weight files.