ermiaazarkhalili/Gemma4-E2B-SFT-Claude-Opus-Reasoning-Unsloth-GGUF
Quantization (fewer layers) vs ermiaazarkhalili/Gemma4-E2B-SFT-Claude-Opus-Reasoning-Unsloth. ermiaazarkhalili/Gemma4-E2B-SFT-Claude-Opus-Reasoning-Unsloth-GGUF matches part of ermiaazarkhalili/Gemma4-E2B-SFT-Claude-Opus-Reasoning-Unsloth (565 of its 2011 tensors; the rest of ermiaazarkhalili/Gemma4-E2B-SFT-Claude-Opus-Reasoning-Unsloth has no counterpart). On the shared part: ermiaazarkhalili/Gemma4-E2B-SFT-Claude-Opus-Reasoning-Unsloth-GGUF is a GGUF quantization (Q8_0) of the same weights: 248 tensors match exactly and 317 within quantization error.
Tensors changed: Embeddings 0/2, Norms 0/247, MLP 0/105, other 0/71, Attention 0/140