convaiinnovations/medgemma-ecg-gguf-only
Quantization (fewer layers) vs convaiinnovations/medgemma-4b-ecginstruct. convaiinnovations/medgemma-ecg-gguf-only matches part of convaiinnovations/medgemma-4b-ecginstruct (444 of its 883 tensors; the rest of convaiinnovations/medgemma-4b-ecginstruct has no counterpart). On the shared part: convaiinnovations/medgemma-ecg-gguf-only is a GGUF quantization (Q4_K_M) of the same weights: 205 tensors match exactly and 239 within quantization error.
Tensors changed: Embeddings 0/1, Norms 0/205, MLP 0/102, Attention 0/136