mradermacher/gemma-3-4b-it-Ko-Reasoning-GGUF

Same weights, converted (fewer layers) vs DimensionSTP/gemma-3-4b-it-Ko-Reasoning. mradermacher/gemma-3-4b-it-Ko-Reasoning-GGUF matches part of DimensionSTP/gemma-3-4b-it-Ko-Reasoning (444 of its 883 tensors; the rest of DimensionSTP/gemma-3-4b-it-Ko-Reasoning has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Embeddings 0/1, Norms 0/205, MLP 0/102, Attention 0/136

On the Hub · Downloads on Model Pulse