mradermacher/Gemma-3-Earthen-v0.2-4B-GGUF

Same weights, converted (fewer layers) vs PJMixers-Dev/Gemma-3-Earthen-v0.2-4B. mradermacher/Gemma-3-Earthen-v0.2-4B-GGUF matches part of PJMixers-Dev/Gemma-3-Earthen-v0.2-4B (444 of its 883 tensors; the rest of PJMixers-Dev/Gemma-3-Earthen-v0.2-4B 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