mradermacher/medgemma-4b-pt-CPT-SFT-GGUF

Same weights, converted (fewer layers) vs medAdapt/medgemma-4b-pt-CPT-SFT. mradermacher/medgemma-4b-pt-CPT-SFT-GGUF matches part of medAdapt/medgemma-4b-pt-CPT-SFT (444 of its 883 tensors; the rest of medAdapt/medgemma-4b-pt-CPT-SFT has no counterpart). On the shared part: Same weights stored in another dtype (F32 → F16/F32).

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

On the Hub · Downloads on Model Pulse