meditsolutions/MSH-v1-Bielik-v2.3-Instruct-MedIT-merge
Full fine-tune vs speakleash/Bielik-11B-v2.3-Instruct. Fine-tune: 99% of the weights changed, across 50 of 50 layers.
Tensors changed: lm_head 1/1, Embeddings 0/1, Norms 35/101, MLP 150/150, Attention 200/200
3 derivatives checked
- meditsolutions/MSH-Lite-7B-v1-Bielik-v2.3-Instruct-Llama-Prune: Same weights, repacked
- mradermacher/MSH-v1-Bielik-v2.3-Instruct-MedIT-merge-i1-GGUF: Quantization
- mradermacher/MSH-v1-Bielik-v2.3-Instruct-MedIT-merge-GGUF: Quantization