Bifrost-AI/Phi-4-bifrost-sol-3.8B

Full fine-tune vs microsoft/Phi-4-mini-instruct. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/1, Norms 10/65, MLP 64/64, Attention 64/64

3 derivatives checked

• mradermacher/Phi-4-bifrost-sol-3.8B-i1-GGUF: Full fine-tune
• Bifrost-AI/Phi-4-bifrost-sol-3.8B-gguf: Same weights, converted
• mradermacher/Phi-4-bifrost-sol-3.8B-GGUF: Full fine-tune

On the Hub · Downloads on Model Pulse