codegood/MPhi_Latest

Partial fine-tune vs microsoft/phi-1_5. Partial change: 29% of the weights changed (attention, lm_head), in 24 of 24 layers.

Tensors changed: lm_head 2/2, Embeddings 0/1, Norms 0/50, MLP 0/96, Attention 72/192

2 derivatives checked

• mradermacher/MPhi_Latest-i1-GGUF: Quantization
• mradermacher/MPhi_Latest-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse