nisten/shqiponja-90

Partial fine-tune vs Qwen/Qwen1.5-72B. Partial change: 6% of the weights changed (attention, mlp, norms), in 5 of 80 layers.

Tensors changed: Embeddings 0/1, MLP 15/240, Attention 35/560, Norms 10/161, lm_head 0/1

2 derivatives checked

• mradermacher/shqiponja-90-i1-GGUF: Quantization
• mradermacher/shqiponja-90-GGUF: Quantization

On the Hub · Downloads on Model Pulse