SupraLabs/Supra-50M-Instruct

Full fine-tune vs SupraLabs/Supra-50M-Base. Fine-tune: 100% of the weights changed, across 12 of 12 layers.

Tensors changed: Embeddings 1/1, Norms 1/25, MLP 36/36, Attention 48/48

1 derivatives checked

• di-zhang-fdu/Supra-50M-Instruct-GGUF: Quantization

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse