Kwoya/Spyra-20B-v.1.1

Full fine-tune vs unsloth/gpt-oss-20b-BF16. Fine-tune: 94% of the weights changed, across 24 of 24 layers.

Tensors changed: Embeddings 0/1, Norms 0/49, MLP 48/144, Attention 96/216, lm_head 0/1

2 derivatives checked

• mradermacher/Spyra-20B-v.1.1-i1-GGUF: Quantization
• mradermacher/Spyra-20B-v.1.1-GGUF: Quantization

On the Hub · Downloads on Model Pulse