Lambent/Mira-1.9-Alternate-27B

Full fine-tune vs google/gemma-3-27b-pt. Fine-tune: 99% of the weights changed, across 62 of 62 layers.

Tensors changed: Norms 384/484, MLP 195/294, Attention 261/464, other 1/1, Embeddings 0/3

2 derivatives checked

• mradermacher/Mira-1.9-Alternate-27B-i1-GGUF: Quantization (fewer layers)
• mradermacher/Mira-1.9-Alternate-27B-GGUF: Quantization (fewer layers)

On the Hub · Downloads on Model Pulse