allura-org/MoE-Girl_400MA_1BT

Full fine-tune vs ibm-granite/granite-3.0-1b-a400m-base. Fine-tune: 100% of the weights changed, across 24 of 24 layers.

Tensors changed: other 72/72, Norms 21/49, Attention 96/96

2 derivatives checked

• mradermacher/MoE-Girl_400MA_1BT-GGUF: Same weights, converted
• DevQuasar-5/allura-org.MoE-Girl_400MA_1BT-GGUF: Quantization

On the Hub · Downloads on Model Pulse