paulml/OmniBeagleSquaredMBX-v3-7B

Full fine-tune vs flemmingmiguel/MBX-7B-v3. Fine-tune: 94% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/1, Norms 65/65, MLP 90/96, Attention 120/128, lm_head 1/1

2 derivatives checked

• mradermacher/OmniBeagleSquaredMBX-v3-7B-GGUF: Quantization
• mradermacher/OmniBeagleSquaredMBX-v3-7B-i1-GGUF: Quantization

On the Hub · Downloads on Model Pulse