TalTechNLP/Voxtral-Mini-3B-2507-estonian

Full fine-tune vs mistralai/Voxtral-Mini-3B-2507. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 2/2, other 6/6, Norms 119/191, MLP 218/218, Attention 344/344, lm_head 1/1

1 derivatives checked

• mradermacher/Voxtral-Mini-3B-2507-estonian-GGUF: Same weights, converted (fewer layers)

On the Hub · Downloads on Model Pulse