Locutusque/TinyMistral-248M-v2.5-Instruct

Full fine-tune vs Locutusque/TinyMistral-248M-v2.5. Fine-tune: 100% of the weights changed, across 12 of 12 layers.

Tensors changed: lm_head 1/1, Embeddings 1/1, Norms 25/25, MLP 36/36, Attention 48/48

3 derivatives checked

• DedeProGames/Chesser-248K-Mini: Full fine-tune
• MaziyarPanahi/TinyMistral-248M-v2.5-Instruct-GGUF: Quantization
• heegyu/TinyMistral-248M-v2.5-Instruct-orpo: Full fine-tune

On the Hub · Downloads on Model Pulse