sail/Sailor2-L-8B

Full fine-tune vs sail/Sailor2-8B. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/1, Norms 59/65, MLP 96/96, Attention 224/224, lm_head 1/1

4 derivatives checked

• mradermacher/Sailor2-L-8B-GGUF: Same weights, converted
• QuantFactory/Sailor2-L-8B-Chat-GGUF: Full fine-tune
• sail/Sailor2-L-8B-Chat: Full fine-tune
• sail/Sailor2-L-8B-SFT: Full fine-tune

On the Hub · Downloads on Model Pulse