sail/Sailor-7B
Full fine-tune vs Qwen/Qwen1.5-7B. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, Norms 65/65, MLP 96/96, Attention 224/224, lm_head 1/1
5 derivatives checked
- mradermacher/Sailor-7B-GGUF: Same weights, converted
- sail/Sailor-7B-Chat-gguf: Full fine-tune
- tensorblock/Sailor-7B-GGUF: Quantization
- sail/Sailor-7B-Chat: Full fine-tune
- hoangcung165/Sailor-7B-Metal-Healt: LoRA-style