KETI-NLP/Qwen3.5-KETI-HAECHI-27B

Full fine-tune vs Qwen/Qwen3.5-27B. Fine-tune: 94% of the weights changed, across 64 of 64 layers.

Tensors changed: lm_head 1/1, Embeddings 0/4, MLP 197/304, Attention 354/556, Norms 111/319

2 derivatives checked

• mradermacher/Qwen3.5-KETI-HAECHI-27B-i1-GGUF: Quantization
• mradermacher/Qwen3.5-KETI-HAECHI-27B-GGUF: Quantization

On the Hub · Downloads on Model Pulse