Hcompany/Holo4-27B
Full fine-tune vs Qwen/Qwen3.8-27B. Fine-tune: 98% of the weights changed, across 64 of 64 layers.
Tensors changed: Norms 142/319, Attention 374/556, MLP 192/304, Embeddings 1/4, lm_head 1/1
9 derivatives checked
- mradermacher/Holo4-27B-i1-GGUF: Quantization
- Hcompany/Holo4-27B-GGUF: Quantization
- mradermacher/Holo4-27B-GGUF: Quantization
- Hcompany/Holo4-27B-FP8: Quantization
- Hcompany/Holo4-27B-NVFP4: Quantization
- DevQuasar/Hcompany.Holo4-27B-GGUF: Quantization
- abenzerps/Holo4-27B-GGUF: Quantization
- arcitech-psp/Holo4-27B-AutoRound-W4A16: Quantization
- timteh673/Holo4-27B-Abliterix: Abliterated