Nanbeige/Nanbeige4.1-3B
Full fine-tune vs Nanbeige/Nanbeige4-3B-Base. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: lm_head 1/1, Embeddings 1/1, Norms 65/65, MLP 96/96, Attention 128/128
24 derivatives checked
- Edge-Quant/Nanbeige4.1-3B-Q8_0-GGUF: Quantization
- Edge-Quant/Nanbeige4.1-3B-Q4_K_M-GGUF: Quantization
- mradermacher/Nanbeige4.1-3B-GGUF: Same weights, converted
- RoroMachine14/nanbeige-4.1-3b-deepseek-distilled: LoRA-style
- seanbailey518/Nanbeige4.1-3B-GGUF: Same weights, converted
- mradermacher/Nanbeige4.1-3B-i1-GGUF: Quantization
- Open4bits/Nanbeige4.1-3B-GGUF: Quantization
- DevQuasar/Nanbeige.Nanbeige4.1-3B-GGUF: Same weights, converted
- SimplySara/Nanbeige4.1-3B-GGUF: Same weights, converted
- deltakitsune/Nanbeige-4.1-Python-DeepThink-3B: LoRA-style
- tantk/Nanbeige4.1-3B-GGUF: Same weights, converted
- SimplySara/Nanbeige4.1-3B-i1-GGUF: Quantization
- ThalisAI/Nanbeige4.1-3B-heretic: Abliterated
- mlx-community/Nanbeige4.1-3B-8bit: Quantization
- Akicou/Nanbeige4.1-3B-GGUF: Quantization
- prithivMLmods/Nanbeige4.1-3B-f32-GGUF: Same weights, converted
- DhruvalLabs/Nanbeige4.1-3B-GGUF: Quantization
- liodon-ai/Nanbeige4.1-3B-imatrix-GGUF: Quantization
- heretic-org/Nanbeige4.1-3B-heretic: Abliterated
- strykes/emberforge-3b-reasoner: LoRA-style
- MuXodious/Nanbeige4.1-3B-PaperWitch-heresy: Abliterated
- SkyAsl/Nanbeige4.1-VLM: LoRA-style
- NeshVerse/Uncensored_Nanbeige-4.1-3B: Abliterated
- DuoNeural/Nanbeige-4.1-3B-Abliterated: Abliterated
2 other repos hold byte-for-byte identical weight files.