empero-ai/Qwen3.8-9B-Distill
Partial fine-tune vs Qwen/Qwen3.5-9B. Partial change: 72% of the weights changed (attention, mlp, norms), in 32 of 32 layers.
Tensors changed: lm_head 0/1, Embeddings 0/5, MLP 96/211, Attention 177/336, other 0/1, Norms 58/221
37 derivatives checked
- empero-ai/Qwen3.8-9B-Distill-GGUF: Same weights, converted
- petruhonk/Qwen3.8-9B-Distill-uncensored-heretic-GGUF: Abliterated
- Noobito45/Qwen3.8-9B-heretic-uncensored-NVFP4-GGUF: Quantization
- Atomic-Germ/Qwen3.8-9B-Distill-GGUF: Quantization
- keXjos/Qwen3.8-9B-mlx-4Bit: Quantization
- Atomic-Germ/Qwen3.8-Distilled-9B-NPU2: Quantization
- Sohailhosseini/Qwen3.8-9B-Distill-AWQ-W4A16: Quantization
- SiddhJagani/Qwen3.8-9B-mlx-8Bit: Quantization
- keXjos/Qwen3.8-9B-Distill-mlx-5Bit: Quantization
- luxuansang/Qwen3.8-9B-heretic-uncensored-NVFP4-GGUF: Quantization
- srmiles/Qwen3.8-9B-Distill-MTP-GGUF: Not comparable yet
- Osakra/Project-Norn-V20-9B: Different architecture
- schsu/Qwen3.8-9B-MLX-MXFP4: Quantization
- hakergeniusz/Qwen3.8-9B-Distill-GGUF: Same weights, converted
- pooyakn73/Qwen3.8-9B-Distill-MTPLX-Q6G64: Quantization
- nicolasramos/Qwen3.8-9B-Distill-MLX-4bit-mtp: Quantization
- ermiaazarkhalili/Qwen3.8-9B-SFT-Fable5-Glint-GGUF: Full fine-tune
- JamieBradfield/Dirk-Qwen3.8-9B-GGUF: Same weights, converted
- Sohailhosseini/Qwen3.8-9B-Distill-FP8: Quantization
- enginil/Qwen3.8-9B-mlx-6Bit: Quantization
- ermiaazarkhalili/Qwen3.8-9B-Function-Calling-xLAM-Unsloth-GGUF: Full fine-tune
- Vishva007/Qwen3.8-9B-Distill-W4A16-AutoRound-GPTQ: Quantization
- SiddhJagani/Qwen3.8-9B-mlx-4Bit: Quantization
- SiddhJagani/Qwen3.8-9B-mlx-3Bit: Quantization
- SiddhJagani/Qwen3.8-9B-mlx-6Bit: Quantization
- SiddhJagani/Qwen3.8-9B-mlx-2Bit: Quantization
- nvythong/Qwen3.8-9B-Distill-mlx-4Bit: Quantization
- nicolasramos/Qwen3.8-9B-Distill-MLX-bf16-mtp: Partial fine-tune (fewer layers)
- DarkKitsune/Qwen3.5-9B-Qworus-V2: Full fine-tune
- SiddhJagani/Qwen3.8-9B-mlx-5Bit: Quantization