Qwen/Qwen3-1.7B
Full fine-tune vs Qwen/Qwen3-1.7B-Base. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 108/113, MLP 84/84, Attention 112/112
380 derivatives checked
- MaziyarPanahi/Qwen3-1.7B-GGUF: Same weights, converted
- unsloth/Qwen3-1.7B-GGUF: Same weights, converted
- Qwen/Qwen3-1.7B-GGUF: Full fine-tune
- ggml-org/Qwen3-1.7B-GGUF: Same weights, converted
- tobil/qmd-query-expansion-1.7B-gguf: LoRA-style
- unsloth/Qwen3-1.7B-unsloth-bnb-4bit: Quantization
- nvidia/canary-qwen-2.5b: Adapter
- bartowski/Qwen_Qwen3-1.7B-GGUF: Same weights, converted
- unsloth/Qwen3-1.7B: Same weights, repacked
- lmstudio-community/Qwen3-1.7B-MLX-8bit: Quantization
- lmstudio-community/Qwen3-1.7B-MLX-4bit: Quantization
- lmstudio-community/Qwen3-1.7B-MLX-bf16: Same weights, repacked
- Qwen/Qwen3-1.7B-FP8: Quantization
- lmstudio-community/Qwen3-1.7B-GGUF: Quantization
- mlx-community/Qwen3-1.7B-4bit: Quantization
- geoffmunn/Qwen3-1.7B-f16: Not comparable yet
- reaperdoesntknow/Shepherd-Alpha: Full fine-tune
- reaperdoesntknow/Qwen3-1.7B-Distilled-30B-A3B: Partial fine-tune
- reaperdoesntknow/Qwen3-1.7B-Coder-Distilled-SFT: Partial fine-tune
- reaperdoesntknow/Qemma-Q1.7B: Different architecture
- blackdeep/knaif: Quantization
- gaianet/Qwen3-1.7B-GGUF: Same weights, converted
- Skywork/Skywork-Reward-V2-Qwen3-1.7B: Full fine-tune
- unsloth/Qwen3-1.7B-bnb-4bit: Quantization
- KKHYA/llavaqwen3-1.7b-finetune: Full fine-tune (fewer layers)
- reaperdoesntknow/DualMind-TKD-Agentic-1.7B: Full fine-tune
- XythicK/Qwen3-1.7B-GGUF: Same weights, converted
- shreyanshp/bitcoin-news-1.7b: Full fine-tune
- rippertnt/Qwen3-1.7B-Q4_K_M-GGUF: Quantization
- OpenMOSS-Team/MOSS-SoundEffect-v2.0: Different architecture
117 other repos hold byte-for-byte identical weight files.