YOYO-AI/Qwen3-VL-4B-YOYO-Instruct
Full fine-tune (fewer layers) vs Qwen/Qwen3-4B-Instruct-2507. Qwen/Qwen3-4B-Instruct-2507 matches part of YOYO-AI/Qwen3-VL-4B-YOYO-Instruct (398 of its 713 tensors; the rest of YOYO-AI/Qwen3-VL-4B-YOYO-Instruct has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 1/1, Norms 145/145, MLP 108/108, Attention 144/144
2 derivatives checked
- mradermacher/Qwen3-VL-4B-YOYO-Instruct-i1-GGUF: Quantization (fewer layers)
- mradermacher/Qwen3-VL-4B-YOYO-Instruct-GGUF: Same weights, converted (fewer layers)