OpenGVLab/InternVL3-2B
Full fine-tune vs OpenGVLab/InternVL3-2B-Instruct. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: lm_head 1/1, Embeddings 5/5, Norms 93/153, MLP 180/186, Attention 288/292, other 0/48
5 derivatives checked
- unsloth/InternVL3-2B-GGUF: Same weights, converted (fewer layers)
- mradermacher/InternVL3-2B-i1-GGUF: Quantization (fewer layers)
- mradermacher/InternVL3-2B-GGUF: Same weights, converted (fewer layers)
- Zoont/InternVL3-2B-4-Bit-GGUF-with-mmproj: Quantization (fewer layers)
- RKNNAI/RK182X-VLM-InternVL3-2B: Quantization
2 other repos hold byte-for-byte identical weight files.