OpenGVLab/InternVL2_5-4B
Full fine-tune (fewer layers) vs OpenGVLab/InternViT-300M-448px-V2_5. OpenGVLab/InternViT-300M-448px-V2_5 matches part of OpenGVLab/InternVL2_5-4B (340 of its 781 tensors; the rest of OpenGVLab/InternVL2_5-4B has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 24 of 24 layers.
Tensors changed: Embeddings 4/4, Attention 96/96, other 48/48, MLP 96/96, Norms 96/96
4 derivatives checked
- OpenGVLab/InternVL2_5-4B-MPO: Full fine-tune
- mradermacher/InternVL2_5-4B-i1-GGUF: Quantization (fewer layers)
- ggml-org/InternVL2_5-4B-GGUF: Same weights, converted (fewer layers)
- mradermacher/InternVL2_5-4B-GGUF: Same weights, converted (fewer layers)
4 other repos hold byte-for-byte identical weight files.