OpenGVLab/InternViT-6B-448px-V2_5
Full fine-tune vs OpenGVLab/InternViT-6B-448px-V1-5. Fine-tune: 100% of the weights changed, across 45 of 45 layers.
Tensors changed: Embeddings 4/4, Norms 180/180, Attention 135/135, other 90/90, MLP 180/180
7 derivatives checked
- OpenGVLab/InternVL2_5-38B: Full fine-tune (fewer layers)
- OpenGVLab/InternVL2_5-26B: Full fine-tune (fewer layers)
- OpenGVLab/InternVL2_5-78B: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3-78B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3_5-38B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3_5-241B-A28B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3-38B-Pretrained: Full fine-tune (fewer layers)