OpenGVLab/InternViT-300M-448px-V2_5
Full fine-tune vs OpenGVLab/InternViT-300M-448px. 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
19 derivatives checked
- OpenGVLab/InternVL3_5-GPT-OSS-20B-A4B-Preview-HF: Different architecture
- OpenGVLab/InternVL2_5-2B: Full fine-tune (fewer layers)
- OpenGVLab/InternVL2_5-8B: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3_5-GPT-OSS-20B-A4B-Preview: Full fine-tune (fewer layers)
- OpenGVLab/InternVL2_5-4B: Full fine-tune (fewer layers)
- OpenGVLab/InternVL2_5-1B: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3_5-8B-Pretrained: Full fine-tune (fewer layers)
- mlx-community/InternVL3-2B-4bit: Quantization
- OpenGVLab/InternVL3_5-4B-Pretrained: Full fine-tune (fewer layers)
- mlx-community/InternVL3-1B-4bit: Quantization
- OpenGVLab/InternVL3_5-2B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3-8B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3-1B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3_5-14B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3_5-1B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3_5-30B-A3B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3-2B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3-14B-Pretrained: Full fine-tune (fewer layers)
- OpenGVLab/InternVL3-9B-Pretrained: Full fine-tune (fewer layers)