OpenGVLab/InternVL3_5-8B
Partial fine-tune vs OpenGVLab/InternVL3_5-8B-MPO. Partial change: 89% of the weights changed (attention, lm_head, mlp, norms), in 36 of 36 layers.
Tensors changed: Embeddings 0/5, Norms 113/241, MLP 113/210, Attention 144/240, other 0/48, lm_head 1/1
9 derivatives checked
- bartowski/OpenGVLab_InternVL3_5-8B-GGUF: Same weights, converted (fewer layers)
- OpenGVLab/InternVL3_5-8B-Flash: Full fine-tune
- lmstudio-community/InternVL3_5-8B-GGUF: Quantization (fewer layers)
- mradermacher/InternVL3_5-8B-GGUF: Same weights, converted (fewer layers)
- mradermacher/InternVL3_5-8B-i1-GGUF: Quantization (fewer layers)
- internlm/CapRL-InternVL3.5-8B: Full fine-tune
- cyankiwi/InternVL3_5-8B-AWQ-4bit: Quantization
- xiaomi-research/SpatioLM-Understanding-InternVL3.5: Same weights, repacked
- cyankiwi/InternVL3_5-8B-AWQ-8bit: Quantization