prithivMLmods/Behemoth-3B-070225-post0.1
Partial fine-tune vs Qwen/Qwen2.5-VL-3B-Instruct. Partial change: 84% of the weights changed (attention, embeddings, mlp, norms), in 36 of 36 layers.
Tensors changed: Embeddings 1/2, Norms 69/138, MLP 116/304, Attention 260/380
2 derivatives checked
- mradermacher/Behemoth-3B-070225-post0.1-i1-GGUF: Quantization (fewer layers)
- mradermacher/Behemoth-3B-070225-post0.1-GGUF: Same weights, converted (fewer layers)