prithivMLmods/Open-R1-Mini-Experimental
LoRA-style vs Qwen/Qwen2-VL-2B-Instruct. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, fc1, fc2, gate_proj, k_proj, o_proj, proj, q_proj, qkv, up_proj, v_proj changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/2, Norms 0/187, MLP 147/216, Attention 176/324
4 derivatives checked
- mradermacher/Open-R1-Mini-Experimental-i1-GGUF: Quantization (fewer layers)
- mradermacher/Open-R1-Mini-Experimental-GGUF: Same weights, converted (fewer layers)
- DevQuasar-7/prithivMLmods.Open-R1-Mini-Experimental-GGUF: Same weights, converted (fewer layers)
- prithivMLmods/Open-R1-Mini-Experimental-GGUF: Same weights, converted (fewer layers)