chaoliangUNSW/Jev-Style-Qwen3.5-2B-Decision-v2
LoRA-style (fewer layers) vs Qwen/Qwen3.5-2B-Base. chaoliangUNSW/Jev-Style-Qwen3.5-2B-Decision-v2 matches part of Qwen/Qwen3.5-2B-Base (320 of its 632 tensors; the rest of Qwen/Qwen3.5-2B-Base has no counterpart). On the shared part: Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, in_proj_a, in_proj_b, in_proj_qkv, in_proj_z, k_proj, o_proj, out_proj, q_proj, up_proj, v_proj changed, in 24 of 24 layers; norms, biases and embeddings are untouched.
Tensors changed: Attention 114/168, Norms 0/79, Embeddings 0/1, MLP 72/72
2 derivatives checked
- chaoliangUNSW/Jev-Style-Qwen3.5-2B-Decision-v2-GGUF: Partial fine-tune
- chaoliangUNSW/Jev-Style-Qwen3.5-2B-Decision-v2-MLX-bf16: Partial fine-tune