flashback2k/FlashModel-Qwen-3.6-35B-A3B-Maxed
LoRA-style (fewer layers) vs Qwen/Qwen3.6-35B-A3B. flashback2k/FlashModel-Qwen-3.6-35B-A3B-Maxed matches part of Qwen/Qwen3.6-35B-A3B (712 of its 1045 tensors; the rest of Qwen/Qwen3.6-35B-A3B 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 40 of 40 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/2, Attention 190/284, MLP 120/287, Norms 0/137, lm_head 0/1, other 0/1
1 derivatives checked
- flashback2k/FlashModel-Qwen-3.6-35B-A3B-Maxed-GGUF: Quantization