beyoru/Qwen3-TTS-12Hz-1.7B-Vi
LoRA-style vs Qwen/Qwen3-TTS-12Hz-1.7B-Base. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, linear_fc1, linear_fc2, o_proj, q_proj, up_proj, v_proj changed, in 28 of 28 layers; norms, biases and embeddings are untouched.
Tensors changed: other 0/79, lm_head 0/15, Embeddings 0/17, Norms 0/134, MLP 101/103, Attention 132/132
2 derivatives checked
- mradermacher/Qwen3-TTS-12Hz-1.7B-Vi-i1-GGUF: Quantization (fewer layers)
- mradermacher/Qwen3-TTS-12Hz-1.7B-Vi-GGUF: Same weights, converted (fewer layers)