LinaTea/Qwen3.5-4B-Fix-GGUF
Same weights, converted vs LinaTea/Qwen3.5-4B-Fix. Same weights stored in another dtype (BF16 → F16/F32), plus 120 new tensors.
Tensors changed: Embeddings 0/1, Norms 0/81, MLP 0/160, Attention 0/32
Same weights, converted vs LinaTea/Qwen3.5-4B-Fix. Same weights stored in another dtype (BF16 → F16/F32), plus 120 new tensors.
Tensors changed: Embeddings 0/1, Norms 0/81, MLP 0/160, Attention 0/32