mradermacher/Elixir-Qwen2.5-Coder-3B-GGUF
Same weights, converted vs jmarceno/Elixir-Qwen2.5-Coder-3B. Same weights stored in another dtype (BF16 → F16/F32).
Tensors changed: Embeddings 0/1, Norms 0/75, MLP 0/111, Attention 0/259
Same weights, converted vs jmarceno/Elixir-Qwen2.5-Coder-3B. Same weights stored in another dtype (BF16 → F16/F32).
Tensors changed: Embeddings 0/1, Norms 0/75, MLP 0/111, Attention 0/259