mradermacher/NuMarkdown-8B-Thinking-GGUF
Same weights, converted (fewer layers) vs numind/NuMarkdown-8B-Thinking. mradermacher/NuMarkdown-8B-Thinking-GGUF matches part of numind/NuMarkdown-8B-Thinking (339 of its 729 tensors; the rest of numind/NuMarkdown-8B-Thinking has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).
Tensors changed: Embeddings 0/1, Norms 0/57, MLP 0/84, Attention 0/196, lm_head 0/1