PrunaAI/gpt2-GGUF-smashed
Quantization vs openai-community/gpt2. PrunaAI/gpt2-GGUF-smashed is a GGUF quantization (Q8_0) of the same weights: 99 tensors match exactly and 1 within quantization error.
Tensors changed: Norms 0/50, Attention 0/24, MLP 0/24, Embeddings 0/2