mradermacher/gpt2-conversational-or-qa-GGUF

Same weights, converted vs Locutusque/gpt2-conversational-or-qa. Same weights stored in another dtype (F32/BOOL → F16/F32), plus 1 new tensors.

Tensors changed: Attention 0/24, Norms 0/50, MLP 0/24, Embeddings 0/2

On the Hub · Downloads on Model Pulse