mradermacher/ThinkTwice-Olmo3-7B-Instruct-GGUF

Same weights, converted vs difanjiao/ThinkTwice-Olmo3-7B-Instruct. Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Embeddings 0/1, MLP 0/96, Attention 0/128, Norms 0/129, lm_head 0/1

On the Hub · Downloads on Model Pulse