mradermacher/InstructLM-500M-GGUF
Same weights, converted vs instruction-pretrain/InstructLM-500M. Same weights stored in another dtype (F32/F16 → F16/F32).
Tensors changed: lm_head 0/1, Embeddings 0/1, MLP 0/48, Attention 0/64, Norms 0/33
Same weights, converted vs instruction-pretrain/InstructLM-500M. Same weights stored in another dtype (F32/F16 → F16/F32).
Tensors changed: lm_head 0/1, Embeddings 0/1, MLP 0/48, Attention 0/64, Norms 0/33