DevQuasar/arcee-ai.Trinity-Mini-GGUF
Same weights, converted vs arcee-ai/Trinity-Mini. Same weights stored in another dtype (BF16 → F16/F32), plus 120 new tensors.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/193, MLP 0/126, Attention 0/160
Same weights, converted vs arcee-ai/Trinity-Mini. Same weights stored in another dtype (BF16 → F16/F32), plus 120 new tensors.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/193, MLP 0/126, Attention 0/160