handwoven8588/CodeRankEmbed-flash-attn
Same weights, converted vs nomic-ai/CodeRankEmbed. Same weights stored in another dtype (F32 → BF16).
Tensors changed: other 0/2, Embeddings 0/2, Attention 0/24, MLP 0/36, Norms 0/48
Same weights, converted vs nomic-ai/CodeRankEmbed. Same weights stored in another dtype (F32 → BF16).
Tensors changed: other 0/2, Embeddings 0/2, Attention 0/24, MLP 0/36, Norms 0/48