danielchalef/Qwen3-Reranker-4B-seq-cls-vllm-fixed

Same weights, converted vs Qwen/Qwen3-Reranker-4B. Same weights stored in another dtype (BF16 → F32), plus 1 new tensors.

Tensors changed: Embeddings 0/1, Norms 0/145, MLP 0/108, Attention 0/144

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse