jinaai/jina-embeddings-v5-omni-small-retrieval
LoRA-style vs jinaai/jina-embeddings-v5-omni-small. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 28 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: other 0/5, MLP 84/308, Norms 0/339, Attention 112/432, Embeddings 0/4
4 derivatives checked
- georgelpreput/jina-embeddings-v5-omni-small-retrieval-NVFP4: Quantization
- jinaai/jina-embeddings-v5-omni-small-retrieval-GGUF: Same weights, converted (fewer layers)
- ewin-reg/jina-embeddings-v5-omni-small-retrieval-Quantized: Quantization
- jinaai/jina-embeddings-v5-omni-small-retrieval-mlx: Partial fine-tune