EximiusLabs/fusion-embedding-1-2b-preview

Different architecture vs Qwen/Qwen3-VL-Embedding-2B. Different architecture: only 0 of 625/131 tensors match by name and shape.

On the Hub · Downloads on Model Pulse