Commit History
-
Fix word_embedding_dimension in 1_Pooling/config.json
Verified signature1_Pooling/config.json declares word_embedding_dimension: 1024, but this model's actual hidden_size (per config.json) is 2560. This mismatch causes sentence-transformers' MSELoss (which sizes its projection layer via model.get_embedding_dimension(), trusting this file) to build a Linear(1024, ...) layer instead of Linear(2560, ...), raising a shape mismatch at runtime when real 2560-dim embeddings are passed in. It also affects any other pooling-dimension-derived Matryoshka dimension defaults. This PR corrects word_embedding_dimension to 2560 to match config.json's hidden_size; all other pooling settings are unchanged.adebe0a
-
Update README.md
Verified signaturecb94392
-
Update README.md
Verified signature1434021
-
Update README.md
Verified signaturea28abc8
-
Upload intermediate checkpoints
Verified signature47d6e2c
-
Upload folder using huggingface_hub
Verified signature8c969e5 -
Update README.md
Verified signature2e39bd6 -
initial commit
Verified signature4a40803