cduk/embeddinggemma-300m-GGUF-with-dense-modules

embeddinggemma-300m in GGUF format with included post-processing dense modules.

Unlike other GGUF which exclude the dense modules, this generates embeddings compatible with TEI and sentence-transformers.

For details, see:

• https://github.com/ggml-org/llama.cpp/pull/16367 and
• https://github.com/ggml-org/llama.cpp/issues/16538

Rank	Model	Size	Similarity	Size Reduction
🥇 1st	embeddinggemma-300M-Q8_0.gguf	318MB	99.96%	45.8%
🥈 2nd	embeddinggemma-300M-Q6_K.gguf	252MB	99.75%	57.1%
🥉 3rd	embeddinggemma-300M-Q5_K.gguf	238MB	99.44%	59.4%
4th	embeddinggemma-300M-QAT-Q8.gguf	318MB	98.68%	45.8%
5th	embeddinggemma-300M-Q4_K.gguf	228MB	98.32%	61.2%
6th	embeddinggemma-300M-QAT-Q4_K.gguf	228MB	95.81%	61.2%