sungo-ganpare/memvid-embedding-models

memvid-embedding-models

Packaged ONNX embedding models for the Japanese-enabled memvid fork.

GitHub repository:

• —https://github.com/ganpare/memvid

This model repo is used by the setup scripts in the GitHub repo:

• —scripts/setup_japanese_models.ps1
• —scripts/setup_japanese_models.sh

Included models

multilingual-e5-large

Files:

• —multilingual-e5-large/multilingual-e5-large.onnx
• —multilingual-e5-large/model.onnx_data
• —multilingual-e5-large/multilingual-e5-large_tokenizer.json

Original model:

• —intfloat/multilingual-e5-large

License:

• —MIT

ruri-pt-large

Files:

• —ruri-pt-large/ruri-pt-large.onnx
• —ruri-pt-large/vocab.txt
• —ruri-pt-large/tokenizer_config.json
• —ruri-pt-large/special_tokens_map.json

Original model:

• —cl-nagoya/ruri-pt-large

License:

• —Apache-2.0

Usage

Windows:

powershell -ExecutionPolicy Bypass -File .\scripts\setup_japanese_models.ps1

Linux / macOS:

bash ./scripts/setup_japanese_models.sh

See the GitHub repo for memvid integration details and the current tokenizer fallback notes.