There are lots of models that fail at the basic step of loading (e.g. #1689). Maybe it would be good to have a unit test that checks that all models can be loaded? One problem may the size of some large models, but maybe there is a way to work around that / we only test models that can be loaded in a reasonable amount of time.

Non-exhaustive list:

• voyageai/voyage-3 fails on RuBQReranking #1652
• infgrad/jasper_en_vision_language_v1 fails on MassiveScenarioClassification #1653
• HIT-TMG/KaLM-embedding-multilingual-mini-instruct-v1 fails #1657
• Hum-Works/lodestone-base-4096-v1 fails #1658
• openbmb/MiniCPM-Embedding fails #1659
• BeastyZ/e5-R-mistral-7b fails #1661
• LLM2Vec fails #1679
• nomic-ai/nomic-embed-text-v1.5 fails #1680
• voyage fails #1681
• jinaai/jina-embeddings-v2-small-en fails #1689
• Lajavaness/bilingual-embedding-large fails due to trust_remote_code #1693
• openbmb/MiniCPM-Embedding still fails #1696