Run MSMARCO dev split on some models

[x-tabdeveloping]

We discovered with @Samoed that some models are completely missing evaluation results on MSMARCO's dev split, and they were included in the old leaderboard due to a bug #1867 . As such, if we want to have their results on the leaderboard, we will have to run them.

I have compiled a list of models that are only missing MSMARCO scores:

models_only_missing_msmarco = [
    "voyageai/voyage-large-2-instruct",
    "Cohere/Cohere-embed-english-v3.0",
    "nvidia/NV-Embed-v1",
    "Haon-Chen/speed-embedding-7b-instruct",
    "Cohere/Cohere-embed-multilingual-v3.0",
    "abhinand/MedEmbed-small-v0.1",
    "Cohere/Cohere-embed-english-light-v3.0",
    "Cohere/Cohere-embed-multilingual-light-v3.0",
    "bigscience/sgpt-bloom-7b1-msmarco",
    "Lajavaness/bilingual-embedding-large",
    "Lajavaness/bilingual-embedding-base",
    "deepvk/USER-bge-m3",
    "manu/bge-m3-custom-fr",
    "Lajavaness/bilingual-embedding-small",
    "manu/sentence_croissant_alpha_v0.4",
    "manu/sentence_croissant_alpha_v0.3",
    "manu/sentence_croissant_alpha_v0.2",
    "minishlab/potion-base-8M",
    "minishlab/M2V_base_glove",
    "minishlab/potion-base-4M",
    "minishlab/M2V_base_glove_subword",
    "minishlab/M2V_base_output",
    "consciousAI/cai-lunaris-text-embeddings",
    "minishlab/potion-base-2M",
    "Omartificial-Intelligence-Space/Arabic-mpnet-base-all-nli-triplet",
    "ai-forever/sbert_large_mt_nlu_ru",
    "ai-forever/sbert_large_nlu_ru",
    "malenia1/ternary-weight-embedding",
    "deepvk/deberta-v1-base",
]

Can we get a run on these @Muennighoff ?