The MTEB(por) benchmark for Brazilian Portuguese (leaderboard + task datasets) and the Colibri embedding model that tops it among small models.
Tardelli Stekel
tardellirs
AI & ML interests
None yet
Recent Activity
liked a model 1 day ago
hotchpotch/bekko-embedding-v1-a25m liked a model 1 day ago
hotchpotch/bekko-embedding-v1-a8m updated a Space 4 days ago
MTEB-BR/leaderboardOrganizations
Contamination-clean ~110M Brazilian-Portuguese embeddings (BERTimbau). #1 in the ~100M class on MTEB(por).
-
tardellirs/brazembed-pt-br
Sentence Similarity • 0.1B • Updated • 186 • 1 -
tardellirs/brazembed-pt-br-clfclu
Sentence Similarity • 0.1B • Updated • 126 -
tardellirs/brazembed-pt-br-pairclf
Sentence Similarity • 0.1B • Updated • 122 -
tardellirs/brazembed-pt-br-sts
Sentence Similarity • 0.1B • Updated • 108
Compact pt-br text-embedding model (~157M) that beats far larger models on MTEB(por): vocabulary trim -> Qwen3-4B distillation -> model soup.
Portuguese EmbeddingGemma-300M, vocab-trimmed 16k–128k. Training-free; 64k ≈ full quality at half the params.
-
tardellirs/embeddinggemma-pt-br
Sentence Similarity • 0.2B • Updated • 634 • 1 -
tardellirs/embeddinggemma-pt-br-128k
Sentence Similarity • 0.2B • Updated • 119 • 1 -
tardellirs/embeddinggemma-pt-br-48k
Sentence Similarity • 0.1B • Updated • 120 • 1 -
tardellirs/embeddinggemma-pt-br-32k
Sentence Similarity • 0.1B • Updated • 132 • 2
The MTEB(por) benchmark for Brazilian Portuguese (leaderboard + task datasets) and the Colibri embedding model that tops it among small models.
Compact pt-br text-embedding model (~157M) that beats far larger models on MTEB(por): vocabulary trim -> Qwen3-4B distillation -> model soup.
Contamination-clean ~110M Brazilian-Portuguese embeddings (BERTimbau). #1 in the ~100M class on MTEB(por).
-
tardellirs/brazembed-pt-br
Sentence Similarity • 0.1B • Updated • 186 • 1 -
tardellirs/brazembed-pt-br-clfclu
Sentence Similarity • 0.1B • Updated • 126 -
tardellirs/brazembed-pt-br-pairclf
Sentence Similarity • 0.1B • Updated • 122 -
tardellirs/brazembed-pt-br-sts
Sentence Similarity • 0.1B • Updated • 108
Portuguese EmbeddingGemma-300M, vocab-trimmed 16k–128k. Training-free; 64k ≈ full quality at half the params.
-
tardellirs/embeddinggemma-pt-br
Sentence Similarity • 0.2B • Updated • 634 • 1 -
tardellirs/embeddinggemma-pt-br-128k
Sentence Similarity • 0.2B • Updated • 119 • 1 -
tardellirs/embeddinggemma-pt-br-48k
Sentence Similarity • 0.1B • Updated • 120 • 1 -
tardellirs/embeddinggemma-pt-br-32k
Sentence Similarity • 0.1B • Updated • 132 • 2