TucanoBR/Tucano-2b4

Text Generation • 2B • Updated Jan 15 • 2.29k • 4

Note 2.4 billion-parameter version of the Tucano series.

TucanoBR/Tucano-2b4-Instruct

Text Generation • 2B • Updated Jan 15 • 1.77k • 2

Note 2.4 billion-parameter version of the Tucano fine-tuned on the TucanoBR/Tucano-SFT dataset.

TucanoBR/Tucano-1b1

Text Generation • 1B • Updated Jan 15 • 1.77k • 2

Note 1.1 billion-parameter version of the Tucano series.

TucanoBR/Tucano-1b1-Instruct

Text Generation • 1B • Updated Jan 15 • 92 • 1

Note 1.1 billion-parameter version of the Tucano fine-tuned on the TucanoBR/Tucano-SFT dataset.

TucanoBR/Tucano-630m

Text Generation • 0.6B • Updated Jan 15 • 101 • 2

Note 630 million-parameter version of the Tucano series.

TucanoBR/Tucano-160m

Text Generation • 0.2B • Updated Jan 15 • 1.83k • 1

Note 160 million-parameter version of the Tucan series.

TucanoBR/BERTimbau-large-text-filter

Text Classification • 0.3B • Updated Nov 13, 2024 • 12

Note BERTimbau-large fine-tuned on the TucanoBR/GigaVerbo-Text-Filter dataset.

TucanoBR/BERTimbau-base-text-filter

Text Classification • 0.1B • Updated Nov 13, 2024 • 20

Note BERTimbau-base fine-tuned on the TucanoBR/GigaVerbo-Text-Filter dataset.

TucanoBR/XGBClassifier-text-filter

Updated Nov 13, 2024

Note XGBClassifier trained on the TucanoBR/GigaVerbo-Text-Filter dataset (requires the embeddings generated by sentence-transformers/LaBSE).

TucanoBR/XGBRegressor-text-filter

Updated Nov 13, 2024

Note XGBRegressor trained on the TucanoBR/GigaVerbo-Text-Filter dataset (requires the embeddings generated by sentence-transformers/LaBSE).

Note GigaVerbo is an extensive dataset comprising 780 GB of Portuguese text, being a concatenated version of several datasets available in Hugging Face, containing over 200 billion tokens.

Note GigaVerbo Text-Filter is a dataset with 110,000 randomly selected samples from 9 subsets of GigaVerbo, all scored by GPT-4o.

Note This is the dataset used to train the "Instruct" versions of the Tucano series.

Note This dataset is a translated version (Portuguese) of the LAMBADA test split as pre-processed by OpenAI.

Note This dataset contains 805 translated samples (Portuguese) from the Alpaca dataset.

Note This dataset contains pairs of completions to prompts. Used for DPO fine-tuning.