Hugging Face's logo

prereasoner
/
prereasoner-weights

Table Question Answering
PEFT
Safetensors
text-to-sql
schema.org
lora
Model card Files
xet
 Community
prereasoner-weights / readers

Commit History

3 days ago

  1. Readers v14: v14b in slot b, so all four reading models are trained with SynSQL-2.5M's databases; the choice over both re-scorers refitted on their pool, threshold 0.0862

    manidoraisamy
    Verified signature
    cd8a03a

  2. Readers v14 with two re-scorers: rescorer_a ('f') and rescorer_b ('g', the v14 recipes' cross-fold candidates) weighed by one choice; v11b back in slot b; the single readers/rescorer/ removed

    manidoraisamy
    Verified signature
    4ad8d30

4 days ago

  1. Readers v14: four reading models trained with SynSQL-2.5M's databases (v14a, v14b, v14b1, v14a1), the re-scorer refreshed on their recipes' cross-fold candidates, the choice refitted, threshold 0.129

    manidoraisamy
    Verified signature
    25a5564

6 days ago

  1. Readers: v13b1 and v13a1 in slots c and d (trained with SELECT * and sheet language); the choice refitted on that pool, threshold 0.1405

    manidoraisamy
    Verified signature
    999b50a

  2. Readers: v12b1 and v12a1 in slots c and d (trained with the grouped structure); the choice refitted on that pool, threshold 0.1434

    manidoraisamy
    Verified signature
    fd08934

7 days ago

  1. Readers v11: the choice (readers/choice.json) replaces the refusal; cards describe the four readers

    manidoraisamy
    Verified signature
    d494912

  2. Readers v11: four reading models (stage2_a to stage2_d); stage 1 and the re-scorer unchanged

    manidoraisamy
    Verified signature
    619d117

8 days ago

  1. Add the own-data readers (readers/: schema items, two reading models, re-scorer, refusal); cards describe them

    manidoraisamy
    Verified signature
    b9bb652