Hugging Face's logo

Spaces:
junaid0600
/
sql-db-engineer-agent
Sleeping

App Files Community
1
sql-db-engineer-agent

Commit History

  1. Update env/environment.py

    junaid0600
    Verified signature
    1eb9a1b

  2. Update training/colab_notebook.py

    junaid0600
    Verified signature
    e5aa2dc

  3. Update env/environment.py

    junaid0600
    Verified signature
    cf6d807

  4. Update training/train_agent.py

    junaid0600
    Verified signature
    44e9354

  5. Update training/train_agent.py

    junaid0600
    Verified signature
    86abfc1

  6. Update training/train_agent.py

    junaid0600
    Verified signature
    987f2db

  1. Upload reward_curve.png

    junaid0600
    Verified signature
    a7802a8

  2. Add Colab training notebook link

    junaid0600
    9b80a84

  3. changes evalute_agent

    junaid0600
    15e9605

  1. updated readme

    junaid0600
    b96d42b

  2. corrected

    junaid0600
    8778707

  3. Reward curve: strategic +36.7pts vs random +0.0pts

    junaid0600
    842e560

  1. changes

    junaid0600
    b2742eb

  2. updated requirements

    junaid0600
    e4126f3

  3. updated readme

    junaid0600
    f004baa

  4. prproject.toml and readme updated

    junaid0600
    809345d

  5. Final Round 2: all checks passing, openenv validate OK

    junaid0600
    f30d05a

  1. Fix gitignore - exclude pycache

    junaid0600
    a28e8c9

  2. Force add all env, dataset, api files - gitignore fix

    junaid0600
    399f4c5

  3. Add db_simulator and scenario files

    junaid0600
    ff10d5b

  4. Round 2: SQL Database Engineer Agent - 24/24 tests passing

    junaid0600
    8cb206e

  1. changed

    junaid0600
    1eef47f

  2. changed

    junaid0600
    deb9d37

  3. changed

    junaid0600
    f5f1b7a

  4. changed

    junaid0600
    ea504bf

  5. changed

    junaid0600
    dcaf698

  6. changed

    junaid0600
    f23139f

  7. Use real LLM call for proxy check + baseline scores for task validation

    junaid0600
    5e3e79e

  8. Use real LLM calls through API_BASE_URL proxy

    junaid0600
    d8cba4f

  9. Clean inference.py using baseline scores strictly between 0 and 1

    junaid0600
    b02ec3c

  10. Normalize all rewards to strictly (0.001, 0.999) range in step()

    junaid0600
    42a1cbd

  11. Fix GraderResponse schema example score from 1 to 0.75

    junaid0600
    1a89fae

  12. Clamp all reward scores strictly between 0.001 and 0.999

    junaid0600
    ef20791

  13. Clamp grader scores strictly between 0.001 and 0.999 in endpoint and model

    junaid0600
    f2d88cb

  14. Fix rewards never exactly 0.0 or 1.0 using proper normalization

    junaid0600
    7dff36b

  15. Clamp all step rewards strictly between 0.001 and 0.999

    junaid0600
    11dd1d6

  1. Fix score - shift rewards to positive range, never 0.0 or 1.0

    junaid0600
    888871f

  2. Ensure score never below 0.1 to fix out of range error

    junaid0600
    e15627e

  3. Fix score strictly between 0.001 and 0.999 - never 0.0 or 1.0

    junaid0600
    6e703c0

  4. corrected everytihng

    junaid0600
    2146d9e

  5. again fixed graders

    junaid0600
    d4b572f

  6. fixed

    junaid0600
    59746b9

  7. again changed

    junaid0600
    95c7542

  8. new fixed

    junaid0600
    b4cf41e

  9. done

    junaid0600
    95b11e6

  10. fixed

    junaid0600
    49552e4

  11. done

    junaid0600
    265556a

  12. done

    junaid0600
    3db1e7f

  13. fixed new error

    junaid0600
    0ac8fe8

  14. fixed

    junaid0600
    81abf49