Hugging Face's logo

Spaces:
ProthamD
/
prothamd-adaptive-world-env
Sleeping

App Files Community
prothamd-adaptive-world-env

Commit History

  1. Update README.md

    ProthamD
    Verified signature
    2ae1f78

  2. Upload blog.md

    ProthamD
    Verified signature
    679cc36

  3. Delete blog.md

    ProthamD
    Verified signature
    b18b3a4

  1. Update static frontend: Experience tab, mobile nav, responsive layout, live run UI⁠

    PROTHAM +1: Cursor
    a36b0c6

  1. Update blog.md

    ProthamD
    Verified signature
    184901b

  2. Serve SPA with no-cache headers so HF UI picks up static/index.html after deploy⁠

    PROTHAM
    0db7f46

  3. Blog/Space: second-person to we/us in analogy and correlation sections⁠

    PROTHAM
    7ff548f

  4. Blog: reconcile correlation narrative; add RNG/game analogy; push protham.ipynb⁠

    PROTHAM
    bee85ba

  5. Refresh training metrics and dashboard image for HF Space⁠

    PROTHAM
    58bb04b

  1. Embed real 77-step GRPO data into training charts⁠

    PROTHAM
    ce4b86c

  2. Bump version to trigger Space rebuild⁠

    PROTHAM
    ad9962b

  3. Update training dashboard charts and stats⁠

    PROTHAM
    4f84b16

  4. Fix Bug 4 and Bug 5 from FIXES.md: _handle_history step_log + asyncio lock for start_episode⁠

    PROTHAM
    65bd8b6

  1. replace unverified % estimates with actual measured numbers from notebook⁠

    PROTHAM
    52e6772

  2. fix: use actual measured numbers from protham.ipynb, remove speculative line from blog⁠

    PROTHAM
    e8782ec

  3. add SPA frontend: Blog, Training, README, API tabs with Chart.js graphs⁠

    PROTHAM
    a0524d0

  4. Update blog.md

    ProthamD
    Verified signature
    9145742

  5. Update blog.md

    ProthamD
    Verified signature
    ab24133

  6. chore: add LFS tracking for PNG files, migrate images to LFS

    PROTHAM
    e8e1506

  7. add training section, result images, and protham notebook to blog

    PROTHAM
    1e85be0

  1. feat: 3-phase belief training — probe evidence → model generates belief from observation⁠

    PROTHAM
    a4ea5bf

  2. fix: model generates belief — parse 2 JSONs from output, raise belief weight to 0.4⁠

    PROTHAM
    455e30f

  3. fix: simplify install to 2 cells — force-reinstall torch (NCCL fix) then unsloth⁠

    PROTHAM
    0297a6f

  4. fix: adopt uv + torch-presence check for unsloth install (mirrors official OpenEnv notebook)⁠

    PROTHAM
    9083784

  5. fix: use pip install unsloth instead of git-based install; split install/verify into two cells⁠

    PROTHAM
    3bf856f

  6. fix: run env.step in thread pool inside /run_episode to prevent event loop deadlock⁠

    PROTHAM
    8ef5b39

  7. fix: add /run_episode endpoint + update reward_fn to use it⁠

    PROTHAM
    55a1039

  8. fix: bypass localhost HTTP for drift/probe — use direct in-process DYNAMIC_CONFIG mutation⁠

    PROTHAM
    dbffc1d

  9. feat: evidence-augmented prompts for true world modeling⁠

    PROTHAM
    c75e589

  10. fix: drift_trigger_step + reward_fn for proper belief scores⁠

    PROTHAM
    ddcfa7f

  11. chore: add uv.lock for openenv deployment

    PROTHAM
    e566985

  1. Fix environment bug: adjust drift triggers and prevent premature auto-termination to avoid reward hacking; update README

    PROTHAM
    0983a18

  1. Remove deprecated __init__ arguments from EnvClient

    PROTHAM
    6d193c2

  2. Update default HF router model to Mistral 7B to ensure standard endpoint availability

    PROTHAM
    d78447a

  3. Fix inference loop for openenv-core 0.2.3 StepResult compatibility

    PROTHAM
    aaf289c

  4. Add HF Spaces YAML metadata to README

    PROTHAM
    ea9cb24

  5. Add .gitignore and remove cached build artifacts

    PROTHAM
    2b433a4

  6. Initial commit: AdaptiveWorld Env v2.0 (Round 2)

    PROTHAM
    7043cc6