Spaces:
Running
Running
Commit History
-
ui: chatbot dark-neon theme — readable bubbles on Watch Agent Play Running
Anurag Agarwal +1: CursorCo-authored-by: Cursor <cursoragent@cursor.com>d90960c -
fix(ui): drop unsupported type= kwarg; dict format works by default
Anurag Agarwal +1: CursorCo-authored-by: Cursor <cursoragent@cursor.com>a435a88 -
fix(ui): Gradio chatbot format error on Watch Agent Play
Anurag Agarwal +1: CursorGradio runtime auto-upgraded; new default chatbot type is "messages" which expects list[dict] with role/content keys. Was returning list[list] tuples → "Data incompatible with messages format" error when selecting a trace. - _format_trace_as_chat now returns list[dict] - _run_live_episode now returns list[dict] - gr.Chatbot widgets now explicitly use type="messages" Co-authored-by: Cursor <cursoragent@cursor.com>d63a291
-
UI: fix '0 runs' chip + instant tab-card flip via js= callback
Anurag Agarwale0be40d -
UI: replace tab strip with big-card navigation
Anurag AgarwalMade-with: Cascade3782303 -
Semantic run names: Probe/Drift/Anchor/Restrain/Champion + regen all plots
Anurag AgarwalMade-with: Cursor84fbeda -
Reframe: drop Run 7 from hero, keep only where appropriate
Anurag AgarwalMade-with: Cursor023e210 -
Polish Blog.md: fix broken images + visual hierarchy
Anurag AgarwalMade-with: Cursorfe94ae7 -
Stronger opening: meeting scheduling with 3 fabricated fields
Anurag AgarwalMade-with: Cursor5df7029 -
Hero Zone redesign + fix Run 7 missing from plot 08
Anurag AgarwalMade-with: Cursor535b7f9 -
Expand Blog.md to comprehensive deep-dive (5.9k words, all evidence)
Anurag AgarwalMade-with: Cursor7fbfbc0 -
Reframe: environment is the contribution, training is validation
Anurag AgarwalMade-with: Cursorea8263a -
Add full plot deck to Blog and README
Anurag AgarwalMade-with: Cursorf436dde -
Submission polish: Run 7 headline across all surfaces
Anurag AgarwalMade-with: Cursor8cf6fc1 -
Run 7 BEATS BASE (+19%) + UI fixes
Anurag AgarwalMade-with: Cursor50d2fcb -
Blog: human hook opening, simple language
Anurag AgarwalMade-with: Cursorca10a3a -
Judge-ready polish: diagram, before/after, curated replays
Anurag AgarwalMade-with: Cursoraf3c208 -
Fix neon CSS injection for Gradio 6.x + Blog.md at root
Anurag AgarwalMade-with: Cursor8fb3486 -
Update blog + slides with Run 5-7 findings
Anurag AgarwalMade-with: Cursor3a0836e -
Port 7860 + neon theme
Anurag AgarwalMade-with: Cursor657f0fa -
Neon cyberpunk theme - dark bg, glowing accents, stat cards
Anurag AgarwalMade-with: Cursorb4f213a -
Fix Gradio compatibility for openenv's bundled version
Anurag AgarwalMade-with: Cursor712275f -
Add rich 4-tab Gradio UI dashboard
Anurag AgarwalMade-with: Cursore65564f -
Run 6 results + training fixes + all plots regenerated
Anurag AgarwalMade-with: Cursoraae07d0 -
Align eval prompts with training: add required_keys to initial context
Anurag AgarwalMade-with: Cursor5c18f41 -
chore: track remaining PNGs via LFS/xet
Anurag AgarwalAdd LFS tracking for plots/02-05,07 so HF Space accepts them. Made-with: Cursor7cd0fee -
plots: add training progression + diagnostics, drop W&B links
Verified signature099bec8
-
docs: README aligned with hackathon judging criteria (Judges 60s tour + storytelling arc + plot captions)
Verified signature310de9a -
docs: sync README.md (slide deck + auto-validator gate update)
Verified signatureb6de3a6 -
docs: sync SUBMISSION_CHECKLIST.md (slide deck + auto-validator gate update)
Verified signature753d688 -
docs: sync docs/slides.md (slide deck + auto-validator gate update)
Verified signature5e0e1b0 -
docs: add detailed model cards for Run 1 / Run 2 / Run 4
Verified signaturef1678ab -
docs: sync docs/trace_demo_run1.md
Verified signature406b310 -
docs: sync docs/trace_demo.md
Verified signature6dbd9a2 -
docs: sync docs/STATUS.md
Verified signaturef35202f -
docs: sync docs/blog.md
Verified signature82693a1 -
Add plots/ for inline embedding in env Space README
Verified signatureac86191 -
Sync canonical README (KL-anchor narrative, embedded plots, deliverable links)
Verified signature7fe6783 -
env: bump max_concurrent_envs from 8 to 64
Verified signature8a9cc57 -
eval: enforce one-tool-call response format on every turn
Verified signaturea22fcfd -
Fix parser: handle quoted commas, balanced parens, ASK:/PROPOSE: prefixes
Verified signature7c0cc92 -
Eval system prompt: align character-for-character with training PROMPT — ensures trained model has zero distribution shift between train and eval
Verified signatured9beb62 -
Eval system prompt: drop misleading software-stack example, align with training PROMPT (forces model to use task-family fields, not copy the example verbatim)
Verified signatureef5498c -
Parser: support ASK:/PROPOSE:/Q:/PLAN: prefix forms produced by Qwen3 GRPO
Verified signatureb8a5922 -
inference: parser fix — handle key=value in func calls + balanced parens
Verified signaturef251890 -
fix(eval): pass enable_thinking=False to disable Qwen3 thinking + bump MAX_TOKENS to 800
Verified signaturee4d1233 -
feat: add run_eval.py to Space (needed by eval_with_vllm.py for trained-model evals)
Verified signature6473a24 -
env: enable concurrent rollout sessions (max_concurrent_envs=8)
Anurag Agarwal895f00d -
rewrite training notebook with cleaner cell-by-cell structure
Anurag Agarwala45e7e7 -
Add training/train_grpo.ipynb — GRPO training notebook (TRL + vLLM + ClarifyEnv)
Anurag Agarwal5e8f794