Natarajan-Networks/grading-env

Add score field to END log line

Fix max_marks to be strictly less than 1

Fix step to return 4 values and clamp all scores strictly between 0 and 1

Restrict marks_awarded to be strictly between 0 and 1

Fix all scores to never be exactly 0 or 1

Fix reward scores to be strictly between 0.05 and 0.95

Fix scores to be strictly between 0.01 and 0.99

Fix scores to be strictly between 0 and 1

Final improvements - ready for submission

Redesign with 3 distinct task types: Factual, Conceptual and Essay grading

Fix question extraction from environment

Fix reset API to return question

Fix step API input handling

Fix question + feedback display

Force rebuild

Add interactive UI with live grading

Fix app.py syntax error and update UI

Update homepage task descriptions

Improve UI, health endpoint and overall code quality

Improve models, openenv.yaml, README, and Dockerfile

Improve inference.py to use LLM grading with HF Space API

Remove test file

Improve environment with random questions, better reward function and updated data

Improve data files, README, and openenv.yaml

Fix inference.py to match hackathon requirements

Simplify inference.py to avoid timeout

Fix log format to match sample inference script

Fix KeyError in inference.py

Fix openenv validation issues

Update openai version requirement

Add inference.py with required log format

Fix PYTHONPATH in Dockerfile

Deploy grading environment

Fix Dockerfile name

Add grading environment

initial commit