Upload README.md

Upload 2 files

feat: Charts + Flood Map tabs

feat: Charts + Flood Map tabs

feat: Charts tab + Flood Map tab with Leaflet + GEE tiles

fix: GEE key_data must be JSON string not dict

feat: dotenv + HF trained model + safe GEE imports

fix: added env

fix: update gee_codegen.py with generate_gee_code function

fix: add gee_codegen.py and gee_client.py to Dockerfile

fix: restore index.html + add map.html separately

fix: adding variable region gee_codegen.py

feat: GEE code download for all 15 basins + flood risk visualisation

feat: full 4-agent pipeline wired into /agent/step

feat: trained RL model inference + compare endpoint

results: all training plots

feat: gold/black frontend + India map + compare

feat: trained RL model inference pipeline

feat: 15 regions seasonal data

fix: remove token + cleanup duplicate multi_agent file

results: SFT loss curve + 3B medium GRPO reward curve

README

Merge branch 'main' of https://github.com/lunasingha1928-oss/Chronostasis

Replace single agent with 4-agent multi-agent system

Add files via upload

results: SFT training loss curve 1.48 -> 0.017

fix: resolve conflict

Add 4 agent multi-agent system scoring 0.99

Add 4 agent multi-agent system scoring 0.99

Add 4 agent multi-agent system scoring 0.99

fix: vague baseline fallbacks + React frontend

add visual report: 4 charts auto-render after episode

add visual report: 4 charts auto-render after episode

add renderer.py to Dockerfile

run server.py directly - has render endpoint

trigger rebuild

inference runs all 3 tasks for validator

fix scores strictly between 0 and 1 (not 0.0 or 1.0)

fix ValueError in tasks.py - safe float parsing

fix inference.py - remove my_env_v4 import, use HTTP client

fix entry point to server.app:main

proper standalone server/app.py with main() function

fix openenv validate: add uv.lock, server/app.py, openenv-core dep

add pyproject.toml for openenv validate

update inference.py

fix multi-replica: agent_step self-contained, passes task+region

fix ValueError empty reward string

force push latest agent_step with error handling

fix model name to llama-3.3-70b-versatile

fix NameError - GROQ_API_KEY variable not defined