garn-garn's picture
Add anonymized executable code package
881f9f2 verified
# Core
torch>=2.0
transformers
datasets
huggingface-hub
# MCP + orchestration client stack
mcp[cli]==1.27.0
openai-agents==0.14.5
litellm==1.81.13
# WatsonX.ai (LLM-as-Judge, scaling comparison)
ibm-watsonx-ai
# Profiling and tracking
wandb
tensorboard
# Data
pandas
numpy
scipy # statistical-fidelity validation: KS, AD, Wasserstein, chi-squared
openpyxl # IEEE DataPort DGA dataset is published as .xlsx; pandas needs this engine to read it
matplotlib
seaborn
# Reproducible notebook execution
nbconvert
nbclient
ipykernel
# Utilities
tqdm
pyyaml
# LLM-as-Judge scoring (scripts/judge_trajectory.py)
litellm
python-dotenv