requirements.txt

# MLLM Adversarial Attack Framework Requirements
# Based on research implementation for "Universal Adversarial Attack on Aligned Multimodal LLMs"

# Core ML and Deep Learning
torch>=2.0.0
torchvision>=0.15.0
transformers>=4.35.0
accelerate>=0.20.0
datasets>=2.14.0

# Image Processing
Pillow>=10.0.0
opencv-python>=4.8.0
imageio>=2.35.0
imageio-ffmpeg>=0.5.0

# Scientific Computing
numpy>=1.26.0
scipy>=1.11.0
pandas>=2.0.0
matplotlib>=3.7.0
seaborn>=0.12.0

# Safety and Evaluation
outlines>=0.0.40
pydantic>=2.0.0
jsonschema>=4.23.0

# Experiment Tracking and Logging
wandb>=0.15.0
tensorboard>=2.13.0
tqdm>=4.66.0

# Utilities
argparse
pathlib
glob
re
json
logging
datetime
random
os
sys
typing
collections
importlib

# Optional: CUDA Support (if using GPU)
# nvidia-ml-py>=12.560.0

# Development and Testing
pytest>=7.4.0
black>=24.10.0
isort>=5.13.0
flake8>=6.0.0

# Additional Dependencies from Conda Environment
absl-py>=2.1.0
addict>=2.4.0
aenum>=3.1.15
aiofiles>=23.2.1
anyio>=4.4.0
apscheduler>=3.11.0
argcomplete>=3.5.1
attrs>=24.2.0
babel>=2.17.0
blis>=0.7.11
blobfile>=2.1.1
botok>=0.8.12
catalogue>=2.0.10
charset-normalizer>=3.3.2
click>=8.1.7
filelock>=3.13.1
fire>=0.7.0
frozenlist>=1.4.1
fsspec>=2024.2.0
genson>=1.3.0
gitdb>=4.0.11
gitpython>=3.1.43
grpcio>=1.70.0
hasami>=0.0.1
hjson>=3.1.0
hopcroftkarp>=1.2.5
httpcore>=1.0.5
httpx>=0.27.2
hydra-core>=1.3.2
imagesize>=1.4.1
importlib-metadata>=7.2.1
importlib-resources>=6.4.5
indic-nlp-library>=0.92
inflect>=5.6.2
interegular>=0.3.3
jinja2>=3.1.3
joblib>=1.4.2
jsonlines>=4.0.0
jsonschema-specifications>=2023.12.1
khmer-nltk>=1.6
langcodes>=3.4.0
language-data>=1.2.0
laonlp>=1.2.0
lark>=1.2.2
llvmlite>=0.43.0
morfessor>=2.0.6
mpmath>=1.3.0
multiprocess>=0.70.16
murmurhash>=1.0.10
nbformat>=5.10.4
networkx>=3.2.1
numba>=0.60.0
nvidia-ml-py>=12.560.0
overrides>=7.7.0
packaging>=23.2
pathspec>=0.12.1
sklearn-crfsuite>=0.5.0
smart-open>=7.0.4
smmap>=5.0.1
srsly>=2.4.8
starlette>=0.38.5
stopes>=2.2.1
submitit>=1.5.2
tomlkit>=0.12.0
typer>=0.12.5
tzlocal>=5.2
urllib3>=2.2.3
wrapt>=1.16.0
xxhash>=3.5.0
yarl>=1.11.1

# Model-Specific Dependencies
# These may be needed depending on which models you use
# sentencepiece>=0.1.99
# protobuf>=3.20.0
# safetensors>=0.3.0

# Optional: For advanced features
# ray>=2.7.0  # For distributed training
# optuna>=3.4.0  # For hyperparameter optimization