mirror of
https://github.com/NicolasBohn/NexQuant.git
synced 2026-07-27 15:37:44 +00:00
TPTBusiness
ea87495a6a
Replace sequential predict() calls with predict_batch() in both build_kronos_factor and evaluate_kronos_model. Up to batch_size windows are processed simultaneously on GPU, reducing per-window time from ~10s to ~0.13s (measured: 10 windows in 1.3s on RTX 5060 Ti). Adds --batch-size / -b option (default 32) to both kronos-factor and kronos-eval CLI commands. Falls back to single inference per window if a batch fails. Refactors timestamp preparation into _build_window_inputs. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
108 lines
1.8 KiB
Plaintext
108 lines
1.8 KiB
Plaintext
# Requirements for runtime.
|
|
pydantic-settings
|
|
|
|
python-Levenshtein
|
|
scikit-learn
|
|
filelock
|
|
loguru
|
|
psutil
|
|
fire
|
|
fuzzywuzzy
|
|
openai
|
|
litellm>=1.73 # to support `from litellm import get_valid_models`
|
|
aiohttp>=3.13.4 # CVE-2026-22815, CVE-2026-34515, CVE-2026-34516, CVE-2026-34525
|
|
azure.identity
|
|
pyarrow
|
|
rich
|
|
tqdm
|
|
typer
|
|
|
|
numpy # we use numpy as default data format. So we have to install numpy
|
|
pandas # we use pandas as default data format. So we have to install pandas
|
|
pandarallel # parallelize pandas
|
|
matplotlib
|
|
langchain
|
|
langchain-community
|
|
tiktoken
|
|
pymupdf # Extract shotsreens from pdf
|
|
|
|
# PDF related
|
|
pypdf
|
|
azure-ai-formrecognizer
|
|
|
|
# factor implementations
|
|
tables
|
|
|
|
# CI Fix Tool
|
|
tree-sitter-python
|
|
tree-sitter
|
|
|
|
python-dotenv
|
|
|
|
# infrastructure related.
|
|
docker
|
|
|
|
# crawler related
|
|
webdriver-manager
|
|
|
|
# demo related
|
|
streamlit>=1.47 # to support input_c.text_area(..., height="content", ...)
|
|
plotly
|
|
st-theme
|
|
randomname
|
|
flask
|
|
flask-cors
|
|
networkx
|
|
|
|
# kaggle crawler
|
|
selenium
|
|
kaggle
|
|
nbformat # also used for notebook conversion
|
|
|
|
# tool
|
|
setuptools-scm
|
|
seaborn
|
|
azure.ai.inference
|
|
|
|
# data folder desc
|
|
humanize
|
|
genson
|
|
|
|
# mlflow
|
|
mlflow
|
|
azureml-mlflow
|
|
types-pytz
|
|
|
|
# Agent
|
|
pydantic-ai-slim[mcp,openai,prefect]
|
|
nest-asyncio
|
|
|
|
# visualize SFT train
|
|
tensorboard # tensorboard --logdir git_ignore_folder/RD-Agent_workspace
|
|
prefect
|
|
|
|
# HuggingFace datasets
|
|
datasets
|
|
|
|
# DuckDuckGo search
|
|
duckduckgo-search
|
|
|
|
# Testing
|
|
pytest
|
|
pytest-cov
|
|
|
|
# Parameter Optimization
|
|
optuna>=3.5.0
|
|
|
|
# News & Data (Polymarket, ForexFactory, CryptoPanic)
|
|
beautifulsoup4>=4.12.0
|
|
|
|
# ML Training Pipeline
|
|
lightgbm>=3.3.0
|
|
scipy>=1.9.0
|
|
|
|
# RL Trading (optional - system works without these)
|
|
# Install for full RL training: pip install stable-baselines3[extra] gymnasium
|
|
# Without these, RL trading uses simple momentum fallback
|
|
# stable-baselines3[extra]>=2.0.0
|
|
# gymnasium>=0.29.0 |