vectrayx-paper-code / requirements.txt
jsantillana's picture
Upload requirements.txt with huggingface_hub
ef44a3b verified
raw
history blame contribute delete
359 Bytes
# Core — exact versions used in the paper experiments
torch==2.11.0
sentencepiece==0.2.1
numpy==2.4.2
# AWS (SageMaker + S3)
boto3==1.42.93
sagemaker==3.10.0
# HuggingFace (external datasets and baselines)
datasets>=2.19.0
transformers>=4.40.0
peft>=0.10.0
# Evaluation
scikit-learn>=1.3.0
# GGUF export (optional, for llama.cpp / Ollama)
# gguf>=0.6.0