miniTorch_sentiment

Assignment 1 of 11-868 LLM Systems.

Goal: A basic deep learning framework capable of CUDA-optimized tensor operations and automatic differentiation.

• Model is a simple neural network for sentiment classification

1. Dependencies

$ python -m pip install -r requirements.txt

$ python -m pip install -r requirements.extra.txt

$ python -m pip install -Ue .

Ensure NVIDIA Drivers are installed

2. Compile the CUDA kernels

$ nvcc -o minitorch/cuda_kernels/combine.so --shared src/combine.cu -Xcompiler -fPIC

3. Train Model

$ python project/run_sentiment.py