Error on Multi-GPU when running fine-tuning: Input module

unsloth multi gpu  How Does Multi GPU Work in Common Deep Learning Frameworks? TensorFlow Multiple GPU; PyTorch Multi GPU; Multi GPU Deployment Models; GPU Server; GPU Cluster Unsloth is planning multi-GPU support in a future release Guidelines FAQ Lists API Security Legal

I run through an error when running my fine-tuning code that uses Unsloth The error only happened when using a multi-gpu setup  multi-gpu setup Also, since apparently a larger batch size was working before I would also recommend trying to reproduce this setup to see

10x faster on a single GPU and up to 32x faster on multiple GPU systems compared to Flash Attention 2 We support NVIDIA GPUs from Tesla T4 to H100, and The repository relies on key technologies such as the Triton compiler for writing high-performance GPU kernels and the PyTorch library for