How to use anindya64/tinyllama-tensorrt with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("anindya64/tinyllama-tensorrt", dtype="auto")
This repo contains the TensorRT LLM version of TinyLlama Model. The conversion is done to support Float16 precision on Nvidia TensorRT.
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("anindya64/tinyllama-tensorrt", dtype="auto")