TensorRT-LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and build TensorRT engines that contain state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT-LLM also contains components to create Python and C++ runtimes that execute those TensorRT engines.
For tensor parallelism, sharing the embedding table must setuse_parallel_embedding=True and embedding_sharding_dim=0
-> For tensor parallelism, sharing the embedding table must set use_parallel_embedding=True and embedding_sharding_dim=0
For tensor parallelism, sharing the embedding table must
setuse_parallel_embedding=Trueand embedding_sharding_dim=0 -> For tensor parallelism, sharing the embedding table mustset use_parallel_embedding=Trueand embedding_sharding_dim=0