text-generation-inference

0e9d249b - feat(benchmark): add support for private tokenizers (#262)