BasicCoder

### System Info - x86_64 - NVIDIA H20 - 96GB - TensorRT-LLM version: 0.11.0.dev2024051400 ### Who can help? @Tracin ### Information - [X] The official example scripts - [ ]...

InternVL2.0 and InternVL1.5 have the same architecture, maybe the download statistics needs to be reconsidered. Similar issue: #1567

Fix prompt_table_data empty tensor shape error

2

When executing LLM in a VLM model alone, the correct prompt_table_data's hidden_size = self.hidden_size * tp_size, because self.hidden_size = pretrained_config.hidden_size // tp_size

The baichuan convert script lacks `scale_y_accum_quant`, `scale_w_quant_orig` value saving.

Update the current default trt_llm version from 0.15.0 to 0.16.0

draft pr about non-streaming output

1

Try to get a valid output of the model, i.e. do not include the input of the model in the case of non-streaming