TensorRT-LLM
TensorRT-LLM copied to clipboard
Error Code 9: Internal Error (Builder was created on device different than current device. qwen-7B INT4-GPTQ on A100
i try to run qwen-7B INT4-GPTQ example, error comes:Error Code 9: Internal Error (Builder was created on device different than current device. my gpu are two A100 80G,please help~~
! python3 /root/TensorRT-LLM/examples/qwen/build.py --hf_model_dir Qwen-7B-Chat-Int4
--quant_ckpt_path Qwen-7B-Chat-Int4
--dtype float16
--remove_input_padding
--use_gpt_attention_plugin float16
--enable_context_fmha
--use_gemm_plugin float16
--use_weight_only
--weight_only_precision int4_gptq
--per_group
--world_size 1
--tp_size 1
--output_dir ./tmp/Qwen/7B/trt_engines/int4-gptq/1-gpu
[01/29/2024-03:10:12] [TRT] [E] 9: [backendBuilderConfig.cpp::validateBasicAssumptions::28] Error Code 9: Internal Error (Builder was created on device different than current device.)
[01/29/2024-03:10:12] [TRT-LLM] [E] Engine building failed, please check the error log.
[01/29/2024-03:10:12] [TRT-LLM] [I] Config saved to ./tmp/Qwen/7B/trt_engines/int4-gptq/1-gpu/config.json.
Traceback (most recent call last):
File "/root/TensorRT-LLM/examples/qwen/build.py", line 710, in
Please follow the issue template to file issue, thank you for cooperation.
faced the same issue
@wujingbo-web @zeeshanvision Try specify CUDA device.
CUDA_VISIBLE_DEVICES=0 python build.py --...
Do you use MIG? Also, please follow the template to share the reproduced steps and the full error logs.