search

OpenPPL / ppq

PPL Quantization Tool (PPQ) is a powerful offline neural network quantization tool.
Apache License 2.0
1.56k stars 236 forks source link

trt onnxParser #511

Closed eezhang123 closed 1 year ago

eezhang123 commented 1 year ago

根据ReadME拉了镜像，启动了容器(不需要再安装trt了吧) 跑例子 run_inference 报错 https://github.com/openppl-public/ppq/blob/master/md_doc/deploy_trt_by_OnnxParser.md

Traceback (most recent call last): File "trt_inference.py", line 77, in inputs_alloc_buf, outputs_alloc_buf, bindings_alloc_buf, stream_alloc_buf = alloc_buf_N(engine,data) File "trt_inference.py", line 20, in alloc_buf_N stream = cuda.Stream() pycuda._driver.LogicError: explicit_context_dependent failed: invalid device context - no currently active context?

eezhang123 commented 1 year ago

解决了 pycuda需要初始化，加一行import pycuda.autoinit

ZhangZhiPku commented 1 year ago

哦豁，完蛋