Skip to content

Comments

Correct version of onnx-tensorrt for trt inference added to Dockerfile. Explicit type conversion.#399

Closed
denti wants to merge 2 commits intoNVIDIA:masterfrom
denti:master
Closed

Correct version of onnx-tensorrt for trt inference added to Dockerfile. Explicit type conversion.#399
denti wants to merge 2 commits intoNVIDIA:masterfrom
denti:master

Conversation

@denti
Copy link

@denti denti commented Feb 13, 2020

No description provided.

Copy link
Contributor

@yzhang123 yzhang123 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for looking at this! The top level Dockerfile should not be the one to be changed,
but the one inside trt/Dockerfile. For this purpose please merge this PR
#403

signal_shape=(args.engine_batch_size, int(args.seq_len), 64)
else:
signal_shape=(args.engine_batch_size, 64, args.seq_len)
signal_shape=(args.engine_batch_size, 64, int(args.seq_len))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

thanks for catching this bug!

COPY trt/requirements.txt ./trt_requirements.txt
RUN pip install --disable-pip-version-check -U -r requirements.txt -r trt_requirements.txt

COPY ./trt/onnx-trt.patch .
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

please do not change this Dockerfile. We have different Dockfiles for different purposes. This Dockerfile is for pure PyTorch training of Jasper
Everything TRT-relased is in the Dockerfile in PyTorch/SpeechRecognition/Jasper/TRT/Dockerfile

@nvpstr nvpstr closed this Mar 9, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants