triton-inference-server / dali_backend

The Triton backend that allows running GPU-accelerated data pre-processing pipelines implemented in DALI's python API.
https://docs.nvidia.com/deeplearning/dali/user-guide/docs/index.html
MIT License
118 stars 28 forks source link

Video Input larger than max #241

Open wq9 opened 3 months ago

wq9 commented 3 months ago

fn.inputs.video is supposed to handle large videos. However, with large videos (>2GB), I get the error below immediately:

[StatusCode.RESOURCE_EXHAUSTED] Sent message larger than max (3316742405 vs. 2147483647)

So I set tritonclient.grpc.InferenceServerClient's channel_args with ("grpc.max_send_message_length", -1). I even added compression_algorithm='gzip' to triton_client.start_stream and the size barely got any smaller.

After waiting a long time, I get error below:

[StatusCode.RESOURCE_EXHAUSTED] Received message larger than max (3303596386 vs. 2147483647)

("grpc.max_receive_message_length", -1) on the triton_client doesn't work.

I suppose the grpc.max_receive_message_length needs to be set on the triton server, so how do I do that?