From 52ad3b383582dc199bc416726d0dc61187b1f734 Mon Sep 17 00:00:00 2001 From: Qing Lan Date: Mon, 5 Dec 2022 09:24:09 -0800 Subject: [PATCH] unset min length --- .github/workflows/llm_integration.yml | 2 +- tests/integration/llm/client.py | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/.github/workflows/llm_integration.yml b/.github/workflows/llm_integration.yml index 7db2a98b1b..3c2e31ae3c 100644 --- a/.github/workflows/llm_integration.yml +++ b/.github/workflows/llm_integration.yml @@ -113,7 +113,7 @@ jobs: run: | sudo python3 llm/prepare.py huggingface gpt-neo-2.7b ./launch_container.sh deepjavalibrary/djl-serving:$DJLSERVING_DOCKER_TAG $PWD/models deepspeed \ - serve + serve -m test:::*=file:///opt/ml/model/test/ python3 llm/client.py huggingface gpt-neo-2.7b docker rm -f $(docker ps -aq) - name: Test bloom-7b diff --git a/tests/integration/llm/client.py b/tests/integration/llm/client.py index d3cadede3a..a5f3ee4d55 100644 --- a/tests/integration/llm/client.py +++ b/tests/integration/llm/client.py @@ -73,7 +73,7 @@ def test_hf_model(model): for batch_size in spec["batch_size"]: for seq_length in spec["seq_length"]: req = {"inputs": batch_generation(batch_size)} - params = {"min_length": seq_length, "max_length": seq_length} + params = {"max_length": seq_length} req["parameters"] = params res = send_json(req) logging.info(f"res {res}")