From 5bf9a80283e56776a5a8d2f215e70886086aca10 Mon Sep 17 00:00:00 2001 From: Adil Hafeez Date: Tue, 10 Dec 2024 21:03:02 -0800 Subject: [PATCH] remove nli model --- arch/tools/cli/consts.py | 1 - crates/common/src/consts.rs | 1 - model_server/Dockerfile | 2 +- model_server/Dockerfile.gpu | 2 +- 4 files changed, 2 insertions(+), 4 deletions(-) diff --git a/arch/tools/cli/consts.py b/arch/tools/cli/consts.py index 598bc29c..292d5e63 100644 --- a/arch/tools/cli/consts.py +++ b/arch/tools/cli/consts.py @@ -2,7 +2,6 @@ KATANEMO_DOCKERHUB_REPO = "katanemo/archgw" KATANEMO_LOCAL_MODEL_LIST = [ "katanemo/Arch-Guard-cpu", "katanemo/Arch-Guard", - "katanemo/bge-large-en-v1.5", ] SERVICE_NAME_ARCHGW = "archgw" SERVICE_NAME_MODEL_SERVER = "model_server" diff --git a/crates/common/src/consts.rs b/crates/common/src/consts.rs index 28db9fb8..87b661ca 100644 --- a/crates/common/src/consts.rs +++ b/crates/common/src/consts.rs @@ -1,4 +1,3 @@ -pub const DEFAULT_EMBEDDING_MODEL: &str = "katanemo/bge-large-en-v1.5"; pub const DEFAULT_INTENT_MODEL: &str = "katanemo/bart-large-mnli"; pub const DEFAULT_PROMPT_TARGET_THRESHOLD: f64 = 0.8; pub const DEFAULT_HALLUCINATED_THRESHOLD: f64 = 0.25; diff --git a/model_server/Dockerfile b/model_server/Dockerfile index d2bd9fa6..1c9d5ff4 100644 --- a/model_server/Dockerfile +++ b/model_server/Dockerfile @@ -15,7 +15,7 @@ WORKDIR /src # specify list of models that will go into the image as a comma separated list # following models have been tested to work with this image # "sentence-transformers/all-MiniLM-L6-v2,sentence-transformers/all-mpnet-base-v2,thenlper/gte-base,thenlper/gte-large,thenlper/gte-small" -ENV MODELS="katanemo/bge-large-en-v1.5-onnx" +ENV MODELS="" COPY ./app ./app COPY ./app/guard_model_config.yaml . diff --git a/model_server/Dockerfile.gpu b/model_server/Dockerfile.gpu index aba65edd..b1b29c06 100644 --- a/model_server/Dockerfile.gpu +++ b/model_server/Dockerfile.gpu @@ -45,7 +45,7 @@ RUN if command -v nvcc >/dev/null 2>&1; then \ COPY . /src # Specify list of models that will go into the image as a comma separated list -ENV MODELS="katanemo/bge-large-en-v1.5-onnx" +ENV MODELS="" ENV DEBIAN_FRONTEND=noninteractive COPY /app /app