Testing

.github/workflows/docker-build-push-backend-container-on-tag.yml

@@ -31,14 +31,5 @@ jobs:
         push: true
         tags: |
           danswer/danswer-backend:${{ github.ref_name }}
-          danswer/danswer-backend:latest
         build-args: |
           DANSWER_VERSION=${{ github.ref_name }}
-
-    - name: Run Trivy vulnerability scanner
-      uses: aquasecurity/trivy-action@master
-      with:
-        # To run locally: trivy image --severity HIGH,CRITICAL danswer/danswer-backend
-        image-ref: docker.io/danswer/danswer-backend:${{ github.ref_name }}
-        severity: 'CRITICAL,HIGH'
-        trivyignores: ./backend/.trivyignore

backend/model_server/custom_models.py

@@ -3,8 +3,10 @@
 import numpy as np
 import tensorflow as tf  # type: ignore
 from fastapi import APIRouter
+from huggingface_hub.constants import HF_HOME
 from transformers import AutoTokenizer  # type: ignore
 from transformers import TFDistilBertForSequenceClassification
+from transformers.utils import TRANSFORMERS_CACHE
 
 from model_server.constants import MODEL_WARM_UP_STRING
 from model_server.utils import simple_log_function_time
@@ -24,6 +26,8 @@
 def get_intent_model_tokenizer(
     model_name: str = INTENT_MODEL_VERSION,
 ) -> "AutoTokenizer":
+    print(f"HF Home: {HF_HOME}")
+    print(f"Cache dir: {TRANSFORMERS_CACHE}")
     global _INTENT_TOKENIZER
     if _INTENT_TOKENIZER is None:
         _INTENT_TOKENIZER = AutoTokenizer.from_pretrained(model_name)
@@ -36,6 +40,8 @@ def get_local_intent_model(
 ) -> TFDistilBertForSequenceClassification:
     global _INTENT_MODEL
     if _INTENT_MODEL is None or max_context_length != _INTENT_MODEL.max_seq_length:
+        print(f"HF Home: {HF_HOME}")
+        print(f"Cache dir: {TRANSFORMERS_CACHE}")
         _INTENT_MODEL = TFDistilBertForSequenceClassification.from_pretrained(
             model_name
         )