Temporarily disable the finetuned till we get it working again.

This is preventing the embedding service from starting.  Via the web browser you will still get a big error if you pick one of the fine tuned models.

embeddings/app/main.py

@@ -1,11 +1,11 @@
 from fastapi import FastAPI
 from clipL14 import router as router_clip
 from minilm import router as router_minilm
-from finetuned import router as router_finetuned
+#from finetuned import router as router_finetuned
 
 app = FastAPI()
 
 app.include_router(router_clip.router, prefix="/clip")
 app.include_router(router_minilm.router, prefix="/minilm")
-app.include_router(router_finetuned.sbert_router, prefix="/ftbert")
-app.include_router(router_finetuned.clip_router, prefix="/ftclip")
+#app.include_router(router_finetuned.sbert_router, prefix="/ftbert")
+#app.include_router(router_finetuned.clip_router, prefix="/ftclip")

embeddings/requirements.txt

@@ -1,3 +1,4 @@
+docarray==0.21.0 # See https://github.com/querqy/chorus/issues/169
 sentence-transformers
 fastapi
 uvicorn[standard]