Switch default deploy model (#18)

deployment/docker-compose.yml

@@ -4,7 +4,7 @@ services:
   init:
     image: tabbyml/tabby
     container_name: tabby-init
-    command: python -m tabby.tools.model_preload --repo_id TabbyML/NeoX-70M
+    command: python -m tabby.tools.model_preload --repo_id TabbyML/J-350M
     volumes:
       - ../data/hf_cache:/root/.cache/huggingface
 
@@ -13,7 +13,7 @@ services:
     container_name: tabby-server
     command: uvicorn tabby.server:app --host 0.0.0.0 --port 5000
     environment:
-      - MODEL_NAME=TabbyML/NeoX-70M
+      - MODEL_NAME=TabbyML/J-350M
       - MODEL_BACKEND=triton
     ports:
       - "5000:5000"
@@ -48,7 +48,7 @@ services:
               count: all
               capabilities: [gpu]
     environment:
-      - MODEL_NAME=TabbyML/NeoX-70M
+      - MODEL_NAME=TabbyML/J-350M
     depends_on:
       init:
         condition: service_completed_successfully

tabby/tools/model_preload.py

@@ -31,7 +31,7 @@ def preload(local_files_only=False):
 
 if __name__ == "__main__":
     args = parse_args()
-    print(f"Loading {args.repo_id} ...")
+    print(f"Loading {args.repo_id}, this will take a while...")
     try:
         preload(local_files_only=args.prefer_local_files)
     except Exception as e: