fix: update huggingface cli commands to 'hf' and ensure correct installation

.github/workflows/deploy-gpu.yml

@@ -22,15 +22,13 @@ jobs:
         with:
           python-version: '3.x'
 
-      - name: Install HuggingFace CLI
-        run: |
-          pip install -U "huggingface_hub[cli]"
-          
+      - name: Install huggingface_hub
+        run: pip install huggingface_hub
+
       - name: Login to HuggingFace
+        run: hf auth login --token $HF_TOKEN
         env:
           HF_TOKEN: ${{ secrets.HF_TOKEN }}
-        run: |
-          huggingface-cli login --token $HF_TOKEN
           
       - name: Prepare deployment files
         env:
@@ -119,7 +117,7 @@ jobs:
           echo "Creating/verifying Space: $HF_SPACE_NAME"
           
           # Try to create Space (will fail if exists, which is OK)
-          huggingface-cli repo create $HF_SPACE_NAME --type space --space_sdk docker || echo "Space already exists or creation failed - continuing..."
+          hf repo create $HF_SPACE_NAME --type space --space_sdk docker || echo "Space already exists or creation failed - continuing..."
           
           echo "✅ Space ready"
 
@@ -132,7 +130,7 @@ jobs:
           echo "Uploading to HuggingFace Space: $HF_SPACE_NAME"
           
           # Use HF CLI to upload entire folder
-          huggingface-cli upload \
+          hf upload \
             "$HF_SPACE_NAME" \
             ./hf-deploy \
             . \