fix logbar api usage (#1345)

* fix logbar api usage

* [CI] install logbar

Author: CSY-ModelCloud

.github/workflows/unit_tests.yml

@@ -387,7 +387,7 @@ jobs:
 
       - name: Install wheel
         run: |
-          uv pip install colorlog git+https://github.com/ModelCloud/Tokenicer -U
+          uv pip install colorlog logbar git+https://github.com/ModelCloud/Tokenicer -U
           echo "===== install optimum bitblas parameterized uvicorn ====="
           uv pip install optimum bitblas==0.0.1.dev13 parameterized uvicorn -i http://${{ needs.check-vm.outputs.ip }}/simple/ --trusted-host ${{ needs.check-vm.outputs.ip }} --extra-index-url https://pypi.org/simple
           echo "===== install dist/whl ====="
@@ -547,7 +547,7 @@ jobs:
 
       - name: Install wheel
         run: |
-          uv pip install -U transformers colorlog
+          uv pip install -U transformers colorlog logbar
           if [ "${{ matrix.test_script }}" == "test_perplexity" ] || \
              [ "${{ matrix.test_script }}" == "test_q4_bitblas" ] || \
              [ "${{ matrix.test_script }}" == "test_save_loaded_quantized_model" ]; then

tests/inference_speed.py

@@ -95,7 +95,7 @@ def inference(self, model_path, backend, tokens_per_second, assert_result=True,
             print(f"New Token Per Second: {avg_tokens_per_second} token/s")
             print(f"****************  {backend} Warm-up Result Info End****************")
 
-        pb = ProgressBar(range(self.NUM_RUNS)).title("Run")
+        pb = logger.pb(range(self.NUM_RUNS)).title("Run")
         for _ in pb:
             start_time = time.time()
             result = model.generate(**inp, max_new_tokens=self.MAX_NEW_TOKENS, pad_token_id=tokenizer.pad_token_id)