Merge branch 'mlc-ai:main' into main

Author: nihalgeorge01

docs/install/mlc_llm.rst

@@ -30,35 +30,35 @@ Select your operating system/compute platform and run the command in your termin
                 .. code-block:: bash
 
                     conda activate your-environment
-                    python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-llm-nightly-cpu mlc-ai-nightly-cpu
+                    python -m pip install --pre -U -f https://mlc.ai/wheels mlc-llm-nightly-cpu mlc-ai-nightly-cpu
 
             .. tab:: CUDA 12.2
 
                 .. code-block:: bash
 
                     conda activate your-environment
-                    python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-llm-nightly-cu122 mlc-ai-nightly-cu122
+                    python -m pip install --pre -U -f https://mlc.ai/wheels mlc-llm-nightly-cu122 mlc-ai-nightly-cu122
 
             .. tab:: CUDA 12.3
 
                 .. code-block:: bash
 
                     conda activate your-environment
-                    python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-llm-nightly-cu123 mlc-ai-nightly-cu123
+                    python -m pip install --pre -U -f https://mlc.ai/wheels mlc-llm-nightly-cu123 mlc-ai-nightly-cu123
 
             .. tab:: ROCm 6.1
 
                 .. code-block:: bash
 
                     conda activate your-environment
-                    python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-llm-nightly-rocm61 mlc-ai-nightly-rocm61
+                    python -m pip install --pre -U -f https://mlc.ai/wheels mlc-llm-nightly-rocm61 mlc-ai-nightly-rocm61
 
             .. tab:: ROCm 6.2
 
                 .. code-block:: bash
 
                     conda activate your-environment
-                    python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-llm-nightly-rocm62 mlc-ai-nightly-rocm62
+                    python -m pip install --pre -U -f https://mlc.ai/wheels mlc-llm-nightly-rocm62 mlc-ai-nightly-rocm62
 
             .. tab:: Vulkan
 
@@ -98,7 +98,7 @@ Select your operating system/compute platform and run the command in your termin
                 .. code-block:: bash
 
                     conda activate your-environment
-                    python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-llm-nightly-cpu mlc-ai-nightly-cpu
+                    python -m pip install --pre -U -f https://mlc.ai/wheels mlc-llm-nightly-cpu mlc-ai-nightly-cpu
 
         .. note::
 
@@ -124,7 +124,7 @@ Select your operating system/compute platform and run the command in your termin
                 .. code-block:: bash
 
                     conda activate your-environment
-                    python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-llm-nightly-cpu mlc-ai-nightly-cpu
+                    python -m pip install --pre -U -f https://mlc.ai/wheels mlc-llm-nightly-cpu mlc-ai-nightly-cpu
 
         .. note::
             Please make sure your conda environment comes with python and pip.

docs/install/tvm.rst

@@ -37,35 +37,35 @@ A nightly prebuilt Python package of Apache TVM Unity is provided.
             .. code-block:: bash
 
               conda activate your-environment
-              python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-ai-nightly-cpu
+              python -m pip install --pre -U -f https://mlc.ai/wheels mlc-ai-nightly-cpu
 
          .. tab:: CUDA 12.2
 
             .. code-block:: bash
 
               conda activate your-environment
-              python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-ai-nightly-cu122
+              python -m pip install --pre -U -f https://mlc.ai/wheels mlc-ai-nightly-cu122
 
          .. tab:: CUDA 12.3
 
             .. code-block:: bash
 
               conda activate your-environment
-              python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-ai-nightly-cu123
+              python -m pip install --pre -U -f https://mlc.ai/wheels mlc-ai-nightly-cu123
 
          .. tab:: ROCm 6.1
 
             .. code-block:: bash
 
               conda activate your-environment
-              python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-ai-nightly-rocm61
+              python -m pip install --pre -U -f https://mlc.ai/wheels mlc-ai-nightly-rocm61
 
          .. tab:: ROCm 6.2
 
             .. code-block:: bash
 
               conda activate your-environment
-              python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-ai-nightly-rocm62
+              python -m pip install --pre -U -f https://mlc.ai/wheels mlc-ai-nightly-rocm62
 
          .. tab:: Vulkan
 
@@ -88,7 +88,7 @@ A nightly prebuilt Python package of Apache TVM Unity is provided.
             .. code-block:: bash
 
               conda activate your-environment
-              python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-ai-nightly-cpu
+              python -m pip install --pre -U -f https://mlc.ai/wheels mlc-ai-nightly-cpu
 
         .. note::
 
@@ -109,7 +109,7 @@ A nightly prebuilt Python package of Apache TVM Unity is provided.
             .. code-block:: bash
 
               conda activate your-environment
-              python -m pip install --pre -U --no-index -f https://mlc.ai/wheels mlc-ai-nightly-cpu
+              python -m pip install --pre -U -f https://mlc.ai/wheels mlc-ai-nightly-cpu
 
       .. note::
         Make sure you also install vulkan loader and clang to avoid vulkan

python/mlc_llm/bench/__main__.py

@@ -187,7 +187,6 @@ def _main():
     parser.add_argument(
         "--dataset-path",
         type=str,
-        required=True,
         help="The dataset file path.",
     )
     parser.add_argument(
@@ -248,6 +247,14 @@ def _main():
         "When specified, the benchmark sends these many new requests each second. "
         'If it is "inf", all requests will be sent together at once.',
     )
+    parser.add_argument(
+        "--replay-timestamp-scale",
+        type=float,
+        help="The timestamp scale when replaying the timestamps in a dataset. "
+        'The dataset replay mode is enabled when neither "--num-concurrent-requests" and '
+        '"--request-rate" is specified. '
+        "The scale is 1 by default in the replay mode.",
+    )
     parser.add_argument(
         "--input-len",
         type=int,