remove incbench

README.md

@@ -167,8 +167,7 @@ model = load(
   </thead>
   <tbody>
       <tr>
-          <td colspan="4" align="center"><a href="./docs/source/3x/autotune.md">Auto Tune</a></td>
-          <td colspan="4" align="center"><a href="./docs/source/3x/benchmark.md">Benchmark</a></td>
+          <td colspan="8" align="center"><a href="./docs/source/3x/autotune.md">Auto Tune</a></td>
       </tr>
   </tbody>
 </table>

docs/source/2x_user_guide.md

@@ -37,8 +37,7 @@ In 2.X API, it's very important to create the `DataLoader` and `Metrics` for you
 <td colspan="2" align="center"><a href="distillation.md">Distillation</a></td>
 </tr>
 <tr>
-<td colspan="2" align="center"><a href="orchestration.md">Orchestration</a></td>
-<td colspan="2" align="center"><a href="benchmark.md">Benchmarking</a></td>
+<td colspan="3" align="center"><a href="orchestration.md">Orchestration</a></td>
 <td colspan="3" align="center"><a href="distributed.md">Distributed Compression</a></td>
 <td colspan="3" align="center"><a href="export.md">Model Export</a></td>
 </tr>

docs/source/get_started.md

@@ -78,8 +78,7 @@ It also provides some common modules across different frameworks. For example, A
   </thead>
   <tbody>
       <tr>
-          <td colspan="4" align="center"><a href="3x/autotune.md">Auto Tune</a></td>
-          <td colspan="4" align="center"><a href="3x/benchmark.md">Benchmark</a></td>
+          <td colspan="8" align="center"><a href="3x/autotune.md">Auto Tune</a></td>
       </tr>
   </tbody>
 </table>

examples/pytorch/cv/mixed_precision/run_benchmark.sh

@@ -70,7 +70,7 @@ function run_benchmark {
                 ${mode_cmd} \
                 ${extra_cmd}
     elif [[ ${mode} == "performance" ]]; then
-        incbench --num_c 4 main.py \
+        python main.py \
                 --pretrained \
                 --tuned_checkpoint ${tuned_checkpoint} \
                 -b ${batch_size} \

examples/pytorch/cv/static_quant/run_benchmark.sh

@@ -85,8 +85,7 @@ function run_benchmark {
                 ${extra_cmd} \
                 ${mode_cmd}
     elif [[ ${mode} == "performance" ]]; then
-        incbench --num_cores_per_instance 4 \
-                main.py \
+        python main.py \
                 --pretrained \
                 -a resnet18 \
                 -b ${batch_size} \

examples/pytorch/nlp/huggingface_models/language-modeling/quantization/smooth_quant/run_benchmark.sh

@@ -93,7 +93,7 @@ function run_benchmark {
             --batch_size ${batch_size} \
             ${extra_cmd} ${mode_cmd}
     elif [[ ${mode} == "performance" ]]; then
-        incbench --num_cores_per_instance 4 run_clm_no_trainer.py \
+        python run_clm_no_trainer.py \
             --model ${model_name_or_path} \
             --approach ${approach} \
             --batch_size ${batch_size} \

examples/pytorch/nlp/huggingface_models/language-modeling/quantization/static_quant/pt2e/run_benchmark.sh

@@ -85,7 +85,7 @@ function run_benchmark {
             --batch_size ${batch_size} \
             ${extra_cmd} ${mode_cmd}
     elif [[ ${mode} == "performance" ]]; then
-        incbench --num_cores_per_instance 4 run_clm_no_trainer.py \
+        python run_clm_no_trainer.py \
             --model ${model_name_or_path} \
             --batch_size ${batch_size} \
             --output_dir ${tuned_checkpoint} \

examples/pytorch/nlp/huggingface_models/language-modeling/quantization/weight_only/run_benchmark.sh

@@ -14,7 +14,6 @@ function init_params {
   batch_size=16
   tuned_checkpoint=saved_results
   task=lambada_openai
-  incbench_cmd="incbench --num_cores_per_instance 4"
   echo ${max_eval_samples}
   for var in "$@"
   do
@@ -107,7 +106,6 @@ function run_benchmark {
     elif [ "${topology}" = "opt_125m_woq_autoround_int4_hpu" ]; then
         model_name_or_path="facebook/opt-125m"
         extra_cmd=$extra_cmd" --woq_algo AutoRound"
-        incbench_cmd="incbench --num_instances 1"
     elif [ "${topology}" = "opt_125m_woq_autotune_int4" ]; then
         model_name_or_path="facebook/opt-125m"
     elif [ "${topology}" = "llama3_1_8b_gptq_int4_hf" ]; then
@@ -134,7 +132,7 @@ function run_benchmark {
             --batch_size ${batch_size} \
             ${extra_cmd} ${mode_cmd}
     elif [[ ${mode} == "performance" ]]; then
-        ${incbench_cmd} run_clm_no_trainer.py \
+        python run_clm_no_trainer.py \
             --model ${model_name_or_path} \
             --batch_size ${batch_size} \
             --output_dir ${tuned_checkpoint} \

examples/pytorch/recommendation/dlrm/static_quant/ipex/run_benchmark.sh

@@ -80,7 +80,7 @@ function run_tuning {
         --save-model ${tuned_checkpoint} --test-freq=2048 --print-auc $ARGS \
         --load-model=${input_model}  --accuracy_only
     elif [[ ${mode} == "performance" ]]; then
-        incbench --num_cores_per_instance 4 $MODEL_SCRIPT \
+        python $MODEL_SCRIPT \
         --raw-data-file=${dataset_location}/day --processed-data-file=${dataset_location}/terabyte_processed.npz \
         --data-set=terabyte --benchmark \
         --memory-map --mlperf-bin-loader --round-targets=True --learning-rate=1.0 \

examples/tensorflow/object_detection/yolo_v5/quantization/ptq/run_benchmark.sh

@@ -46,7 +46,7 @@ function run_benchmark {
             --batch_size ${batch_size} \
             --benchmark
     elif [[ ${mode} == "performance" ]]; then
-        incbench --num_c 4 main.py \
+        python main.py \
             --input_model ${input_model} \
             --dataset_location ${dataset_location} \
             --mode ${mode} \