[LLM] Add qwen1.5-7B in iGPU perf (#10127)

* Add qwen1.5 test config yaml with transformers 4.37.0

* Update for yaml file

.github/workflows/llm_performance_tests.yml

@@ -426,6 +426,30 @@ jobs:
 
           call conda deactivate
 
+      - name: Prepare igpu perf test for Qwen1.5 (32-32)
+        shell: bash
+        run: |
+          sed -i 's/{today}_test2/{today}_test3/g' python/llm/dev/benchmark/all-in-one/run.py
+          sed -i "s/path to your local model hub/$MODEL_HUB_DIR/g" python/llm/test/benchmark/igpu-perf/32-32_437.yaml
+
+      - name: Test on igpu for Qwen1.5 (32-32)
+        shell: cmd
+        run: |
+          call conda activate igpu-perf
+          pip install transformers==4.37.0
+
+          call "C:\Program Files (x86)\Intel\oneAPI\setvars.bat"
+          set SYCL_CACHE_PERSISTENT=1
+          set BIGDL_LLM_XMX_DISABLED=1
+
+          cd python\llm\dev\benchmark\all-in-one
+          move ..\..\..\test\benchmark\igpu-perf\32-32_437.yaml config.yaml
+          set PYTHONIOENCODING=utf-8
+          python run.py >> %CSV_SAVE_PATH%\32-32\log\%LOG_FILE% 2>&1
+          if %ERRORLEVEL% neq 0 (exit /b 1)
+
+          call conda deactivate
+
       - name: Concat csv and generate html (32-32)
         shell: cmd
         run: |
@@ -449,7 +473,7 @@ jobs:
         shell: bash
         run: |
           sed -i 's/32-32/1024-128/g' python/llm/dev/benchmark/all-in-one/run.py
-          sed -i 's/{today}_test2/{today}_test1/g' python/llm/dev/benchmark/all-in-one/run.py
+          sed -i 's/{today}_test3/{today}_test1/g' python/llm/dev/benchmark/all-in-one/run.py
           sed -i "s/path to your local model hub/$MODEL_HUB_DIR/g" python/llm/test/benchmark/igpu-perf/1024-128.yaml
 
       - name: Test on igpu (1024-128)
@@ -496,6 +520,30 @@ jobs:
 
           call conda deactivate
 
+      - name: Prepare igpu perf test for Qwen 1.5 (1024-128)
+        shell: bash
+        run: |
+          sed -i 's/{today}_test2/{today}_test3/g' python/llm/dev/benchmark/all-in-one/run.py
+          sed -i "s/path to your local model hub/$MODEL_HUB_DIR/g" python/llm/test/benchmark/igpu-perf/1024-128_437.yaml
+
+      - name: Test on igpu for Qwen 1.5 (1024-128)
+        shell: cmd
+        run: |
+          call conda activate igpu-perf
+          pip install transformers==4.37.0
+
+          call "C:\Program Files (x86)\Intel\oneAPI\setvars.bat"
+          set SYCL_CACHE_PERSISTENT=1
+          set BIGDL_LLM_XMX_DISABLED=1
+
+          cd python\llm\dev\benchmark\all-in-one
+          move ..\..\..\test\benchmark\igpu-perf\1024-128_437.yaml config.yaml
+          set PYTHONIOENCODING=utf-8
+          python run.py >> %CSV_SAVE_PATH%\1024-128\log\%LOG_FILE% 2>&1
+          if %ERRORLEVEL% neq 0 (exit /b 1)
+
+          call conda deactivate
+
       - name: Concat csv and generate html (1024-128)
         shell: cmd
         run: |
@@ -518,7 +566,7 @@ jobs:
         shell: bash
         run: |
           sed -i 's/1024-128/2048-256/g' python/llm/dev/benchmark/all-in-one/run.py
-          sed -i 's/{today}_test2/{today}_test1/g' python/llm/dev/benchmark/all-in-one/run.py
+          sed -i 's/{today}_test3/{today}_test1/g' python/llm/dev/benchmark/all-in-one/run.py
           sed -i "s/path to your local model hub/$MODEL_HUB_DIR/g" python/llm/test/benchmark/igpu-perf/2048-256.yaml
 
       - name: Test on igpu (2048-256)
@@ -565,6 +613,30 @@ jobs:
 
           call conda deactivate
 
+      - name: Prepare igpu perf test for Qwen 1.5 (2048-256)
+        shell: bash
+        run: |
+          sed -i 's/{today}_test2/{today}_test3/g' python/llm/dev/benchmark/all-in-one/run.py
+          sed -i "s/path to your local model hub/$MODEL_HUB_DIR/g" python/llm/test/benchmark/igpu-perf/2048-256_437.yaml
+
+      - name: Test on igpu for Qwen 1.5 (2048-256)
+        shell: cmd
+        run: |
+          call conda activate igpu-perf
+          pip install transformers==4.37.0
+
+          call "C:\Program Files (x86)\Intel\oneAPI\setvars.bat"
+          set SYCL_CACHE_PERSISTENT=1
+          set BIGDL_LLM_XMX_DISABLED=1
+
+          cd python\llm\dev\benchmark\all-in-one
+          move ..\..\..\test\benchmark\igpu-perf\2048-256_437.yaml config.yaml
+          set PYTHONIOENCODING=utf-8
+          python run.py >> %CSV_SAVE_PATH%\2048-256\log\%LOG_FILE% 2>&1
+          if %ERRORLEVEL% neq 0 (exit /b 1)
+
+          call conda deactivate
+
       - name: Concat csv and generate html (2048-256)
         shell: cmd
         run: |
@@ -588,7 +660,7 @@ jobs:
         run: |
           # hide time info
           sed -i 's/2048-256/32-512/g' python/llm/dev/benchmark/all-in-one/run.py
-          sed -i 's/{today}_test2/{today}_test1/g' python/llm/dev/benchmark/all-in-one/run.py
+          sed -i 's/{today}_test3/{today}_test1/g' python/llm/dev/benchmark/all-in-one/run.py
           sed -i "s/path to your local model hub/$MODEL_HUB_DIR/g" python/llm/test/benchmark/igpu-perf/32-512.yaml
 
       - name: Test on igpu (32-512)
@@ -635,6 +707,30 @@ jobs:
 
           call conda deactivate
 
+      - name: Prepare igpu perf test for Qwen 1.5 (32-512)
+        shell: bash
+        run: |
+          sed -i 's/{today}_test2/{today}_test3/g' python/llm/dev/benchmark/all-in-one/run.py
+          sed -i "s/path to your local model hub/$MODEL_HUB_DIR/g" python/llm/test/benchmark/igpu-perf/32-512_437.yaml
+
+      - name: Test on igpu for Qwen 1.5 (32-512)
+        shell: cmd
+        run: |
+          call conda activate igpu-perf
+          pip install transformers==4.37.0
+
+          call "C:\Program Files (x86)\Intel\oneAPI\setvars.bat"
+          set SYCL_CACHE_PERSISTENT=1
+          set BIGDL_LLM_XMX_DISABLED=1
+
+          cd python\llm\dev\benchmark\all-in-one
+          move ..\..\..\test\benchmark\igpu-perf\32-512_437.yaml config.yaml
+          set PYTHONIOENCODING=utf-8
+          python run.py >> %CSV_SAVE_PATH%\32-512\log\%LOG_FILE% 2>&1
+          if %ERRORLEVEL% neq 0 (exit /b 1)
+
+          call conda deactivate
+
       - name: Concat csv and generate html (32-512)
         shell: cmd
         run: |

python/llm/test/benchmark/concat_csv.py

@@ -36,7 +36,7 @@ def main():
 
     merged_df = pd.concat([pd.read_csv(file, index_col=0) for file in csv_files], ignore_index=True)
     merged_df.reset_index(drop=True, inplace=True)
-    merged_csv = csv_files[0].replace("_test1", "").replace("_test2", "")
+    merged_csv = csv_files[0].replace("_test1", "").replace("_test2", "").replace("_test3", "")
     merged_df.to_csv(merged_csv)
 
 if __name__ == "__main__":

python/llm/test/benchmark/igpu-perf/1024-128_437.yaml

@@ -0,0 +1,13 @@
+repo_id:
+  - 'Qwen/Qwen1.5-7B-Chat'
+local_model_hub: 'path to your local model hub'
+warm_up: 1
+num_trials: 3
+num_beams: 1 # default to greedy search
+low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
+batch_size: 1 # default to 1
+in_out_pairs:
+  - '1024-128'
+test_api:
+  - "transformer_int4_gpu_win" # on Intel GPU for Windows (catch GPU peak memory)
+cpu_embedding: True # whether put embedding to CPU (only avaiable now for gpu win related test_api)

python/llm/test/benchmark/igpu-perf/2048-256_437.yaml

@@ -0,0 +1,13 @@
+repo_id:
+  - 'Qwen/Qwen1.5-7B-Chat'
+local_model_hub: 'path to your local model hub'
+warm_up: 1
+num_trials: 3
+num_beams: 1 # default to greedy search
+low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
+batch_size: 1 # default to 1
+in_out_pairs:
+  - '2048-256'
+test_api:
+  - "transformer_int4_gpu_win" # on Intel GPU for Windows (catch GPU peak memory)
+cpu_embedding: True # whether put embedding to CPU (only avaiable now for gpu win related test_api)

python/llm/test/benchmark/igpu-perf/32-32_437.yaml

@@ -0,0 +1,13 @@
+repo_id:
+  - 'Qwen/Qwen1.5-7B-Chat'
+local_model_hub: 'path to your local model hub'
+warm_up: 3
+num_trials: 5
+num_beams: 1 # default to greedy search
+low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
+batch_size: 1 # default to 1
+in_out_pairs:
+  - '32-32'
+test_api:
+  - "transformer_int4_gpu_win" # on Intel GPU for Windows (catch GPU peak memory)
+cpu_embedding: True # whether put embedding to CPU (only avaiable now for gpu win related test_api)

python/llm/test/benchmark/igpu-perf/32-512_437.yaml

@@ -0,0 +1,13 @@
+repo_id:
+  - 'Qwen/Qwen1.5-7B-Chat'
+local_model_hub: 'path to your local model hub'
+warm_up: 1
+num_trials: 3
+num_beams: 1 # default to greedy search
+low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
+batch_size: 1 # default to 1
+in_out_pairs:
+  - '32-512'
+test_api:
+  - "transformer_int4_gpu_win" # on Intel GPU for Windows (catch GPU peak memory)
+cpu_embedding: True # whether put embedding to CPU (only avaiable now for gpu win related test_api)