Merge remote-tracking branch 'origin/main' into rope

.github/scripts/eval_base_config.py

@@ -89,6 +89,17 @@
         models as lmdeploy_qwen1_5_7b  # noqa: F401, E501
     from opencompass.configs.models.qwen.lmdeploy_qwen2_7b import \
         models as lmdeploy_qwen2_7b  # noqa: F401, E501
+    # Summary Groups
+    from opencompass.configs.summarizers.groups.cmmlu import \
+        cmmlu_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.GaokaoBench import \
+        GaokaoBench_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.mathbench_v1_2024 import \
+        mathbench_2024_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.mmlu import \
+        mmlu_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.mmlu_pro import \
+        mmlu_pro_summary_groups  # noqa: F401, E501
 
     # read models
 race_datasets = [race_datasets[1]]

.github/scripts/eval_chat_config.py

@@ -98,6 +98,27 @@
         models as lmdeploy_qwen2_7b_instruct  # noqa: F401, E501
     from opencompass.configs.models.qwen.lmdeploy_qwen_7b_chat import \
         models as lmdeploy_qwen_7b_chat  # noqa: F401, E501
+    # Summary Groups
+    from opencompass.configs.summarizers.groups.bbh import \
+        bbh_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.cmmlu import \
+        cmmlu_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.ds1000 import \
+        ds1000_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.GaokaoBench import \
+        GaokaoBench_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.humanevalx import \
+        humanevalx_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.mathbench_v1_2024 import \
+        mathbench_2024_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.mmlu import \
+        mmlu_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.mmlu_pro import \
+        mmlu_pro_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.scicode import \
+        scicode_summary_groups  # noqa: F401, E501
+    from opencompass.configs.summarizers.groups.teval import \
+        teval_summary_groups  # noqa: F401, E501
 
 llama2_meta_template = dict(round=[
     dict(role='HUMAN', begin='[INST] ', end=' [/INST]'),

.github/workflows/daily_ete_test.yml

@@ -17,10 +17,15 @@ on:
         required: true
         description: 'Set backend testcase filter: turbomind or pytorch or turbomind, pytorch. Default is "["turbomind", "pytorch"]"'
         type: string
-        default: '["turbomind", "pytorch", "turbomind_vl"]'
+        default: "['turbomind', 'pytorch']"
       model:
         required: true
-        description: 'Set testcase module filter: chat, restful, pipeline, quantization. Default contains all models'
+        description: 'Set testcase module filter: llm, vllm. Default contains all models'
+        type: string
+        default: "['llm','mllm']"
+      function:
+        required: true
+        description: 'Set testcase function filter: chat, restful, pipeline. Default contains all functions'
         type: string
         default: '["pipeline", "restful", "chat"]'
       offline_mode:
@@ -206,14 +211,20 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        backend: ${{ fromJSON(inputs.backend || '["turbomind", "pytorch", "turbomind_vl"]')}}
-        model: ${{ fromJSON(inputs.model || '["pipeline", "restful", "chat"]')}}
+        backend: ${{ fromJSON(inputs.backend || '["turbomind", "pytorch"]')}}
+        model: ${{ fromJSON(inputs.model || '["llm", "mllm"]')}}
+        function: ${{ fromJSON(inputs.function || '["pipeline","restful","chat"]')}}
         exclude:
-          - backend: turbomind_vl
-            model: chat
+          - backend: turbomind
+            model: mllm
+            function: chat
+          - backend: pytorch
+            model: mllm
+            function: chat
         include:
           - backend: turbomind
-            model: local_case
+            model: llm
+            function: local_case
     env:
       PYTHONPATH: /nvme/qa_test_models/offline_pkg/LLaVA
       MODELSCOPE_CACHE: /root/modelscope_hub
@@ -261,46 +272,46 @@ jobs:
           ln -s ${{env.REPORT_DIR}}/.pytest_cache autotest
       - name: Test lmdeploy - chat workspace
         continue-on-error: true
-        if: matrix.backend == 'turbomind' && matrix.model == 'chat'
+        if: matrix.backend == 'turbomind' && matrix.model == 'llm' && matrix.function == 'chat'
         run: |
           pytest autotest/tools/chat/test_command_chat_workspace.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
           pytest autotest/tools/chat/test_command_chat_workspace.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - chat
         continue-on-error: true
-        if: (matrix.backend == 'pytorch' || matrix.backend == 'turbomind') && matrix.model == 'chat'
+        if: (matrix.backend == 'pytorch' || matrix.backend == 'turbomind') && matrix.model == 'llm' && matrix.function == 'chat'
         run: |
           pytest autotest/tools/chat/test_command_chat_hf_${{matrix.backend}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
           pytest autotest/tools/chat/test_command_chat_hf_${{matrix.backend}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - pipeline
         continue-on-error: true
-        if: matrix.model == 'pipeline'
+        if: matrix.function == 'pipeline'
         run: |
-          pytest autotest/tools/pipeline/test_pipeline_chat_${{matrix.backend}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
+          pytest autotest/tools/pipeline/test_pipeline_chat_${{matrix.backend}}_${{matrix.model}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
-          pytest autotest/tools/pipeline/test_pipeline_chat_${{matrix.backend}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
+          pytest autotest/tools/pipeline/test_pipeline_chat_${{matrix.backend}}_${{matrix.model}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - restful
         continue-on-error: true
-        if: matrix.model == 'restful'
+        if: matrix.function == 'restful'
         run: |
-          pytest autotest/tools/restful/test_restful_chat_hf_${{matrix.backend}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
+          pytest autotest/tools/restful/test_restful_chat_hf_${{matrix.backend}}_${{matrix.model}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
-          pytest autotest/tools/restful/test_restful_chat_hf_${{matrix.backend}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
+          pytest autotest/tools/restful/test_restful_chat_hf_${{matrix.backend}}_${{matrix.model}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - restful workspace
         continue-on-error: true
-        if: matrix.backend == 'turbomind' && matrix.model == 'restful'
+        if: matrix.backend == 'turbomind' && matrix.model == 'llm' && matrix.function == 'restful'
         run: |
           pytest autotest/tools/restful/test_restful_chat_workspace.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
           pytest autotest/tools/restful/test_restful_chat_workspace.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - local testcase
-        if: matrix.backend == 'turbomind' && matrix.model == 'local_case'
+        if: matrix.backend == 'turbomind' && matrix.model == 'llm' && matrix.function == 'local_case'
         run: |
           pytest /local_case/issue_regression --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}}|| true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
@@ -321,7 +332,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        backend: ['turbomind', 'pytorch']
+        backend: ${{ fromJSON(inputs.backend || '["turbomind", "pytorch"]')}}
     timeout-minutes: 60
     container:
       image: openmmlab/lmdeploy:latest-cu11

.github/workflows/daily_ete_test_v100.yml

@@ -17,10 +17,15 @@ on:
         required: true
         description: 'Set backend testcase filter: turbomind or pytorch or turbomind, pytorch. Default is "["turbomind", "pytorch"]"'
         type: string
-        default: '["turbomind", "pytorch", "turbomind_vl"]'
+        default: "['turbomind', 'pytorch']"
       model:
         required: true
-        description: 'Set testcase module filter: chat, restful, pipeline, quantization. Default contains all models'
+        description: 'Set testcase module filter: llm, vllm. Default contains all models'
+        type: string
+        default: "['llm','mllm']"
+      function:
+        required: true
+        description: 'Set testcase function filter: chat, restful, pipeline. Default contains all functions'
         type: string
         default: '["pipeline", "restful", "chat"]'
       offline_mode:
@@ -201,14 +206,20 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        backend: ${{ fromJSON(inputs.backend || '["turbomind", "pytorch", "turbomind_vl"]')}}
-        model: ${{ fromJSON(inputs.model || '["pipeline", "restful", "chat"]')}}
+        backend: ${{ fromJSON(inputs.backend || '["turbomind", "pytorch"]')}}
+        model: ${{ fromJSON(inputs.model || '["llm", "mllm"]')}}
+        function: ${{ fromJSON(inputs.function || '["pipeline","restful","chat"]')}}
         exclude:
-          - backend: turbomind_vl
-            model: chat
+          - backend: turbomind
+            model: mllm
+            function: chat
+          - backend: pytorch
+            model: mllm
+            function: chat
         include:
           - backend: turbomind
-            model: local_case
+            model: llm
+            function: local_case
     env:
       PYTHONPATH: /nvme/qa_test_models/offline_pkg/LLaVA
       MODELSCOPE_CACHE: /root/modelscope_hub
@@ -255,46 +266,46 @@ jobs:
           ln -s ${{env.REPORT_DIR}}/.pytest_cache autotest
       - name: Test lmdeploy - chat workspace
         continue-on-error: true
-        if: matrix.backend == 'turbomind' && matrix.model == 'chat'
+        if: matrix.backend == 'turbomind' && matrix.model == 'llm' && matrix.function == 'chat'
         run: |
           pytest autotest/tools/chat/test_command_chat_workspace.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
           pytest autotest/tools/chat/test_command_chat_workspace.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - chat
         continue-on-error: true
-        if: (matrix.backend == 'pytorch' || matrix.backend == 'turbomind') && matrix.model == 'chat'
+        if: (matrix.backend == 'pytorch' || matrix.backend == 'turbomind') && matrix.model == 'llm' && matrix.function == 'chat'
         run: |
           pytest autotest/tools/chat/test_command_chat_hf_${{matrix.backend}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
           pytest autotest/tools/chat/test_command_chat_hf_${{matrix.backend}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - pipeline
         continue-on-error: true
-        if: matrix.model == 'pipeline'
+        if: matrix.function == 'pipeline'
         run: |
-          pytest autotest/tools/pipeline/test_pipeline_chat_${{matrix.backend}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
+          pytest autotest/tools/pipeline/test_pipeline_chat_${{matrix.backend}}_${{matrix.model}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
-          pytest autotest/tools/pipeline/test_pipeline_chat_${{matrix.backend}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
+          pytest autotest/tools/pipeline/test_pipeline_chat_${{matrix.backend}}_${{matrix.model}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - restful
         continue-on-error: true
-        if: matrix.model == 'restful'
+        if: matrix.function == 'restful'
         run: |
-          pytest autotest/tools/restful/test_restful_chat_hf_${{matrix.backend}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
+          pytest autotest/tools/restful/test_restful_chat_hf_${{matrix.backend}}_${{matrix.model}}.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
-          pytest autotest/tools/restful/test_restful_chat_hf_${{matrix.backend}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
+          pytest autotest/tools/restful/test_restful_chat_hf_${{matrix.backend}}_${{matrix.model}}.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - restful workspace
         continue-on-error: true
-        if: matrix.backend == 'turbomind' && matrix.model == 'restful'
+        if: matrix.backend == 'turbomind' && matrix.model == 'llm' && matrix.function == 'restful'
         run: |
           pytest autotest/tools/restful/test_restful_chat_workspace.py -m 'gpu_num_1 and not pr_test' -n 8 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S') || true
           pytest autotest/tools/restful/test_restful_chat_workspace.py -m 'gpu_num_2 and not pr_test' -n 4 --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}} || true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
       - name: Test lmdeploy - local testcase
-        if: matrix.backend == 'turbomind' && matrix.model == 'local_case'
+        if: matrix.backend == 'turbomind' && matrix.model == 'llm' && matrix.function == 'local_case'
         run: |
           pytest /local_case/issue_regression --alluredir=${{env.REPORT_DIR}} ${{env.COV_PARAM}}|| true
           mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
@@ -315,7 +326,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        backend: ['turbomind', 'pytorch']
+        backend: ${{ fromJSON(inputs.backend || '["turbomind", "pytorch"]')}}
     timeout-minutes: 120
     container:
       image: openmmlab/lmdeploy:latest-cu12

README.md

@@ -157,6 +157,7 @@ For detailed inference benchmarks in more devices and more settings, please refe
   <li>InternVL-Chat (v1.1-v1.5)</li>
   <li>InternVL2 (1B-76B)</li>
   <li>Mono-InternVL (2B)</li>
+  <li>ChemVLM (8B-26B)</li>
   <li>MiniGeminiLlama (7B)</li>
   <li>CogVLM-Chat (17B)</li>
   <li>CogVLM2-Chat (19B)</li>
@@ -166,6 +167,7 @@ For detailed inference benchmarks in more devices and more settings, please refe
   <li>Phi-3.5-vision (4.2B)</li>
   <li>GLM-4V (9B)</li>
   <li>Llama3.2-vision (11B, 90B)</li>
+  <li>Molmo (7B-D,72B)</li>
 </ul>
 </td>
 </tr>

README_ja.md

@@ -152,6 +152,8 @@ LMDeploy TurboMindエンジンは卓越した推論能力を持ち、さまざ
   <li>DeepSeek-VL (7B)</li>
   <li>InternVL-Chat (v1.1-v1.5)</li>
   <li>InternVL2 (1B-76B)</li>
+  <li>Mono-InternVL (2B)</li>
+  <li>ChemVLM (8B-26B)</li>
   <li>MiniGeminiLlama (7B)</li>
   <li>CogVLM-Chat (17B)</li>
   <li>CogVLM2-Chat (19B)</li>
@@ -161,6 +163,7 @@ LMDeploy TurboMindエンジンは卓越した推論能力を持ち、さまざ
   <li>Phi-3.5-vision (4.2B)</li>
   <li>GLM-4V (9B)</li>
   <li>Llama3.2-vision (11B, 90B)</li>
+  <li>Molmo (7B-D,72B)</li>
 </ul>
 </td>
 </tr>

README_zh-CN.md

@@ -158,6 +158,7 @@ LMDeploy TurboMind 引擎拥有卓越的推理能力，在各种规模的模型
   <li>InternVL-Chat (v1.1-v1.5)</li>
   <li>InternVL2 (1B-76B)</li>
   <li>Mono-InternVL (2B)</li>
+  <li>ChemVLM (8B-26B)</li>
   <li>MiniGeminiLlama (7B)</li>
   <li>CogVLM-Chat (17B)</li>
   <li>CogVLM2-Chat (19B)</li>
@@ -167,6 +168,7 @@ LMDeploy TurboMind 引擎拥有卓越的推理能力，在各种规模的模型
   <li>Phi-3.5-vision (4.2B)</li>
   <li>GLM-4V (9B)</li>
   <li>Llama3.2-vision (11B, 90B)</li>
+  <li>Molmo (7B-D,72B)</li>
 </ul>
 </td>
 </tr>

autotest/config-v100.yaml

@@ -22,6 +22,7 @@ tp_config:
 
 turbomind_chat_model:
     - meta-llama/Llama-3.2-1B-Instruct
+    - meta-llama/Llama-3.2-3B-Instruct
     - meta-llama/Meta-Llama-3-1-8B-Instruct
     - meta-llama/Meta-Llama-3-1-8B-Instruct-AWQ
     - meta-llama/Meta-Llama-3-8B-Instruct
@@ -48,6 +49,7 @@ pytorch_chat_model:
     - meta-llama/Meta-Llama-3-1-8B-Instruct
     - internlm/internlm2_5-7b-chat
     - internlm/internlm2_5-20b-chat
+    - OpenGVLab/InternVL2-1B
     - OpenGVLab/InternVL2-2B
     - OpenGVLab/InternVL2-4B
     - OpenGVLab/InternVL2-8B
@@ -71,16 +73,25 @@ pytorch_base_model:
     - internlm/internlm2_5-7b
     - internlm/internlm2_5-20b
 
-vl_model:
+turbomind_vl_model:
     - OpenGVLab/InternVL2-1B
     - OpenGVLab/InternVL2-2B
-    - OpenGVLab/InternVL2-4B
     - OpenGVLab/InternVL2-8B
     - OpenGVLab/InternVL2-26B
     - Qwen/Qwen2-VL-2B-Instruct
     - Qwen/Qwen2-VL-7B-Instruct
     - internlm/internlm-xcomposer2d5-7b
     - THUDM/glm-4v-9b
+
+pytorch_vl_model:
+    - OpenGVLab/InternVL2-1B
+    - OpenGVLab/InternVL2-4B
+    - OpenGVLab/InternVL2-8B
+    - OpenGVLab/InternVL2-26B
+    - OpenGVLab/Mono-InternVL-2B
+    - Qwen/Qwen2-VL-2B-Instruct
+    - Qwen/Qwen2-VL-7B-Instruct
+    - THUDM/glm-4v-9b
     - microsoft/Phi-3.5-vision-instruct
 
 turbomind_quatization:
@@ -107,10 +118,13 @@ pytorch_quatization:
         - internlm/internlm2_5-7b-chat
         - internlm/internlm2_5-7b
     no_kvint4:
+        - OpenGVLab/InternVL2-1B
         - OpenGVLab/InternVL2-4B
         - deepseek-ai/DeepSeek-V2-Lite-Chat
         - microsoft/Phi-3-mini-4k-instruct
         - microsoft/Phi-3-vision-128k-instruct
+        - microsoft/Phi-3.5-vision-instruct
+        - openbmb/MiniCPM-V-2_6
     no_kvint8:
         - deepseek-ai/DeepSeek-V2-Lite-Chat