xinference 0.15.3__py3-none-any.whl → 0.15.4__py3-none-any.whl

xinference/model/llm/llm_family.json

@@ -1111,7 +1111,8 @@
       "th"
     ],
     "model_ability": [
-      "chat"
+      "chat",
+      "tools"
     ],
     "model_description": "The Llama 3.1 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks..",
     "model_specs": [
@@ -1299,14 +1300,16 @@
         "model_id": "hugging-quants/Meta-Llama-3.1-405B-Instruct-AWQ-INT4"
       }
     ],
-    "chat_template": "{{- '<|begin_of_text|>' }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\n\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\n\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\n\" }}\n{{- \"Today Date: \" + date_string + \"\n\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\n\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\n\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\n\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\n\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\n\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\n\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\n\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' }}\n{%- endif %}\n",
+    "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
     "stop_token_ids": [
       128001,
+      128008,
       128009
     ],
     "stop": [
       "<|end_of_text|>",
-      "<|eot_id|>"
+      "<|eot_id|>",
+      "<|eom_id|>"
     ]
   },
   {
@@ -6906,18 +6909,15 @@
         "model_id":"Qwen/Qwen2-VL-72B-Instruct-GPTQ-{quantization}"
       }
     ],
-    "prompt_style":{
-      "style_name":"QWEN",
-      "system_prompt":"You are a helpful assistant",
-      "roles":[
-        "user",
-        "assistant"
-      ],
-      "stop": [
-        "<|im_end|>",
-        "<|endoftext|>"
-      ]
-    }
+    "chat_template": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}",
+    "stop_token_ids": [
+      151645,
+      151643
+    ],
+    "stop": [
+      "<|im_end|>",
+      "<|endoftext|>"
+    ]
   },
   {
     "version": 1,
@@ -7925,7 +7925,7 @@
         }
       }
     ],
-    "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+    "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
     "stop_token_ids": [
       151643,
       151644,

xinference/model/llm/llm_family_modelscope.json

@@ -246,7 +246,8 @@
       "th"
     ],
     "model_ability": [
-      "chat"
+      "chat",
+      "tools"
     ],
     "model_description": "The Llama 3.1 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks..",
     "model_specs": [
@@ -350,14 +351,16 @@
         "model_hub": "modelscope"
       }
     ],
-    "chat_template": "{{- '<|begin_of_text|>' }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\n\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\n\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\n\" }}\n{{- \"Today Date: \" + date_string + \"\n\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\n\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\n\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\n\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\n\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\n\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\n\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\n\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' }}\n{%- endif %}\n",
+    "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n    {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n    {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n    {#- Extract the first user message so we can plug it in here #}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n    {{- \"Do not use variables.\\n\\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\\n\\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n            {%- for arg_name, arg_val in tool_call.arguments | items %}\n                {{- arg_name + '=\"' + arg_val + '\"' }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- endif %}\n                {%- endfor %}\n            {{- \")\" }}\n        {%- else  %}\n            {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n            {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n            {{- '\"parameters\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- \"}\" }}\n        {%- endif %}\n        {%- if builtin_tools is defined %}\n            {#- This means we're in ipython mode #}\n            {{- \"<|eom_id|>\" }}\n        {%- else %}\n            {{- \"<|eot_id|>\" }}\n        {%- endif %}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
     "stop_token_ids": [
       128001,
+      128008,
       128009
     ],
     "stop": [
       "<|end_of_text|>",
-      "<|eot_id|>"
+      "<|eot_id|>",
+      "<|eom_id|>"
     ]
   },
   {
@@ -4624,14 +4627,15 @@
         "model_hub": "modelscope"
       }
     ],
-    "prompt_style": {
-      "style_name": "QWEN",
-      "system_prompt": "You are a helpful assistant",
-      "roles": [
-        "user",
-        "assistant"
-      ]
-    }
+    "chat_template": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}",
+    "stop_token_ids": [
+      151645,
+      151643
+    ],
+    "stop": [
+      "<|im_end|>",
+      "<|endoftext|>"
+    ]
   },
   {
     "version": 1,
@@ -5679,7 +5683,7 @@
         }
       }
     ],
-    "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+    "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
     "stop_token_ids": [
       151643,
       151644,

xinference/model/llm/transformers/cogvlm2.py

@@ -29,7 +29,7 @@ from ..utils import (
     parse_messages,
 )
 from .core import PytorchChatModel, PytorchGenerateConfig
-from .utils import get_max_src_len
+from .utils import cache_clean, get_max_src_len
 
 logger = logging.getLogger(__name__)
 
@@ -176,6 +176,7 @@ class CogVLM2Model(PytorchChatModel):
             query = content
         return query, image, history
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],

xinference/model/llm/transformers/cogvlm2_video.py

@@ -28,6 +28,7 @@ from ..utils import (
     parse_messages,
 )
 from .core import PytorchChatModel, PytorchGenerateConfig
+from .utils import cache_clean
 
 logger = logging.getLogger(__name__)
 
@@ -227,6 +228,7 @@ class CogVLM2VideoModel(PytorchChatModel):
 
         return query, image, video, history
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],

xinference/model/llm/transformers/core.py

@@ -40,7 +40,7 @@ from ....types import (
 from ...utils import select_device
 from ..core import LLM
 from ..llm_family import LLMFamilyV1, LLMSpecV1
-from ..utils import QWEN_TOOL_CALL_FAMILY, ChatModelMixin
+from ..utils import LLAMA3_TOOL_CALL_FAMILY, QWEN_TOOL_CALL_FAMILY, ChatModelMixin
 from .utils import get_context_length, get_max_src_len, pad_prefill_tokens
 
 logger = logging.getLogger(__name__)
@@ -733,7 +733,11 @@ class PytorchChatModel(PytorchModel, ChatModelMixin):
         tools = generate_config.pop("tools", []) if generate_config else None
         model_family = self.model_family.model_family or self.model_family.model_name
         full_context_kwargs = {}
-        if tools and model_family in QWEN_TOOL_CALL_FAMILY:
+        if (
+            tools
+            and model_family in QWEN_TOOL_CALL_FAMILY
+            or model_family in LLAMA3_TOOL_CALL_FAMILY
+        ):
             full_context_kwargs["tools"] = tools
         assert self.model_family.chat_template is not None
         full_prompt = self.get_full_context(

xinference/model/llm/transformers/deepseek_vl.py

@@ -28,6 +28,7 @@ from ....types import ChatCompletion, ChatCompletionChunk, CompletionChunk
 from ..llm_family import LLMFamilyV1, LLMSpecV1
 from ..utils import generate_chat_completion, generate_completion_chunk
 from .core import PytorchChatModel, PytorchGenerateConfig
+from .utils import cache_clean
 
 logger = logging.getLogger(__name__)
 
@@ -137,6 +138,7 @@ class DeepSeekVLChatModel(PytorchChatModel):
             return "".join(new_content), images
         return content, []
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],

xinference/model/llm/transformers/glm4v.py

@@ -26,7 +26,7 @@ from ...utils import select_device
 from ..llm_family import LLMFamilyV1, LLMSpecV1
 from ..utils import _decode_image, generate_chat_completion, generate_completion_chunk
 from .core import PytorchChatModel, PytorchGenerateConfig
-from .utils import get_max_src_len
+from .utils import cache_clean, get_max_src_len
 
 logger = logging.getLogger(__name__)
 
@@ -129,6 +129,7 @@ class Glm4VModel(PytorchChatModel):
                     res.append({"role": role, "content": text})
         return res
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],

xinference/model/llm/transformers/intern_vl.py

@@ -27,6 +27,7 @@ from ..utils import (
     parse_messages,
 )
 from .core import PytorchChatModel, PytorchGenerateConfig
+from .utils import cache_clean
 
 logger = logging.getLogger(__name__)
 
@@ -326,6 +327,7 @@ class InternVLChatModel(PytorchChatModel):
             use_fast=False,
         )
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],

xinference/model/llm/transformers/minicpmv25.py

@@ -29,6 +29,7 @@ from ..utils import (
     parse_messages,
 )
 from .core import PytorchChatModel, PytorchGenerateConfig
+from .utils import cache_clean
 
 logger = logging.getLogger(__name__)
 
@@ -119,6 +120,7 @@ class MiniCPMV25Model(PytorchChatModel):
                 raise RuntimeError("Only one image per message is supported")
         return content, []
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],

xinference/model/llm/transformers/minicpmv26.py

@@ -30,6 +30,7 @@ from ..utils import (
     parse_messages,
 )
 from .core import PytorchChatModel, PytorchGenerateConfig
+from .utils import cache_clean
 
 logger = logging.getLogger(__name__)
 
@@ -198,6 +199,7 @@ class MiniCPMV26Model(PytorchChatModel):
         msgs.append({"role": "user", "content": images_chat + [content]})
         return msgs, video_existed
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],

xinference/model/llm/transformers/omnilmm.py

@@ -24,6 +24,7 @@ from ...utils import select_device
 from ..llm_family import LLMFamilyV1, LLMSpecV1
 from ..utils import generate_chat_completion, parse_messages
 from .core import PytorchChatModel, PytorchGenerateConfig
+from .utils import cache_clean
 
 logger = logging.getLogger(__name__)
 
@@ -87,6 +88,7 @@ class OmniLMMModel(PytorchChatModel):
             return images, other_content
         return [], [{"type": "text", "text": content}]
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],

xinference/model/llm/transformers/qwen2_audio.py

@@ -14,16 +14,22 @@
 import logging
 import uuid
 from io import BytesIO
-from typing import Dict, Iterator, List, Optional, Union
+from typing import Iterator, List, Optional, Union
 from urllib.request import urlopen
 
 import numpy as np
 
 from ....model.utils import select_device
-from ....types import ChatCompletion, ChatCompletionChunk, CompletionChunk
+from ....types import (
+    ChatCompletion,
+    ChatCompletionChunk,
+    ChatCompletionMessage,
+    CompletionChunk,
+)
 from ..llm_family import LLMFamilyV1, LLMSpecV1
 from ..utils import generate_chat_completion, generate_completion_chunk
 from .core import PytorchChatModel, PytorchGenerateConfig
+from .utils import cache_clean
 
 logger = logging.getLogger(__name__)
 
@@ -68,7 +74,7 @@ class Qwen2AudioChatModel(PytorchChatModel):
 
     def _transform_messages(
         self,
-        messages: List[Dict],
+        messages: List[ChatCompletionMessage],
     ):
         import librosa
 
@@ -89,9 +95,10 @@ class Qwen2AudioChatModel(PytorchChatModel):
 
         return text, audios
 
+    @cache_clean
     def chat(
         self,
-        messages: List[Dict],
+        messages: List[ChatCompletionMessage],
         generate_config: Optional[PytorchGenerateConfig] = None,
     ) -> Union[ChatCompletion, Iterator[ChatCompletionChunk]]:
         text, audios = self._transform_messages(messages)

xinference/model/llm/transformers/qwen2_vl.py

@@ -27,6 +27,7 @@ from ....types import (
 from ..llm_family import LLMFamilyV1, LLMSpecV1
 from ..utils import generate_chat_completion, generate_completion_chunk
 from .core import PytorchChatModel, PytorchGenerateConfig
+from .utils import cache_clean
 
 logger = logging.getLogger(__name__)
 
@@ -75,34 +76,7 @@ class Qwen2VLChatModel(PytorchChatModel):
                 self.model_path, device_map=device, trust_remote_code=True
             ).eval()
 
-    def _transform_messages(
-        self,
-        messages: List[ChatCompletionMessage],
-    ):
-        transformed_messages = []
-        for msg in messages:
-            new_content = []
-            role = msg["role"]
-            content = msg["content"]
-            if isinstance(content, str):
-                new_content.append({"type": "text", "text": content})
-            elif isinstance(content, List):
-                for item in content:  # type: ignore
-                    if "text" in item:
-                        new_content.append({"type": "text", "text": item["text"]})
-                    elif "image_url" in item:
-                        new_content.append(
-                            {"type": "image", "image": item["image_url"]["url"]}
-                        )
-                    elif "video_url" in item:
-                        new_content.append(
-                            {"type": "video", "video": item["video_url"]["url"]}
-                        )
-            new_message = {"role": role, "content": new_content}
-            transformed_messages.append(new_message)
-
-        return transformed_messages
-
+    @cache_clean
     def chat(
         self,
         messages: List[ChatCompletionMessage],  # type: ignore

xinference/model/llm/transformers/qwen_vl.py

@@ -28,7 +28,7 @@ from ....types import ChatCompletion, ChatCompletionChunk, CompletionChunk
 from ..llm_family import LLMFamilyV1, LLMSpecV1
 from ..utils import generate_chat_completion, generate_completion_chunk
 from .core import PytorchChatModel, PytorchGenerateConfig
-from .utils import pad_prefill_tokens
+from .utils import cache_clean, pad_prefill_tokens
 
 logger = logging.getLogger(__name__)
 
@@ -137,6 +137,7 @@ class QwenVLChatModel(PytorchChatModel):
         prompt = self._message_content_to_qwen(messages[-1]["content"])
         return prompt, qwen_history
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],

xinference/model/llm/transformers/utils.py

@@ -11,7 +11,8 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-
+import asyncio
+import functools
 import gc
 import logging
 import os
@@ -126,7 +127,8 @@ def generate_stream(
     echo = bool(generate_config.get("echo", False))
     stop_str = generate_config.get("stop", None)
     stop_token_ids = generate_config.get("stop_token_ids", None) or []
-    stop_token_ids.append(tokenizer.eos_token_id)
+    if tokenizer.eos_token_id not in stop_token_ids:
+        stop_token_ids.append(tokenizer.eos_token_id)
     chunk_id = str(uuid.uuid4())
 
     logits_processor = prepare_logits_processor(
@@ -776,3 +778,34 @@ def batch_inference_one_step(
         for r in req_list:
             r.stopped = True
             r.error_msg = str(e)
+
+
+def cache_clean(fn):
+    @functools.wraps(fn)
+    async def _async_wrapper(self, *args, **kwargs):
+        import gc
+
+        from ....device_utils import empty_cache
+
+        result = await fn(self, *args, **kwargs)
+
+        gc.collect()
+        empty_cache()
+        return result
+
+    @functools.wraps(fn)
+    def _wrapper(self, *args, **kwargs):
+        import gc
+
+        from ....device_utils import empty_cache
+
+        result = fn(self, *args, **kwargs)
+
+        gc.collect()
+        empty_cache()
+        return result
+
+    if asyncio.iscoroutinefunction(fn):
+        return _async_wrapper
+    else:
+        return _wrapper

xinference/model/llm/transformers/yi_vl.py

@@ -29,6 +29,7 @@ from ..utils import (
     parse_messages,
 )
 from .core import PytorchChatModel, PytorchGenerateConfig
+from .utils import cache_clean
 
 logger = logging.getLogger(__name__)
 
@@ -99,6 +100,7 @@ class YiVLChatModel(PytorchChatModel):
                 raise RuntimeError("Only one image per message is supported by Yi VL.")
         return content
 
+    @cache_clean
     def chat(
         self,
         messages: List[Dict],