Qwen3-30B-A3B-int8_asym-ov / openvino_detokenizer.xml
Echo9Zulu's picture
Upload 14 files
a1d4525 verified
<?xml version="1.0"?>
<net name="detokenizer" version="11">
<layers>
<layer id="0" name="Parameter_6888385" type="Parameter" version="opset1">
<data shape="?,?" element_type="i64" />
<output>
<port id="0" precision="I64" names="Parameter_6888385">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="1" name="Convert_6888555" type="Convert" version="opset1">
<data destination_type="i32" />
<input>
<port id="0" precision="I64">
<dim>-1</dim>
<dim>-1</dim>
</port>
</input>
<output>
<port id="1" precision="I32">
<dim>-1</dim>
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="2" name="Constant_6888387" type="Const" version="opset1">
<data element_type="i32" shape="151669" offset="0" size="606676" />
<output>
<port id="0" precision="I32">
<dim>151669</dim>
</port>
</output>
</layer>
<layer id="3" name="Constant_6888389" type="Const" version="opset1">
<data element_type="i32" shape="151669" offset="606676" size="606676" />
<output>
<port id="0" precision="I32">
<dim>151669</dim>
</port>
</output>
</layer>
<layer id="4" name="Constant_6888391" type="Const" version="opset1">
<data element_type="u8" shape="976309" offset="1213352" size="976309" />
<output>
<port id="0" precision="U8">
<dim>976309</dim>
</port>
</output>
</layer>
<layer id="5" name="Slice_6888396" type="Const" version="opset1">
<data element_type="i32" shape="14" offset="2189661" size="56" />
<output>
<port id="0" precision="I32">
<dim>14</dim>
</port>
</output>
</layer>
<layer id="6" name="VocabDecoder_6888398" type="VocabDecoder" version="extension">
<data skip_tokens="" />
<input>
<port id="0" precision="I32">
<dim>-1</dim>
<dim>-1</dim>
</port>
<port id="1" precision="I32">
<dim>151669</dim>
</port>
<port id="2" precision="I32">
<dim>151669</dim>
</port>
<port id="3" precision="U8">
<dim>976309</dim>
</port>
<port id="4" precision="I32">
<dim>14</dim>
</port>
</input>
<output>
<port id="5" precision="I32">
<dim>-1</dim>
</port>
<port id="6" precision="I32">
<dim>-1</dim>
</port>
<port id="7" precision="I32">
<dim>-1</dim>
</port>
<port id="8" precision="I32">
<dim>-1</dim>
</port>
<port id="9" precision="U8">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="7" name="FuzeRagged_6888399" type="FuzeRagged" version="extension">
<input>
<port id="0" precision="I32">
<dim>-1</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="I32">
<dim>-1</dim>
</port>
<port id="3" precision="I32">
<dim>-1</dim>
</port>
</input>
<output>
<port id="4" precision="I32">
<dim>-1</dim>
</port>
<port id="5" precision="I32">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="8" name="UTF8Validate_6888400" type="UTF8Validate" version="extension">
<data replace_mode="true" />
<input>
<port id="0" precision="I32">
<dim>-1</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="U8">
<dim>-1</dim>
</port>
</input>
<output>
<port id="3" precision="I32">
<dim>-1</dim>
</port>
<port id="4" precision="I32">
<dim>-1</dim>
</port>
<port id="5" precision="U8">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="9" name="StringTensorPack_6888401" type="StringTensorPack" version="opset15">
<input>
<port id="0" precision="I32">
<dim>-1</dim>
</port>
<port id="1" precision="I32">
<dim>-1</dim>
</port>
<port id="2" precision="U8">
<dim>-1</dim>
</port>
</input>
<output>
<port id="3" precision="STRING" names="Result_6888402,string_output">
<dim>-1</dim>
</port>
</output>
</layer>
<layer id="10" name="Result_6888402" type="Result" version="opset1" output_names="Result_6888402,string_output">
<input>
<port id="0" precision="STRING">
<dim>-1</dim>
</port>
</input>
</layer>
</layers>
<edges>
<edge from-layer="0" from-port="0" to-layer="1" to-port="0" />
<edge from-layer="1" from-port="1" to-layer="6" to-port="0" />
<edge from-layer="2" from-port="0" to-layer="6" to-port="1" />
<edge from-layer="3" from-port="0" to-layer="6" to-port="2" />
<edge from-layer="4" from-port="0" to-layer="6" to-port="3" />
<edge from-layer="5" from-port="0" to-layer="6" to-port="4" />
<edge from-layer="6" from-port="7" to-layer="7" to-port="2" />
<edge from-layer="6" from-port="9" to-layer="8" to-port="2" />
<edge from-layer="6" from-port="8" to-layer="7" to-port="3" />
<edge from-layer="6" from-port="6" to-layer="7" to-port="1" />
<edge from-layer="6" from-port="5" to-layer="7" to-port="0" />
<edge from-layer="7" from-port="4" to-layer="8" to-port="0" />
<edge from-layer="7" from-port="5" to-layer="8" to-port="1" />
<edge from-layer="8" from-port="3" to-layer="9" to-port="0" />
<edge from-layer="8" from-port="4" to-layer="9" to-port="1" />
<edge from-layer="8" from-port="5" to-layer="9" to-port="2" />
<edge from-layer="9" from-port="3" to-layer="10" to-port="0" />
</edges>
<rt_info>
<add_attention_mask value="True" />
<add_prefix_space />
<add_special_tokens value="True" />
<chat_template value="{%- if tools %}&#10; {{- '&lt;|im_start|>system\n' }}&#10; {%- if messages[0].role == 'system' %}&#10; {{- messages[0].content + '\n\n' }}&#10; {%- endif %}&#10; {{- &quot;# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within &lt;tools>&lt;/tools> XML tags:\n&lt;tools>&quot; }}&#10; {%- for tool in tools %}&#10; {{- &quot;\n&quot; }}&#10; {{- tool | tojson }}&#10; {%- endfor %}&#10; {{- &quot;\n&lt;/tools>\n\nFor each function call, return a json object with function name and arguments within &lt;tool_call>&lt;/tool_call> XML tags:\n&lt;tool_call>\n{\&quot;name\&quot;: &lt;function-name>, \&quot;arguments\&quot;: &lt;args-json-object>}\n&lt;/tool_call>&lt;|im_end|>\n&quot; }}&#10;{%- else %}&#10; {%- if messages[0].role == 'system' %}&#10; {{- '&lt;|im_start|>system\n' + messages[0].content + '&lt;|im_end|>\n' }}&#10; {%- endif %}&#10;{%- endif %}&#10;{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}&#10;{%- for message in messages[::-1] %}&#10; {%- set index = (messages|length - 1) - loop.index0 %}&#10; {%- if ns.multi_step_tool and message.role == &quot;user&quot; and not(message.content.startswith('&lt;tool_response>') and message.content.endswith('&lt;/tool_response>')) %}&#10; {%- set ns.multi_step_tool = false %}&#10; {%- set ns.last_query_index = index %}&#10; {%- endif %}&#10;{%- endfor %}&#10;{%- for message in messages %}&#10; {%- if (message.role == &quot;user&quot;) or (message.role == &quot;system&quot; and not loop.first) %}&#10; {{- '&lt;|im_start|>' + message.role + '\n' + message.content + '&lt;|im_end|>' + '\n' }}&#10; {%- elif message.role == &quot;assistant&quot; %}&#10; {%- set content = message.content %}&#10; {%- set reasoning_content = '' %}&#10; {%- if message.reasoning_content is defined and message.reasoning_content is not none %}&#10; {%- set reasoning_content = message.reasoning_content %}&#10; {%- else %}&#10; {%- if '&lt;/think>' in message.content %}&#10; {%- set content = message.content.split('&lt;/think>')[-1].lstrip('\n') %}&#10; {%- set reasoning_content = message.content.split('&lt;/think>')[0].rstrip('\n').split('&lt;think>')[-1].lstrip('\n') %}&#10; {%- endif %}&#10; {%- endif %}&#10; {%- if loop.index0 > ns.last_query_index %}&#10; {%- if loop.last or (not loop.last and reasoning_content) %}&#10; {{- '&lt;|im_start|>' + message.role + '\n&lt;think>\n' + reasoning_content.strip('\n') + '\n&lt;/think>\n\n' + content.lstrip('\n') }}&#10; {%- else %}&#10; {{- '&lt;|im_start|>' + message.role + '\n' + content }}&#10; {%- endif %}&#10; {%- else %}&#10; {{- '&lt;|im_start|>' + message.role + '\n' + content }}&#10; {%- endif %}&#10; {%- if message.tool_calls %}&#10; {%- for tool_call in message.tool_calls %}&#10; {%- if (loop.first and content) or (not loop.first) %}&#10; {{- '\n' }}&#10; {%- endif %}&#10; {%- if tool_call.function %}&#10; {%- set tool_call = tool_call.function %}&#10; {%- endif %}&#10; {{- '&lt;tool_call>\n{&quot;name&quot;: &quot;' }}&#10; {{- tool_call.name }}&#10; {{- '&quot;, &quot;arguments&quot;: ' }}&#10; {%- if tool_call.arguments is string %}&#10; {{- tool_call.arguments }}&#10; {%- else %}&#10; {{- tool_call.arguments | tojson }}&#10; {%- endif %}&#10; {{- '}\n&lt;/tool_call>' }}&#10; {%- endfor %}&#10; {%- endif %}&#10; {{- '&lt;|im_end|>\n' }}&#10; {%- elif message.role == &quot;tool&quot; %}&#10; {%- if loop.first or (messages[loop.index0 - 1].role != &quot;tool&quot;) %}&#10; {{- '&lt;|im_start|>user' }}&#10; {%- endif %}&#10; {{- '\n&lt;tool_response>\n' }}&#10; {{- message.content }}&#10; {{- '\n&lt;/tool_response>' }}&#10; {%- if loop.last or (messages[loop.index0 + 1].role != &quot;tool&quot;) %}&#10; {{- '&lt;|im_end|>\n' }}&#10; {%- endif %}&#10; {%- endif %}&#10;{%- endfor %}&#10;{%- if add_generation_prompt %}&#10; {{- '&lt;|im_start|>assistant\n' }}&#10; {%- if enable_thinking is defined and enable_thinking is false %}&#10; {{- '&lt;think>\n\n&lt;/think>\n\n' }}&#10; {%- endif %}&#10;{%- endif %}" />
<clean_up_tokenization_spaces />
<detokenizer_input_type value="i64" />
<eos_token_id value="151645" />
<handle_special_tokens_with_re />
<max_length />
<number_of_inputs value="1" />
<openvino_tokenizers_version value="2025.1.0.0-523-710ddf14de8" />
<openvino_version value="2025.1.0-18503-6fec06580ab-releases/2025/1" />
<original_post_processor_template value="{&quot;type&quot;: &quot;ByteLevel&quot;, &quot;add_prefix_space&quot;: false, &quot;trim_offsets&quot;: false, &quot;use_regex&quot;: false}" />
<original_tokenizer_class value="&lt;class 'transformers.models.qwen2.tokenization_qwen2_fast.Qwen2TokenizerFast'>" />
<pad_token_id value="151643" />
<sentencepiece_version value="0.2.0" />
<skip_special_tokens value="True" />
<streaming_detokenizer value="False" />
<tokenizer_output_type value="i64" />
<tokenizers_version value="0.21.1" />
<transformers_version value="4.51.3" />
<use_max_padding value="False" />
<use_sentencepiece_backend value="False" />
<utf8_replace_mode value="replace" />
<with_detokenizer value="True" />
</rt_info>
</net>