Team Ai
Modelpublic

ai21labs/Jamba-tiny-dev

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
14likes352kdownloads
tokenizer_config.json197 linesDownload Raw Back to root
1{2  "add_bos_token": true,3  "add_eos_token": false,4  "add_prefix_space": null,5  "added_tokens_decoder": {6    "0": {7      "content": "<|pad|>",8      "lstrip": false,9      "normalized": false,10      "rstrip": false,11      "single_word": false,12      "special": true13    },14    "1": {15      "content": "<|startoftext|>",16      "lstrip": false,17      "normalized": false,18      "rstrip": false,19      "single_word": false,20      "special": true21    },22    "2": {23      "content": "<|endoftext|>",24      "lstrip": false,25      "normalized": false,26      "rstrip": false,27      "single_word": false,28      "special": true29    },30    "3": {31      "content": "<|unk|>",32      "lstrip": false,33      "normalized": false,34      "rstrip": false,35      "single_word": false,36      "special": true37    },38    "518": {39      "content": "<|eom|>",40      "lstrip": false,41      "normalized": false,42      "rstrip": false,43      "single_word": false,44      "special": true45    },46    "519": {47      "content": "<|bom|>",48      "lstrip": false,49      "normalized": false,50      "rstrip": false,51      "single_word": false,52      "special": true53    },54    "520": {55      "content": "<|system|>",56      "lstrip": false,57      "normalized": false,58      "rstrip": false,59      "single_word": false,60      "special": true61    },62    "521": {63      "content": "<|user|>",64      "lstrip": false,65      "normalized": false,66      "rstrip": false,67      "single_word": false,68      "special": true69    },70    "522": {71      "content": "<|assistant|>",72      "lstrip": false,73      "normalized": false,74      "rstrip": false,75      "single_word": false,76      "special": true77    },78    "523": {79      "content": "<|tool|>",80      "lstrip": false,81      "normalized": false,82      "rstrip": false,83      "single_word": false,84      "special": true85    },86    "524": {87      "content": "<documents>",88      "lstrip": false,89      "normalized": false,90      "rstrip": false,91      "single_word": false,92      "special": true93    },94    "525": {95      "content": "</documents>",96      "lstrip": false,97      "normalized": false,98      "rstrip": false,99      "single_word": false,100      "special": true101    },102    "526": {103      "content": "<tool_definitions>",104      "lstrip": false,105      "normalized": false,106      "rstrip": false,107      "single_word": false,108      "special": true109    },110    "527": {111      "content": "</tool_definitions>",112      "lstrip": false,113      "normalized": false,114      "rstrip": false,115      "single_word": false,116      "special": true117    },118    "528": {119      "content": "<active_output_modes>",120      "lstrip": false,121      "normalized": false,122      "rstrip": false,123      "single_word": false,124      "special": true125    },126    "529": {127      "content": "</active_output_modes>",128      "lstrip": false,129      "normalized": false,130      "rstrip": false,131      "single_word": false,132      "special": true133    },134    "530": {135      "content": "<citations>",136      "lstrip": false,137      "normalized": false,138      "rstrip": false,139      "single_word": false,140      "special": true141    },142    "531": {143      "content": "</citations>",144      "lstrip": false,145      "normalized": false,146      "rstrip": false,147      "single_word": false,148      "special": true149    },150    "532": {151      "content": "<tool_calls>",152      "lstrip": false,153      "normalized": false,154      "rstrip": false,155      "single_word": false,156      "special": true157    },158    "533": {159      "content": "</tool_calls>",160      "lstrip": false,161      "normalized": false,162      "rstrip": false,163      "single_word": false,164      "special": true165    }166  },167  "additional_special_tokens": [168    "<|eom|>",169    "<|bom|>",170    "<|system|>",171    "<|user|>",172    "<|assistant|>",173    "<|tool|>",174    "<documents>",175    "</documents>",176    "<tool_definitions>",177    "</tool_definitions>",178    "<active_output_modes>",179    "</active_output_modes>",180    "<citations>",181    "</citations>",182    "<tool_calls>",183    "</tool_calls>"184  ],185  "bos_token": "<|startoftext|>",186  "chat_template": "{# Variables #}\n{% set ns = namespace(message_count=0, is_last_checked_defined=False) %}\n{##}\n{% set bom_str = bom_str or \"<|bom|>\" %}\n{% set eom_str = eom_str or \"<|eom|>\" %}\n{% set default_system_message = \"\" %}\n{##}\n{% set documents_prefix = \"<documents>\" %}\n{% set documents_suffix = \"</documents>\" %}\n{% set tool_definitions_prefix = \"<tool_definitions>\" %}\n{% set tool_definitions_suffix = \"</tool_definitions>\" %}\n{% set active_modes_prefix = \"<active_output_modes>\" %}\n{% set active_modes_suffix = \"</active_output_modes>\" %}\n{##}\n{% set tool_calls_prefix = \"<tool_calls>\" %}\n{% set tool_calls_suffix = \"</tool_calls>\" %}\n{% set citations_prefix = \"<citations>\" %}\n{% set citations_suffix = \"</citations>\" %}\n{##}\n{% if add_generation_prompt is not defined %}\n  {% set add_generation_prompt = True %}\n{% endif %}\n{% set role_to_predict = role_to_predict or \"assistant\" %}\n{% if messages|length > 0 and messages[0].role == \"system\" %}\n  {% set system_message = messages[0].content %}\n  {% set loop_messages = messages[1:] %}\n{% else %}\n  {% set system_message = default_system_message %}\n  {% set loop_messages = messages %}\n{% endif %}\n{##}\n{##}\n{# Macros #}\n{% macro handle_tool_definitions(tools) %}\n  {{- tool_definitions_prefix -}}\n  {{- \"\\n# Tools\" -}}\n  {{- \"\\n\\n## Functions\" -}}\n  {% for tool in tools %}\n    {% set _ = is_param_set(tool, field=\"type\") %}\n    {% set is_tool_type_set = ns.is_last_checked_defined %}\n    {% if is_tool_type_set %}\n      {% if tool.type == \"function\" %}\n        {% set tool = tool.function %}\n      {% else %}\n        {{ raise_exception(\"Currently, the only supported tool type is `function`\") }}\n      {% endif %}\n    {% endif %}\n    {{- \"\\n\\n\" + (tool|tojson(indent=2)) -}}\n  {% endfor %}\n  {{- \"\\n\" + tool_definitions_suffix -}}\n{% endmacro %}\n{##}\n{% macro handle_first_system_message(system_message, tools) %}\n  {{- bom_str + handle_role(\"system\") -}}\n  {% set _ = is_param_set(system_message) %}\n  {% set is_system_message_set = ns.is_last_checked_defined %}\n  {% if is_system_message_set %}\n    {{- system_message -}}\n  {% endif %}\n  {% set _ = is_param_set(tools, is_list=True) %}\n  {% set is_tools_set = ns.is_last_checked_defined %}\n  {% if is_tools_set %}\n    {% if system_message %}\n      {{- \"\\n\\n\" -}}\n    {% endif %}\n    {{- handle_tool_definitions(tools) -}}\n  {% endif %}\n  {% set ns.message_count = ns.message_count + 1 %}\n{% endmacro %}\n{##}\n{% macro handle_tool_calls(tool_calls) %}\n  {{- tool_calls_prefix + \"[\\n\" -}}\n  {% for tool_call in tool_calls %}\n    {% set _ = is_param_set(tool_call, field=\"function\") %}\n    {% set is_tool_call_function_set = ns.is_last_checked_defined %}\n    {% if is_tool_call_function_set %}\n      {%- set tool_call = tool_call.function %}\n    {%- endif %}\n    {% set arguments = tool_call.arguments %}\n    {% if arguments is not string %}\n      {%- set arguments = arguments|tojson -%}\n    {%- endif %}\n    {{ \"{\\\"name\\\": \\\"\" + tool_call.name + \"\\\", \\\"arguments\\\": \" + arguments + \"}\" -}}\n    {% if not loop.last %}\n      {{- \",\" }}\n    {% endif %}\n  {% endfor %}\n  {{- \"\\n]\" + tool_calls_suffix -}}\n{% endmacro %}\n{##}\n{% macro handle_documents(documents) %}\n  {{- documents_prefix -}}\n  {{- \"\\n# Documents\" -}}\n  {{- \"\\n\\nYou can use the following documents for reference:\" -}}\n  {% for doc in documents %}\n    {{- \"\\n\\n## Document ID: \" + loop.index0|string -}}\n    {% set _ = is_param_set(doc, field=\"title\") %}\n    {% set is_doc_title_set = ns.is_last_checked_defined %}\n    {% if is_doc_title_set %}\n      {{- \"\\nTitle: \" + doc.title -}}\n    {% endif %}\n    {% for key, value in doc.items() %}\n      {% if key not in [\"title\", \"text\"] %}\n        {{- \"\\n\" + key|title + \": \" + value|string -}}\n      {% endif %}\n    {% endfor %}\n    {{- \"\\nText: \" + doc.text -}}\n  {% endfor %}\n  {{- \"\\n\" + documents_suffix -}}\n{% endmacro %}\n{##}\n{% macro handle_knobs(knobs) %}\n  {{- active_modes_prefix -}}\n  {{- \"\\n# Active Modes\" -}}\n  {{ \"\\n\\nThe following modes configure the format or style of your responses. You should adhere to all currently\" -}}\n  {{ \" active modes simultaneously.\" -}}\n  {% if knobs.citation_mode == \"fast\" %}\n    {{- \"\\n\\n## Citation Mode\" -}}\n    {{- \"\\n\\nProvide a list of references only for the documents you base your response on. Format your response\" -}}\n    {{ \" with the original answer followed by a citation section. Use this template:\" -}}\n    {{ \" `{answer}\" + citations_prefix + \"DOCUMENT_IDS\" + citations_suffix + \"`, where DOCUMENT_IDS are the relevant document numbers\" -}}\n    {{ \" (e.g. [2, 5, 9]), or [] if the answer cannot be supported by the provided documents.\" -}}\n  {% endif %}\n  {% if knobs.response_format == \"json_object\" %}\n    {{- \"\\n\\n## JSON Mode\" -}}\n    {{ \"\\n\\nProvide your response in JSON format. Adhere strictly to any schema given by the user.\" -}}\n    {{ \" If an appropriate JSON format exists, use it without modification.\" -}}\n  {% endif %}\n  {{- \"\\n\" + active_modes_suffix -}}\n{% endmacro %}\n{##}\n{% macro get_last_user_index(messages) %}\n  {% set ns.last_user_index = 0 %}\n  {% for message in messages %}\n    {% if message.role == 'user' %}\n      {% set ns.last_user_index = loop.index0 %}\n    {% endif %}\n  {% endfor %}\n  {{- ns.last_user_index -}}\n{% endmacro %}\n{##}\n{% macro handle_last_system_message(documents, knobs, use_documents, use_knobs) %}\n  {{- bom_str + handle_role(\"system\") -}}\n  {% set macros_to_call = [] %}\n  {% set params_for_macros = [] %}\n  {% if use_documents %}\n    {% set macros_to_call = macros_to_call + [handle_documents] %}\n    {% set params_for_macros = params_for_macros + [[documents]] %}\n  {% endif %}\n  {% if use_knobs %}\n    {% set macros_to_call = macros_to_call + [handle_knobs] %}\n    {% set params_for_macros = params_for_macros + [[knobs]] %}\n  {% endif %}\n  {% for i in range(macros_to_call|length) %}\n    {% if i > 0 %}\n      {{- \"\\n\\n\" -}}\n    {% endif %}\n    {{- macros_to_call[i](*params_for_macros[i]) -}}\n  {% endfor %}\n  {% set ns.message_count = ns.message_count + 1 %}\n{% endmacro %}\n{##}\n{% macro handle_role(role, add_space=True) %}\n  {{- \"<|\" + role + \"|>\" -}}\n  {% if add_space %}\n    {{- \" \" -}}\n  {% endif %}\n{% endmacro %}\n{##}\n{% macro is_param_set(param, field=none, is_list=False) %}\n  {% if field is not none %}\n    {% if field in param %}\n      {% set param = param[field] %}\n    {% else %}\n      {% set param = none %}\n    {% endif %}\n  {% endif %}\n  {% set is_defined = param is defined and param is not none %}\n  {% if is_list %}\n    {% set ns.is_last_checked_defined = is_defined and param|length > 0 %}\n  {% else %}\n    {% set ns.is_last_checked_defined = is_defined %}\n  {% endif %}\n{% endmacro %}\n{##}\n{##}\n{# Template #}\n{{- \"<|startoftext|>\" -}}\n{% set _ = is_param_set(system_message) %}\n{% set is_system_message_set = ns.is_last_checked_defined %}\n{% set _ = is_param_set(tools, is_list=True) %}\n{% set is_tools_set = ns.is_last_checked_defined %}\n{% set has_system_message = (is_system_message_set or is_tools_set) %}\n{% if has_system_message %}\n  {{- handle_first_system_message(system_message, tools) -}}\n{% endif %}\n{% set last_user_index = get_last_user_index(loop_messages)|int %}\n{% for message in loop_messages %}\n  {% if loop.index0 == last_user_index %}\n    {% set _ = is_param_set(documents, is_list=True) %}\n    {% set use_documents = ns.is_last_checked_defined %}\n    {% set _ = is_param_set(knobs) %}\n    {% set use_knobs = ns.is_last_checked_defined and knobs.is_set %}\n    {% set add_last_system_message = use_documents or use_knobs %}\n    {% if add_last_system_message %}\n      {% if ns.message_count > 0 %}\n        {{- eom_str -}}\n      {% endif %}\n      {{- handle_last_system_message(documents, knobs, use_documents, use_knobs) -}}\n    {% endif %}\n  {% endif %}\n  {% set role = message.role %}\n  {% set _ = is_param_set(message, field=\"name\") %}\n  {% set is_message_name_set = ns.is_last_checked_defined %}\n  {% if is_message_name_set %}\n    {% set message_prefix = handle_role(role) + \"(\" + message.name + \")\" %}\n  {% else %}\n    {% set message_prefix = handle_role(role) %}\n  {% endif %}\n  {% set content = (message.content or \"\") %}\n  {% if content is not string %}\n    {% set content = content|tojson %}\n  {% endif %}\n  {% if ns.message_count > 0 %}\n    {{- eom_str -}}\n  {% endif %}\n  {{- bom_str + message_prefix + content -}}\n  {% set _ = is_param_set(message, field=\"tool_calls\", is_list=True) %}\n  {% set is_tool_calls_set = ns.is_last_checked_defined %}\n  {% if role == \"assistant\" and is_tool_calls_set %}\n    {{- handle_tool_calls(message.tool_calls) -}}\n  {% endif %}\n  {% set _ = is_param_set(message, field=\"citations\", is_list=True) %}\n  {% set is_citations_set = ns.is_last_checked_defined %}\n  {% if role == \"assistant\" and is_citations_set %}\n    {{- citations_prefix + message.citations|map(attribute=\"document_id\")|list|string + citations_suffix -}}\n  {% endif %}\n  {% set ns.message_count = ns.message_count + 1 %}\n{% endfor %}\n{% if add_generation_prompt %}\n  {% if ns.message_count > 0 %}\n    {{- eom_str -}}\n  {% endif %}\n  {{- bom_str + handle_role(role_to_predict, add_space=False) -}}\n  {% set _ = is_param_set(generation_preamble) %}\n  {% set is_generation_preamble_set = ns.is_last_checked_defined %}\n  {% if is_generation_preamble_set and generation_preamble.strip() != \"\" %}\n    {{- \" \" + generation_preamble -}}\n  {% endif %}\n  {% set ns.message_count = ns.message_count + 1 %}\n{% else %}\n  {% if ns.message_count > 0 %}\n    {{- eom_str -}}\n  {% endif %}\n{% endif %}\n",187  "clean_up_tokenization_spaces": false,188  "eos_token": "<|endoftext|>",189  "legacy": true,190  "model_max_length": 1000000000000000019884624838656,191  "pad_token": "<|pad|>",192  "spaces_between_special_tokens": false,193  "tokenizer_class": "LlamaTokenizer",194  "unk_token": "<|unk|>",195  "use_default_system_prompt": false196}197