From ff8db563315e2118a2911e012eba4ee0ccb0215f Mon Sep 17 00:00:00 2001 From: alpayariyak Date: Fri, 6 Sep 2024 18:20:10 -0700 Subject: [PATCH 1/3] Add Llama 3.1 Instruct Chat template, Ensure Correctness of 3, Refactors for Chat Templates --- src/axolotl/cli/preprocess.py | 23 +++-------------- src/axolotl/cli/train.py | 21 +++------------- src/axolotl/prompt_strategies/sharegpt.py | 25 +++++++++++++++---- src/axolotl/utils/chat_templates.py | 1 + .../config/models/input/v0_4_1/__init__.py | 1 + 5 files changed, 28 insertions(+), 43 deletions(-) diff --git a/src/axolotl/cli/preprocess.py b/src/axolotl/cli/preprocess.py index aab29e2670..4b32796cf6 100644 --- a/src/axolotl/cli/preprocess.py +++ b/src/axolotl/cli/preprocess.py @@ -23,10 +23,7 @@ ) from axolotl.common.cli import PreprocessCliArgs from axolotl.common.const import DEFAULT_DATASET_PREPARED_PATH -from axolotl.prompt_strategies.sharegpt import ( - register_chatml_template, - register_llama3_template, -) +from axolotl.prompt_strategies.sharegpt import register_chat_template from axolotl.utils.trainer import disable_datasets_caching LOG = logging.getLogger("axolotl.cli.preprocess") @@ -44,22 +41,8 @@ def do_cli(config: Union[Path, str] = Path("examples/"), **kwargs): return_remaining_strings=True ) - if parsed_cfg.chat_template == "chatml": - if parsed_cfg.default_system_message: - LOG.info( - f"ChatML set. Adding default system message: {parsed_cfg.default_system_message}" - ) - register_chatml_template(parsed_cfg.default_system_message) - else: - register_chatml_template() - elif parsed_cfg.chat_template == "llama3": - if parsed_cfg.default_system_message: - LOG.info( - f"LLaMA-3 set. Adding default system message: {parsed_cfg.default_system_message}" - ) - register_llama3_template(parsed_cfg.default_system_message) - else: - register_llama3_template() + if parsed_cfg.chat_template: + register_chat_template(parsed_cfg.chat_template, parsed_cfg.default_system_message) if not parsed_cfg.dataset_prepared_path: msg = ( diff --git a/src/axolotl/cli/train.py b/src/axolotl/cli/train.py index 16d66a82f0..b7b214ea50 100644 --- a/src/axolotl/cli/train.py +++ b/src/axolotl/cli/train.py @@ -19,10 +19,7 @@ ) from axolotl.common.cli import TrainerCliArgs from axolotl.integrations.base import PluginManager -from axolotl.prompt_strategies.sharegpt import ( - register_chatml_template, - register_llama3_template, -) +from axolotl.prompt_strategies.sharegpt import register_chat_template from axolotl.train import train LOG = logging.getLogger("axolotl.cli.train") @@ -42,21 +39,9 @@ def do_train(cfg, cli_args) -> None: print_axolotl_text_art() check_accelerate_default_config() check_user_token() - if cfg.chat_template == "chatml" and cfg.default_system_message: - LOG.info( - f"ChatML set. Adding default system message: {cfg.default_system_message}" - ) - register_chatml_template(cfg.default_system_message) - else: - register_chatml_template() - if cfg.chat_template == "llama3" and cfg.default_system_message: - LOG.info( - f"LLaMA-3 set. Adding default system message: {cfg.default_system_message}" - ) - register_llama3_template(cfg.default_system_message) - else: - register_llama3_template() + if cfg.chat_template: + register_chat_template(cfg.chat_template, cfg.default_system_message) if cfg.rl: # and cfg.rl != "orpo": dataset_meta = load_rl_datasets(cfg=cfg, cli_args=cli_args) diff --git a/src/axolotl/prompt_strategies/sharegpt.py b/src/axolotl/prompt_strategies/sharegpt.py index 4565c35d5d..d87d7e3b9e 100644 --- a/src/axolotl/prompt_strategies/sharegpt.py +++ b/src/axolotl/prompt_strategies/sharegpt.py @@ -39,12 +39,12 @@ def register_chatml_template(system_message=None): ) -def register_llama3_template(system_message=None): - system_message = system_message or "You are a helpful assistant." +def register_llama3x_template(version, system_message=None): + extra_system_content_to_insert = "Cutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n" if version == "3.1" else "" register_conv_template( Conversation( - name="llama3", - system_template="<|start_header_id|>system<|end_header_id|>\n\n{system_message}<|eot_id|>", + name=f"llama{version.replace('.', '')}", + system_template="<|start_header_id|>system<|end_header_id|>\n\n{extra_system_content_to_insert}{system_message}<|eot_id|>", system_message=system_message, roles=("user", "assistant"), sep_style=SeparatorStyle.LLAMA3, @@ -95,6 +95,21 @@ def _load(tokenizer, cfg, ds_cfg: Optional[Dict[str, Any]] = None): return _load +def register_chat_template(chat_template, default_system_message=None): + if chat_template == "chatml" or chat_template.startswith("llama"): + if default_system_message: + LOG.info(f"Adding default system message: {default_system_message}") + + if chat_template == "chatml": + register_chatml_template(default_system_message) + LOG.info(f"Using ChatML template") + else: + version = ".".join(list(chat_template.split("llama"))[1]) + if version not in ["3", "3.1"]: + raise ValueError(f"Invalid name for Llama 3x template: {chat_template}, only llama3 and llama31 are supported") + LOG.info(f"Using Llama {version} template") + register_llama3x_template(version, default_system_message) + class SimpleShareGPTPromptTokenizingStrategy(ShareGPTPromptTokenizingStrategy): """ @@ -103,7 +118,7 @@ class SimpleShareGPTPromptTokenizingStrategy(ShareGPTPromptTokenizingStrategy): _strict = False _messages = "conversations" - +q @property def strict(self): return self._strict diff --git a/src/axolotl/utils/chat_templates.py b/src/axolotl/utils/chat_templates.py index 2443f56f93..30e40d7749 100644 --- a/src/axolotl/utils/chat_templates.py +++ b/src/axolotl/utils/chat_templates.py @@ -12,6 +12,7 @@ "gemma": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '' + role + '\n' + message['content'] | trim + '\n' }}{% endfor %}{% if add_generation_prompt %}{{'model\n'}}{% endif %}", "cohere": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif false == true %}{% set loop_messages = messages %}{% set system_message = 'You are Command-R, a brilliant, sophisticated, AI-assistant trained to assist human users by providing thorough responses. You are trained by Cohere.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% if system_message != false %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' + system_message + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|START_OF_TURN_TOKEN|><|USER_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'assistant' %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' }}{% endif %}", "llama3": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}", + "llama31": "{{- 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", "llama3_2_vision": '{{- 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 {%- if strftime_now is defined %}\n {%- set date_string = strftime_now("%d %b %Y") %}\n {%- else %}\n {%- set date_string = "26 Jul 2024" %}\n {%- endif %}\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{#- Find out if there are any images #}\n{% set image_ns = namespace(has_images=false) %} \n{%- for message in messages %}\n {%- for content in message[\'content\'] %}\n {%- if content[\'type\'] == \'image\' %}\n {%- set image_ns.has_images = true %}\n {%- endif %}\n {%- endfor %}\n{%- endfor %}\n\n{#- Error out if there are images and system message #}\n{%- if image_ns.has_images and not system_message == "" %}\n {{- raise_exception("Prompting with images is incompatible with system messages.") }}\n{%- endif %}\n\n{#- System message if there are no images #}\n{%- if not image_ns.has_images %}\n {{- "<|start_header_id|>system<|end_header_id|>\\n\\n" }}\n {%- if tools is not none %}\n {{- "Environment: ipython\\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{%- endif %}\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\' }}\n {%- if message[\'content\'] is string %}\n {{- message[\'content\'] }}\n {%- else %}\n {%- for content in message[\'content\'] %}\n {%- if content[\'type\'] == \'image\' %}\n {{- \'<|image|>\' }}\n {%- elif content[\'type\'] == \'text\' %}\n {{- content[\'text\'] }}\n {%- endif %}\n {%- endfor %}\n {%- endif %}\n {{- \'<|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 {{- \'<|start_header_id|>assistant<|end_header_id|>\\n\\n\' -}}\n {{- \'{"name": "\' + tool_call.name + \'", \' }}\n {{- \'"parameters": \' }}\n {{- tool_call.arguments | tojson }}\n {{- "}" }}\n {{- "<|eot_id|>" }}\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', "phi_3": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'system') %}{{'<|system|>' + '\n' + message['content'] + '<|end|>' + '\n'}}{% elif (message['role'] == 'user') %}{{'<|user|>' + '\n' + message['content'] + '<|end|>' + '\n' + '<|assistant|>' + '\n'}}{% elif message['role'] == 'assistant' %}{{message['content'] + '<|end|>' + '\n'}}{% endif %}{% endfor %}", "phi_35": "{% for message in messages %}{% if message['role'] == 'system' and message['content'] %}{{'<|system|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'user' %}{{'<|user|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'assistant' %}{{'<|assistant|>\n' + message['content'] + '<|end|>\n'}}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}", diff --git a/src/axolotl/utils/config/models/input/v0_4_1/__init__.py b/src/axolotl/utils/config/models/input/v0_4_1/__init__.py index 4831da3c8a..34b6cb77e7 100644 --- a/src/axolotl/utils/config/models/input/v0_4_1/__init__.py +++ b/src/axolotl/utils/config/models/input/v0_4_1/__init__.py @@ -195,6 +195,7 @@ class ChatTemplate(str, Enum): gemma = "gemma" # pylint: disable=invalid-name cohere = "cohere" # pylint: disable=invalid-name llama3 = "llama3" # pylint: disable=invalid-name + llama31 = "llama31" # pylint: disable=invalid-name llama3_2_vision = "llama3_2_vision" # pylint: disable=invalid-name phi_3 = "phi_3" # pylint: disable=invalid-name phi_35 = "phi_35" # pylint: disable=invalid-name From ab414701d6c820a9ca5d31b200747887ff3eb9c0 Mon Sep 17 00:00:00 2001 From: Wing Lian Date: Tue, 10 Sep 2024 10:39:43 -0400 Subject: [PATCH 2/3] lint, fix templating, default version --- src/axolotl/cli/preprocess.py | 4 +++- src/axolotl/prompt_strategies/sharegpt.py | 19 ++++++++++++++----- src/axolotl/utils/chat_templates.py | 2 +- tests/prompt_strategies/test_sharegpt.py | 4 ++-- 4 files changed, 20 insertions(+), 9 deletions(-) diff --git a/src/axolotl/cli/preprocess.py b/src/axolotl/cli/preprocess.py index 4b32796cf6..592b6c2cf4 100644 --- a/src/axolotl/cli/preprocess.py +++ b/src/axolotl/cli/preprocess.py @@ -42,7 +42,9 @@ def do_cli(config: Union[Path, str] = Path("examples/"), **kwargs): ) if parsed_cfg.chat_template: - register_chat_template(parsed_cfg.chat_template, parsed_cfg.default_system_message) + register_chat_template( + parsed_cfg.chat_template, parsed_cfg.default_system_message + ) if not parsed_cfg.dataset_prepared_path: msg = ( diff --git a/src/axolotl/prompt_strategies/sharegpt.py b/src/axolotl/prompt_strategies/sharegpt.py index d87d7e3b9e..7ca5559fea 100644 --- a/src/axolotl/prompt_strategies/sharegpt.py +++ b/src/axolotl/prompt_strategies/sharegpt.py @@ -39,12 +39,18 @@ def register_chatml_template(system_message=None): ) -def register_llama3x_template(version, system_message=None): - extra_system_content_to_insert = "Cutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n" if version == "3.1" else "" +def register_llama3x_template(version="", system_message=None): + extra_system_content_to_insert = ( + "Cutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n" + if version == "3.1" + else "" + ) register_conv_template( Conversation( name=f"llama{version.replace('.', '')}", - system_template="<|start_header_id|>system<|end_header_id|>\n\n{extra_system_content_to_insert}{system_message}<|eot_id|>", + system_template="<|start_header_id|>system<|end_header_id|>\n\n" + + extra_system_content_to_insert + + "{system_message}<|eot_id|>", system_message=system_message, roles=("user", "assistant"), sep_style=SeparatorStyle.LLAMA3, @@ -95,6 +101,7 @@ def _load(tokenizer, cfg, ds_cfg: Optional[Dict[str, Any]] = None): return _load + def register_chat_template(chat_template, default_system_message=None): if chat_template == "chatml" or chat_template.startswith("llama"): if default_system_message: @@ -102,11 +109,13 @@ def register_chat_template(chat_template, default_system_message=None): if chat_template == "chatml": register_chatml_template(default_system_message) - LOG.info(f"Using ChatML template") + LOG.info("Using ChatML template") else: version = ".".join(list(chat_template.split("llama"))[1]) if version not in ["3", "3.1"]: - raise ValueError(f"Invalid name for Llama 3x template: {chat_template}, only llama3 and llama31 are supported") + raise ValueError( + f"Invalid name for Llama 3x template: {chat_template}, only llama3 and llama31 are supported" + ) LOG.info(f"Using Llama {version} template") register_llama3x_template(version, default_system_message) diff --git a/src/axolotl/utils/chat_templates.py b/src/axolotl/utils/chat_templates.py index 30e40d7749..073bd0538a 100644 --- a/src/axolotl/utils/chat_templates.py +++ b/src/axolotl/utils/chat_templates.py @@ -12,7 +12,7 @@ "gemma": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '' + role + '\n' + message['content'] | trim + '\n' }}{% endfor %}{% if add_generation_prompt %}{{'model\n'}}{% endif %}", "cohere": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif false == true %}{% set loop_messages = messages %}{% set system_message = 'You are Command-R, a brilliant, sophisticated, AI-assistant trained to assist human users by providing thorough responses. You are trained by Cohere.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% if system_message != false %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' + system_message + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|START_OF_TURN_TOKEN|><|USER_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'assistant' %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' }}{% endif %}", "llama3": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}", - "llama31": "{{- 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", + "llama31": '{{- 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', "llama3_2_vision": '{{- 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 {%- if strftime_now is defined %}\n {%- set date_string = strftime_now("%d %b %Y") %}\n {%- else %}\n {%- set date_string = "26 Jul 2024" %}\n {%- endif %}\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{#- Find out if there are any images #}\n{% set image_ns = namespace(has_images=false) %} \n{%- for message in messages %}\n {%- for content in message[\'content\'] %}\n {%- if content[\'type\'] == \'image\' %}\n {%- set image_ns.has_images = true %}\n {%- endif %}\n {%- endfor %}\n{%- endfor %}\n\n{#- Error out if there are images and system message #}\n{%- if image_ns.has_images and not system_message == "" %}\n {{- raise_exception("Prompting with images is incompatible with system messages.") }}\n{%- endif %}\n\n{#- System message if there are no images #}\n{%- if not image_ns.has_images %}\n {{- "<|start_header_id|>system<|end_header_id|>\\n\\n" }}\n {%- if tools is not none %}\n {{- "Environment: ipython\\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{%- endif %}\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\' }}\n {%- if message[\'content\'] is string %}\n {{- message[\'content\'] }}\n {%- else %}\n {%- for content in message[\'content\'] %}\n {%- if content[\'type\'] == \'image\' %}\n {{- \'<|image|>\' }}\n {%- elif content[\'type\'] == \'text\' %}\n {{- content[\'text\'] }}\n {%- endif %}\n {%- endfor %}\n {%- endif %}\n {{- \'<|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 {{- \'<|start_header_id|>assistant<|end_header_id|>\\n\\n\' -}}\n {{- \'{"name": "\' + tool_call.name + \'", \' }}\n {{- \'"parameters": \' }}\n {{- tool_call.arguments | tojson }}\n {{- "}" }}\n {{- "<|eot_id|>" }}\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', "phi_3": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'system') %}{{'<|system|>' + '\n' + message['content'] + '<|end|>' + '\n'}}{% elif (message['role'] == 'user') %}{{'<|user|>' + '\n' + message['content'] + '<|end|>' + '\n' + '<|assistant|>' + '\n'}}{% elif message['role'] == 'assistant' %}{{message['content'] + '<|end|>' + '\n'}}{% endif %}{% endfor %}", "phi_35": "{% for message in messages %}{% if message['role'] == 'system' and message['content'] %}{{'<|system|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'user' %}{{'<|user|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'assistant' %}{{'<|assistant|>\n' + message['content'] + '<|end|>\n'}}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}", diff --git a/tests/prompt_strategies/test_sharegpt.py b/tests/prompt_strategies/test_sharegpt.py index e7a73a0de5..a3d3253a9e 100644 --- a/tests/prompt_strategies/test_sharegpt.py +++ b/tests/prompt_strategies/test_sharegpt.py @@ -12,12 +12,12 @@ GlaiveShareGPTPromptTokenizingStrategy, SimpleShareGPTPromptTokenizingStrategy, register_chatml_template, - register_llama3_template, + register_llama3x_template, ) from axolotl.prompters import ShareGPTPrompterV2 register_chatml_template() -register_llama3_template() +register_llama3x_template() @pytest.fixture(name="sharegpt_dataset") From 1f5cd4e081b12ac7d3c6f67eadf7af920dfeab5e Mon Sep 17 00:00:00 2001 From: Wing Lian Date: Tue, 10 Sep 2024 11:21:52 -0400 Subject: [PATCH 3/3] ensure the default is llama3 --- src/axolotl/prompt_strategies/sharegpt.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/axolotl/prompt_strategies/sharegpt.py b/src/axolotl/prompt_strategies/sharegpt.py index 7ca5559fea..cc7ccf1190 100644 --- a/src/axolotl/prompt_strategies/sharegpt.py +++ b/src/axolotl/prompt_strategies/sharegpt.py @@ -39,7 +39,7 @@ def register_chatml_template(system_message=None): ) -def register_llama3x_template(version="", system_message=None): +def register_llama3x_template(version="3", system_message=None): extra_system_content_to_insert = ( "Cutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n" if version == "3.1"