Functions — github.com/anarchy-ai/LLM-VM

hub / github.com/anarchy-ai/LLM-VM / functions

Functions280 in github.com/anarchy-ai/LLM-VM

⨍ Functions 280◇ Types & classes 98 ↳ Endpoints 2

↓ 25 callersFunctionprint_op

(*kargs, **kwargs)

src/llm_vm/agents/REBEL/utils.py:19

↓ 22 callersFunctionprint_big

(data, label = "")

src/llm_vm/agents/FLAT/agent_helper/utils.py:49

↓ 18 callersMethodcreate

(regex, type, choices, grammar_type, *, default=None)

src/llm_vm/guided_completion.py:23

↓ 12 callersMethodcomplete

This function is Anarchy's completion entry point Parameters: prompt (str): Prompt to send to LLM for generation

src/llm_vm/client.py:104

↓ 10 callersFunctionMSG

(u, c)

src/llm_vm/agents/REBEL/utils.py:27

↓ 10 callersFunction__shuffled

(tool_id: int)

src/llm_vm/agents/FLAT/models/utils/tool_picker_model/get_training_tools.py:10

↓ 8 callersFunctioncall_llm

(llm_request: LLMCallParams)

src/llm_vm/agents/FLAT/agent_helper/requests/call_llm.py:4

↓ 8 callersMethodmakeInteraction

Formats the tool description to contain the relevant tags according to the dynamic params Parameters ---------- p

src/llm_vm/agents/REBEL/agent.py:310

↓ 7 callersFunctionprint_op

(*kargs, **kwargs)

src/llm_vm/agents/FLAT/agent_helper/utils.py:22

↓ 7 callersMethodset_training_in_progress

(self, c_id, is_training)

src/llm_vm/completion/optimize.py:78

↓ 6 callersMethodgenerate

This function uses openAI's API to generate a response from the prompt Parameters: prompt (str): Prompt to send to LLM

src/llm_vm/onsite_llm.py:739

↓ 6 callersMethodinit_if_null

(self, c_id)

src/llm_vm/completion/optimize.py:100

↓ 6 callersMethodrun

(self, question: str, memory: TupleList)

src/llm_vm/agents/FLAT/agent.py:39

↓ 5 callersFunctioncalcCost

(p)

src/llm_vm/agents/REBEL/bothandler.py:33

↓ 5 callersFunctioncall_ChatGPT

(state, cur_prompt, stop=None, max_tokens=20, temperature=0.2)

src/llm_vm/agents/REBEL/utils.py:31

↓ 5 callersMethodget_model

(self, c_id)

src/llm_vm/completion/optimize.py:96

↓ 5 callersFunctionmake_interaction

(human_question, ai_response, data = '', Q = L_QUESTION, A = L_ANSWE

src/llm_vm/agents/FLAT/agent_helper/utils.py:95

↓ 4 callersFunctionget_newest_decision_model

(model: DecisionStep, default_model = OpenAIModel.DAVINCI_TEXT)

src/llm_vm/agents/FLAT/models/get_decision_model.py:6

↓ 4 callersMethodset_model

(self, c_id, model_id)

src/llm_vm/completion/optimize.py:92

↓ 4 callersMethodset_tools

(self, tools)

src/llm_vm/agents/FLAT/agent.py:22

↓ 3 callersFunctionasyncAwait

(t)

src/llm_vm/completion/optimize.py:52

↓ 3 callersFunctionasyncStart

(foo)

src/llm_vm/completion/optimize.py:41

↓ 3 callersMethodcomplete

Runs a completion using the string stable_context+dynamic_prompt. Returns an optional training closure to use if the caller decides

src/llm_vm/completion/optimize.py:111

↓ 3 callersFunctionformat_simple_value

(value: str)

src/llm_vm/agents/FLAT/agent_helper/replacer.py:18

↓ 3 callersMethodget_data

(self, c_id)

src/llm_vm/completion/optimize.py:62

↓ 3 callersMethodload_finetune

(self, model_filename=None, small_model=False)

src/llm_vm/client.py:201

↓ 3 callersFunctionload_model_closure

(model_name)

src/llm_vm/onsite_llm.py:62

↓ 3 callersFunctionprepPrintPromptContext

(p)

src/llm_vm/agents/REBEL/utils.py:23

↓ 3 callersFunctionprompt_for_answer

(question: str)

src/llm_vm/agents/FLAT/agent_helper/tool_utils.py:57

↓ 3 callersMethodquery

(self, **kwargs)

src/llm_vm/vector_db.py:32

↓ 2 callersFunction__create_tool_tag

(tool: SingleTool)

src/llm_vm/agents/FLAT/agent_helper/tool_utils.py:19

↓ 2 callersMethod_prefix_state

(self, prefix_str=None, last_token=None)

src/llm_vm/guided_completion.py:193

↓ 2 callersMethod_prefix_state

(self, prefix_str=None, last_token=None)

src/llm_vm/guided_completion.py:327

↓ 2 callersMethodadd_example

(self, c_id, example)

src/llm_vm/completion/optimize.py:74

↓ 2 callersFunctionbuildExampleTools

(tools=GENERIC_TOOLS)

src/llm_vm/agents/REBEL/agent.py:73

↓ 2 callersMethodcreate_index

(self)

src/llm_vm/vector_db.py:12

↓ 2 callersMethoddata_synthesis

This method generates QA pairs using the larger LLM to be used as training data for fine-tuning the smaller LLM. Parameters

src/llm_vm/completion/data_synthesis.py:20

↓ 2 callersFunctiondo_format

(x)

src/llm_vm/utils/print_types.py:10

↓ 2 callersFunctiondo_format

(x)

src/llm_vm/agents/FLAT/agent_helper/utils.py:50

↓ 2 callersFunctiongenerate_convo_history

(memory: TupleList = [], facts: TupleList = [])

src/llm_vm/agents/FLAT/agent_helper/tool_utils.py:43

↓ 2 callersMethodget_ram_usage

()

src/llm_vm/utils/ram.py:23

↓ 2 callersFunctionisOpenAIModel

(str)

src/llm_vm/config.py:47

↓ 2 callersMethodlist_indexes

(self)

src/llm_vm/vector_db.py:16

↓ 2 callersMethodmodel_dtype

(self, dtype)

src/llm_vm/onsite_llm.py:278

↓ 2 callersMethodprint_progress_bar

(self, percentage)

src/llm_vm/utils/ram.py:26

↓ 2 callersFunctionprogress_label

(label: str)

src/llm_vm/agents/FLAT/models/helpers/upload_model.py:22

↓ 2 callersMethodrun

Runs the Agent on the given inputs Parameters ---------- question the user's input question memo

src/llm_vm/agents/REBEL/agent.py:278

↓ 2 callersFunctionset_api_key

Set the API key in the environment variable. Parameters: - key: The API key to set. Defaults to OPENAI_DEFAULT_KEY. There a

src/llm_vm/utils/keys.py:30

↓ 2 callersMethodset_tools

Adds all the available tools when the class is initialized. Parameters ---------- tools a list of availa

src/llm_vm/agents/REBEL/agent.py:224

↓ 2 callersFunctionsplitter_prompt

(elements: List[QuestionSplitModelData])

src/llm_vm/agents/FLAT/agent_helper/tool_utils.py:83

↓ 2 callersFunctionsquared_sum

return 3 rounded square rooted value

eval_finetuned.py:34

↓ 2 callersFunctionsquared_sum

return 3 rounded square rooted value

src/llm_vm/agents/REBEL/agent.py:165

↓ 2 callersMethodstart

(self)

src/llm_vm/utils/ram.py:12

↓ 2 callersMethodupsert

(self, **kwargs)

src/llm_vm/vector_db.py:28

↓ 2 callersFunctionverbose_answer

(data, answer)

src/llm_vm/agents/FLAT/agent_helper/utils.py:25

↓ 1 callersFunction__construct_answer_from_memory_jsonl

(data: List[QuestionSplitModelData])

src/llm_vm/agents/FLAT/models/utils/answer_from_memory_model/get_asm_as_jsonl.py:6

↓ 1 callersFunction__construct_question_split_jsonl

(data: List[QuestionSplitModelData])

src/llm_vm/agents/FLAT/models/utils/question_split_model/get_qs_as_jsonl.py:6

↓ 1 callersFunction__construct_tool_picker_jsonl

(data: List[ToolpickerInputModelData])

src/llm_vm/agents/FLAT/models/utils/tool_picker_model/get_tp_as_jsonl.py:6

↓ 1 callersFunction__format_tool_url

(url: str)

src/llm_vm/agents/FLAT/agent_helper/use_tool.py:22

↓ 1 callersFunction__get_example_tools

()

src/llm_vm/utils/tools.py:22

↓ 1 callersFunction__get_generic_tools

()

src/llm_vm/agents/FLAT/agent_helper/tools.py:18

↓ 1 callersFunction__get_mem_tuple

(json_mem: List[List[Union[str, dict]]])

src/llm_vm/agents/FLAT/models/utils/tool_input_model/tool_input_model_data.py:12

↓ 1 callersFunction__get_pure_key

(key: str)

src/llm_vm/agents/FLAT/agent_helper/replacer.py:13

↓ 1 callersFunction__get_random_tool_subset

(tool_id: int, shuffle_value: int = 0, shuffle_modulo: int = 1e9)

src/llm_vm/agents/FLAT/models/utils/tool_picker_model/tool_picker_model_data.py:15

↓ 1 callersFunction__get_tool_input

( tool: SingleTool, mem: TupleList, question: str, verbose: int, max_tokens: int = 20,

src/llm_vm/agents/FLAT/agent_helper/business_logic.py:20

↓ 1 callersFunction__get_tool_input_jsonl

(data: List[ToolInputModelData])

src/llm_vm/agents/FLAT/models/utils/tool_input_model/get_tool_input_as_jsonl.py:8

↓ 1 callersFunction__get_tool_input_model

(model: OpenAIModel)

src/llm_vm/agents/FLAT/models/utils/tool_input_model/tool_input_model_data.py:6

↓ 1 callersFunction__get_toolpicker_model

(tools_input_model: ToolpickerInputModel)

src/llm_vm/agents/FLAT/models/utils/tool_picker_model/tool_picker_model_data.py:21

↓ 1 callersFunction__is_pure_interpolation

(key: str)

src/llm_vm/agents/FLAT/agent_helper/replacer.py:8

↓ 1 callersFunction__question_splitter_data

()

src/llm_vm/agents/FLAT/models/utils/question_split_model/question_split_model_data.py:5

↓ 1 callersFunction__split_into_subquestions_instructions

(extra: str = "")

src/llm_vm/agents/FLAT/agent_helper/tool_utils.py:31

↓ 1 callersFunction__tokens_to_dollars

(usage)

src/llm_vm/agents/FLAT/agent_helper/requests/call_open_ai.py:5

↓ 1 callersFunction__tools_we_have_access_to

(tools: str = __DEFAULT_TOOLS)

src/llm_vm/agents/FLAT/agent_helper/tool_utils.py:27

↓ 1 callersMethod_get_model_layers

(self)

src/llm_vm/onsite_llm.py:928

↓ 1 callersMethod_get_model_size

(self)

src/llm_vm/onsite_llm.py:931

↓ 1 callersMethod_pdf_loader

(self, pdf_file_path)

src/llm_vm/client.py:217

↓ 1 callersFunctioncalcCost

(p)

src/llm_vm/agents/REBEL/utils.py:39

↓ 1 callersFunctioncall_agent

()

src/llm_vm/agents/agent_interface.py:12

↓ 1 callersFunctioncall_open_ai

(request: LLMCallParams)

src/llm_vm/agents/FLAT/agent_helper/requests/call_open_ai.py:8

↓ 1 callersMethodchange_model_dtype

(self, big_model_dtype=None, small_model_dtype=None)

src/llm_vm/client.py:207

↓ 1 callersFunctioncheck_can_answer_from_memory

(question: str, memory: TupleList = [], facts: TupleList = [])

src/llm_vm/agents/FLAT/agent_helper/bothandler.py:111

↓ 1 callersFunctioncheck_model_status

( job_id: str, label: str )

src/llm_vm/agents/FLAT/models/helpers/check_model_status.py:5

↓ 1 callersMethodchoices_completion

(choices)

src/llm_vm/guided_completion.py:58

↓ 1 callersFunctioncli

This function is the entry point for the project and allows the user to specify an option network address and port number when launching from th

src/llm_vm/server/main.py:40

↓ 1 callersMethodcomplete

(self, stable_context, dynamic_prompt, data_synthesis = False, finetune = False, regex = None, type = None, ch

src/llm_vm/completion/optimize.py:190

↓ 1 callersMethodcomplete_delay_train

Runs a completion using the string stable_context+dynamic_prompt. Returns an optional training closure to use if the caller decides

src/llm_vm/completion/optimize.py:198

↓ 1 callersMethodconstruct_filter_set

(self, expression)

src/llm_vm/guided_completion.py:250

↓ 1 callersMethodconstruct_final_filter_set

(self, prefix_ids, terminals_map)

src/llm_vm/guided_completion.py:355

↓ 1 callersFunctioncos_similarity

return cosine similarity between two lists

src/llm_vm/agents/REBEL/agent.py:171

↓ 1 callersFunctioncreate_jsonl_file

(data_list)

src/llm_vm/onsite_llm.py:75

↓ 1 callersMethoddelete_index

(self, index_name)

src/llm_vm/vector_db.py:24

↓ 1 callersMethoddescribe_index

(self, index_name)

src/llm_vm/vector_db.py:20

↓ 1 callersMethodend

(self)

src/llm_vm/utils/ram.py:17

↓ 1 callersMethodfinetune

(self,data, optimizer, c_id, model_filename=None)

src/llm_vm/onsite_llm.py:170

↓ 1 callersMethodfinetune

(self, dataset, optimizer, c_id, small_model_filename=None)

src/llm_vm/onsite_llm.py:760

↓ 1 callersFunctionflat_main

()

src/llm_vm/agents/FLAT/agent.py:63

↓ 1 callersMethodgenerate_examples

(self, final_prompt, openai_key, example_delim="<END>", model="gpt-4", max_tokens=1000, temperature=1, complet

src/llm_vm/completion/data_synthesis.py:58

↓ 1 callersFunctiongenerate_hash

(input_string)

src/llm_vm/completion/optimize.py:36

↓ 1 callersFunctionget_randomised_training_tools

( generic_tools: ToolList, shuffled_by: int = 0, shuffle_by_modulo: int = 1e9 )

src/llm_vm/agents/FLAT/models/utils/tool_picker_model/get_training_tools.py:4

↓ 1 callersFunctionget_tool_by_id

(tools, tool_id)

src/llm_vm/agents/FLAT/agent_helper/utils.py:103

next →1–100 of 280, ranked by callers