MCPcopy Create free account

hub / github.com/FastMAS/KVCOMM / functions

Functions446 in github.com/FastMAS/KVCOMM

⨍ Functions 446◇ Types & classes 71
↓ 1 callersFunctionget_output
(func: str, assert_statement: str, timeout: int = 5)
KVCOMM/tools/coding/python_executor.py:21
↓ 1 callersFunctionget_pred_value
(pred_str)
datasets/gsm8k_dataset.py:72
↓ 1 callersFunctiongsm_data_process
(dataset)
datasets/gsm8k_dataset.py:9
↓ 1 callersMethodhas_active_anchor
Determine whether an anchor should trigger dense prefill.
KVCOMM/llm/gpt_chat.py:557
↓ 1 callersMethodinit_nodes
Creates and adds new nodes to the graph.
KVCOMM/graph/graph.py:136
↓ 1 callersMethodinit_potential_edges
Creates and potential edges to the graph.
KVCOMM/graph/graph.py:147
↓ 1 callersMethodkeys
(cls)
KVCOMM/prompt/prompt_set_registry.py:14
↓ 1 callersFunctionload_result
(result_file: Path)
experiments/run_humaneval.py:33
↓ 1 callersFunctionload_result
(result_file: Path)
experiments/run_gsm8k.py:36
↓ 1 callersMethodlocate_placeholder
Locate placeholder token spans in a templated prompt. Args: original_text: Templated prompt with placeholders such as
KVCOMM/llm/gpt_chat.py:945
↓ 1 callersFunctionmain
()
KVCOMM/utils/log.py:76
↓ 1 callersFunctionmain
()
experiments/run_humaneval.py:94
↓ 1 callersFunctionmain
()
experiments/run_mmlu.py:78
↓ 1 callersFunctionmain
()
experiments/benchmark_TTFT.py:144
↓ 1 callersFunctionmain
()
experiments/run_gsm8k.py:102
↓ 1 callersMethodoffset_kv_cache_pair
Blend base caches with anchor deltas weighted by similarity.
KVCOMM/llm/kvcomm_engine.py:868
↓ 1 callersFunctionparse_args
()
experiments/run_humaneval.py:46
↓ 1 callersFunctionparse_args
()
experiments/run_mmlu.py:31
↓ 1 callersFunctionparse_args
()
experiments/benchmark_TTFT.py:31
↓ 1 callersFunctionparse_args
()
experiments/run_gsm8k.py:49
↓ 1 callersMethodpostprocess_answer
Normalise the model output into a plain string answer.
datasets/mmlu_dataset.py:71
↓ 1 callersFunctionprocess_tasks
(prompts, api_key)
KVCOMM/tools/vgen/dalle3.py:8
↓ 1 callersMethodrecord_to_target_answer
(record: pd.DataFrame)
datasets/mmlu_dataset.py:84
↓ 1 callersFunctionreplace
Returns a new TokenInput where the slice [start, end) is replaced by `real`.
KVCOMM/llm/token_ops.py:157
↓ 1 callersMethodrotate_tensor
Apply RoPE rotation using provided cos/sin tables.
KVCOMM/llm/kvcomm_engine.py:1178
↓ 1 callersMethodsearch
(self, query: str, item_num: int = 3)
KVCOMM/tools/search/search.py:26
↓ 1 callersFunctionsearch_wiki_api
(query, session)
KVCOMM/tools/search/wiki.py:65
↓ 1 callersMethodset_anchor
Populate or update anchor store with per-placeholder deltas for a message.
KVCOMM/llm/kvcomm_engine.py:1082
↓ 1 callersMethodset_reader
(self, suffix)
KVCOMM/tools/reader/readers.py:307
↓ 1 callersFunctionswarmlog
Custom log function for swarm operations. Includes dynamic global variables. Args: sender (str): The name of the sender. tex
KVCOMM/utils/log.py:49
↓ 1 callersMethodupdate_anchor
Update the anchor list by filtering out the least frequent anchors in the oldest anchor set.
KVCOMM/llm/kvcomm_engine.py:1062
↓ 1 callersMethodupdate_memory_multirequest
Update per-message memories when running in multi-request mode.
KVCOMM/graph/node.py:117
FunctionYoutube
(url, has_subtitles)
KVCOMM/tools/web/youtube.py:5
Method__call__
(self, input_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs: Any)
KVCOMM/llm/gpt_chat.py:98
Method__contains__
(self, key: object)
KVCOMM/llm/kvcomm_engine.py:629
Method__deepcopy__
(self, memo)
KVCOMM/llm/gpt_chat.py:1813
Method__deepcopy__
(self, memo)
KVCOMM/graph/graph.py:455
Method__delitem__
(self, key: str)
KVCOMM/llm/kvcomm_engine.py:568
Method__getitem__
(self, key: str)
KVCOMM/llm/kvcomm_engine.py:551
Method__getitem__
(self, index: int)
datasets/mmlu_dataset.py:54
Method__init__
(self)
KVCOMM/tools/search/wiki.py:8
Method__init__
(self)
KVCOMM/tools/search/arXiv.py:4
Method__init__
(self)
KVCOMM/tools/search/search.py:10
Method__init__
(self)
KVCOMM/tools/reader/readers.py:323
Method__init__
(self)
KVCOMM/utils/metrics.py:25
Method__init__
(self)
KVCOMM/utils/globals.py:14
Method__init__
(self)
KVCOMM/utils/globals.py:18
Method__init__
(self)
KVCOMM/utils/globals.py:22
Method__init__
(self)
KVCOMM/utils/globals.py:26
Method__init__
(self)
KVCOMM/utils/globals.py:30
Method__init__
(self, base: Dict[str, Any])
KVCOMM/llm/kvcomm_engine.py:541
Method__init__
( self, request_uid: str, anchor_dict: Dict[str, Any], anchor_len_dict: Dict[s
KVCOMM/llm/kvcomm_engine.py:648
Method__init__
(self, llm: "LLMChat")
KVCOMM/llm/kvcomm_engine.py:690
Method__init__
(self, prompt_length: int)
KVCOMM/llm/gpt_chat.py:88
Method__init__
(self, model_name: str)
KVCOMM/llm/gpt_chat.py:129
Method__init__
Create a chat model instance and initialize shared resources. Args: model_name: HF model identifier. prefix: Optional
KVCOMM/llm/gpt_chat.py:206
Method__init__
Initializes a new Node instance.
KVCOMM/graph/node.py:44
Method__init__
(self, domain: str, llm_name: Optional[str], agent_names: List
KVCOMM/graph/graph.py:45
Method__init__
( self, id: str | None = None, role: str = None, domain: str = "", llm
KVCOMM/agents/copy_machine.py:22
Method__init__
( self, id: str | None = None, domain: str = "", llm_name: str = "", l
KVCOMM/agents/final_decision.py:15
Method__init__
Used for Directed IO
KVCOMM/agents/final_decision.py:612
Method__init__
Used for Directed IO
KVCOMM/agents/final_decision.py:649
Method__init__
( self, id: str | None = None, role: str = None, domain: str = "", llm
KVCOMM/agents/code_writing.py:15
Method__init__
( self, id: str | None = None, role: str = None, domain: str = "", llm
KVCOMM/agents/analyze_agent.py:23
Method__init__
( self, id: str | None = None, role: str = None, domain: str = "", llm
KVCOMM/agents/math_solver.py:14
Method__init__
(self)
experiments/accuracy.py:4
Method__init__
(self, split: Union[Literal['dev'], Literal['val'], Literal['test']], )
datasets/mmlu_dataset.py:10
Method__iter__
(self)
KVCOMM/llm/kvcomm_engine.py:571
Method__len__
(self)
KVCOMM/llm/kvcomm_engine.py:574
Method__len__
(self)
datasets/mmlu_dataset.py:51
Method__setitem__
(self, key: str, value: Any)
KVCOMM/llm/kvcomm_engine.py:565
Method_async_execute
Handle asynchronous execution across different KV cache modes.
KVCOMM/agents/copy_machine.py:137
Method_async_execute
To be overriden by the descendant class
KVCOMM/agents/final_decision.py:289
Method_async_execute
Handle asynchronous execution across different KV cache strategies.
KVCOMM/agents/final_decision.py:558
Method_async_execute
To be overriden by the descendant class
KVCOMM/agents/final_decision.py:633
Method_async_execute
To be overriden by the descendant class
KVCOMM/agents/final_decision.py:676
Method_async_execute
To be overriden by the descendant class
KVCOMM/agents/code_writing.py:190
Method_async_execute
Handle asynchronous execution across different KV cache modes.
KVCOMM/agents/analyze_agent.py:182
Method_async_execute
To be overriden by the descendant class
KVCOMM/agents/math_solver.py:142
Method_execute
To be overriden by the descendant class
KVCOMM/agents/copy_machine.py:118
Method_execute
To be overriden by the descendant class
KVCOMM/agents/final_decision.py:270
Method_execute
To be overriden by the descendant class
KVCOMM/agents/final_decision.py:541
Method_execute
To be overriden by the descendant class
KVCOMM/agents/final_decision.py:622
Method_execute
To be overriden by the descendant class
KVCOMM/agents/final_decision.py:659
Method_execute
To be overriden by the descendant class
KVCOMM/agents/code_writing.py:171
Method_execute
To be overriden by the descendant class
KVCOMM/agents/analyze_agent.py:163
Method_execute
To be overriden by the descendant class
KVCOMM/agents/math_solver.py:124
Method_get_slot
(bucket_key: str)
KVCOMM/llm/kvcomm_engine.py:1238
Method_make_anchor
(i, ph_id, real_ph, base_ph, real_pf, base_pf)
KVCOMM/llm/kvcomm_engine.py:1109
Method_process_inputs
To be overriden by the descendant class
KVCOMM/agents/final_decision.py:617
Method_process_inputs
To be overriden by the descendant class
KVCOMM/agents/final_decision.py:654
Function_runner
(q)
KVCOMM/tools/coding/python_executor.py:45
Method_should_materialise
(ph_id: str)
KVCOMM/llm/kvcomm_engine.py:1106
Method_summarize_agent_outputs
Summarize peer outputs, running tests on code blocks when present.
KVCOMM/agents/final_decision.py:66
Methodadd_predecessor
(self, operation: 'Node', st='spatial')
KVCOMM/graph/node.py:73
Methodagen
Asynchronously generate a response via hosted chat API.
KVCOMM/llm/gpt_chat.py:132
Methodagen
( self, messages: List[Message] = None, max_tokens: Optional[int] = None, temp
KVCOMM/llm/gpt_chat.py:1019
Methodbegin_of_text
(self)
KVCOMM/llm/gpt_chat.py:309
Methodbuild_prompt
Create model inputs from system/user prompts and optional assistant suffix.
KVCOMM/llm/gpt_chat.py:499
Functionbytes_to_data_url
(image_bytes: bytes, mime_type: str)
KVCOMM/tools/web/screenshot.py:11
← previousnext →201–300 of 446, ranked by callers