instruction stringclasses 100
values | code stringlengths 78 193k | response stringlengths 259 170k | file stringlengths 59 203 |
|---|---|---|---|
Add standardized docstrings across the file |
from __future__ import annotations
import asyncio
import json
from typing import Any
from sqlalchemy import (
TIMESTAMP,
Column,
ForeignKey,
Index,
Integer,
MetaData,
String,
Table,
Text,
delete,
insert,
select,
text as sql_text,
update,
)
from sqlalchemy.ext.a... | --- +++ @@ -1,3 +1,25 @@+"""SQLAlchemy-powered Session backend.
+
+Usage::
+
+ from agents.extensions.memory import SQLAlchemySession
+
+ # Create from SQLAlchemy URL (uses asyncpg driver under the hood for Postgres)
+ session = SQLAlchemySession.from_url(
+ session_id="user-123",
+ url="postgres... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/extensions/memory/sqlalchemy_session.py |
Add well-formatted docstrings | from __future__ import annotations
import logging
from typing import TYPE_CHECKING, Any, Callable, Literal
from openai import AsyncOpenAI
from ..models._openai_shared import get_default_openai_client
from .openai_conversations_session import OpenAIConversationsSession
from .session import (
OpenAIResponsesCompac... | --- +++ @@ -27,6 +27,10 @@ def select_compaction_candidate_items(
items: list[TResponseInputItem],
) -> list[TResponseInputItem]:
+ """Select compaction candidate items.
+
+ Excludes user messages and compaction items.
+ """
def _is_user_message(item: TResponseInputItem) -> bool:
if not i... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/memory/openai_responses_compaction_session.py |
Turn comments into proper docstrings | from __future__ import annotations
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any
if TYPE_CHECKING:
from .agent import Agent
from .guardrail import InputGuardrailResult, OutputGuardrailResult
from .items import ModelResponse, RunItem, TResponseInputItem
from .run_context impor... | --- +++ @@ -19,6 +19,7 @@
@dataclass
class RunErrorDetails:
+ """Data collected from an agent run when an exception occurs."""
input: str | list[TResponseInputItem]
new_items: list[RunItem]
@@ -33,6 +34,7 @@
class AgentsException(Exception):
+ """Base class for all exceptions in the Agents SDK.... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/exceptions.py |
Provide docstrings following PEP 257 |
from __future__ import annotations
import copy
import dataclasses
import json
from collections import deque
from collections.abc import Callable, Mapping, Sequence
from dataclasses import dataclass, field
from typing import TYPE_CHECKING, Any, Generic, Literal, Optional, Union, cast
from uuid import uuid4
from opena... | --- +++ @@ -1,3 +1,4 @@+"""RunState class for serializing and resuming agent runs with human-in-the-loop support."""
from __future__ import annotations
@@ -136,6 +137,19 @@
@dataclass
class RunState(Generic[TContext, TAgent]):
+ """Serializable snapshot of an agent run, including context, usage, and interrup... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_state.py |
Add docstrings to improve readability |
from __future__ import annotations
import json
import logging
from dataclasses import dataclass, field
from typing import TYPE_CHECKING, Any, cast
from .._tool_identity import get_tool_call_name, get_tool_call_trace_name
if TYPE_CHECKING:
from ..run_config import CallModelData, ModelInputData
logger = logging.... | --- +++ @@ -1,3 +1,29 @@+"""Built-in call_model_input_filter that trims large tool outputs from older turns.
+
+Agentic applications often accumulate large tool outputs (search results, code execution
+output, error analyses) that consume significant tokens but lose relevance as the
+conversation progresses. This modul... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/extensions/tool_output_trimmer.py |
Provide docstrings following PEP 257 | from __future__ import annotations
from dataclasses import dataclass, field, fields
from typing import TYPE_CHECKING, Any, cast
from openai.types.responses import ResponseFunctionToolCall
from ._tool_identity import get_tool_call_namespace, tool_trace_name
from .agent_tool_state import get_agent_tool_state_scope, se... | --- +++ @@ -34,6 +34,7 @@
@dataclass
class ToolContext(RunContextWrapper[TContext]):
+ """The context of a tool call."""
tool_name: str = field(default_factory=_assert_must_pass_tool_name)
"""The name of the tool being invoked."""
@@ -72,6 +73,7 @@ _approvals: dict[str, _ApprovalRecord] | None... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tool_context.py |
Please document this code using docstrings | from __future__ import annotations
import os
from dataclasses import dataclass, field
from typing import TYPE_CHECKING, Any, Callable, Generic, Literal, Optional
from typing_extensions import NotRequired, TypedDict
from .guardrail import InputGuardrail, OutputGuardrail
from .handoffs import HandoffHistoryMapper, Han... | --- +++ @@ -28,12 +28,14 @@
def _default_trace_include_sensitive_data() -> bool:
+ """Return the default for trace_include_sensitive_data based on environment."""
val = os.getenv("OPENAI_AGENTS_TRACE_INCLUDE_SENSITIVE_DATA", "true")
return val.strip().lower() in ("1", "true", "yes", "on")
@datacla... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_config.py |
Improve documentation using docstrings | from __future__ import annotations
from dataclasses import dataclass, field
from typing import TYPE_CHECKING, Any, Generic
from typing_extensions import TypeVar
from ._tool_identity import (
FunctionToolLookupKey,
get_function_tool_approval_keys,
get_function_tool_lookup_key,
is_reserved_synthetic_to... | --- +++ @@ -27,6 +27,11 @@
@dataclass(eq=False)
class _ApprovalRecord:
+ """Tracks approval/rejection state for a tool.
+
+ ``approved`` and ``rejected`` are either booleans (permanent allow/deny)
+ or lists of call IDs when approval is scoped to specific tool calls.
+ """
approved: bool | list[st... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_context.py |
Add docstrings explaining edge cases | from __future__ import annotations
import asyncio
import inspect
from collections.abc import Awaitable, Callable, Mapping, Sequence
from typing import Any, Literal, cast
from openai.types.responses import (
ResponseCompactionItem,
ResponseComputerToolCall,
ResponseCustomToolCall,
ResponseFileSearchToo... | --- +++ @@ -233,6 +233,7 @@ ]
| None = None,
) -> SingleStepResult:
+ """Finalize a turn once final output is known and run end hooks."""
final_output_hooks = run_final_output_hooks_fn or run_final_output_hooks
await final_output_hooks(agent, hooks, context_wrapper, final_output)
@@ -265,6 +266... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/turn_resolution.py |
Create Google-style docstrings for my code | from .config import TracingConfig
from .context import TraceCtxManager
from .create import (
agent_span,
custom_span,
function_span,
generation_span,
get_current_span,
get_current_trace,
guardrail_span,
handoff_span,
mcp_tools_span,
response_span,
speech_group_span,
speec... | --- +++ @@ -83,16 +83,28 @@
def add_trace_processor(span_processor: TracingProcessor) -> None:
+ """
+ Adds a new trace processor. This processor will receive all traces/spans.
+ """
get_trace_provider().register_processor(span_processor)
def set_trace_processors(processors: list[TracingProcessor... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/__init__.py |
Write docstrings for backend logic |
from __future__ import annotations
import asyncio
import inspect
from collections.abc import Mapping
from typing import Any
def _get_awaitable_to_wait_on(awaitable: Any) -> Any | None:
if inspect.iscoroutine(awaitable):
return awaitable.cr_await
if inspect.isgenerator(awaitable):
return awai... | --- +++ @@ -1,3 +1,11 @@+"""Best-effort progress inspection for cancelled function-tool tasks.
+
+These helpers prefer public coroutine introspection first, then fall back to a
+small set of private asyncio attributes for patterns that still hide their
+driving tasks or deadlines (`Task._fut_waiter`, gather `_children`... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/_asyncio_progress.py |
Create docstrings for API functions | from __future__ import annotations
import json
import math
import os
import queue
import random
import threading
import time
from functools import cached_property
from typing import Any
import httpx
from ..logger import logger
from .processor_interface import TracingExporter, TracingProcessor
from .spans import Span... | --- +++ @@ -19,6 +19,7 @@
class ConsoleSpanExporter(TracingExporter):
+ """Prints the traces and spans to the console."""
def export(self, items: list[Trace | Span[Any]]) -> None:
for item in items:
@@ -50,6 +51,19 @@ base_delay: float = 1.0,
max_delay: float = 30.0,
):
+ ... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/processors.py |
Generate consistent documentation across files | from __future__ import annotations
import logging
import os
import threading
import uuid
from abc import ABC, abstractmethod
from datetime import datetime, timezone
from typing import Any
from ..logger import logger
from .config import TracingConfig
from .processor_interface import TracingProcessor
from .scope import... | --- +++ @@ -17,6 +17,7 @@
def _safe_debug(message: str) -> None:
+ """Best-effort debug logging that tolerates closed streams during shutdown."""
def _has_closed_stream_handler(log: logging.Logger) -> bool:
current: logging.Logger | None = log
@@ -41,6 +42,9 @@
class SynchronousMultiTracingPr... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/provider.py |
Write docstrings describing functionality | from __future__ import annotations
from typing import Any
from .config import TracingConfig
from .create import get_current_trace, trace
from .traces import (
Trace,
TraceState,
_hash_tracing_api_key,
_trace_id_was_started,
reattach_trace,
)
def _get_tracing_api_key(tracing: TracingConfig | None... | --- +++ @@ -55,6 +55,7 @@ trace_state: TraceState | None = None,
reattach_resumed_trace: bool = False,
) -> Trace | None:
+ """Return a trace object for this run when one is not already active."""
current_trace = get_current_trace()
if current_trace:
return None
@@ -88,6 +89,7 @@
cla... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/context.py |
Write documentation strings for class attributes |
from __future__ import annotations
from collections.abc import Sequence
from typing import Any
from openai.types.responses import ResponseFunctionToolCall
from ..agent import Agent
from ..items import ItemHelpers, RunItem, ToolApprovalItem, ToolCallOutputItem, TResponseInputItem
from .items import ReasoningItemIdPo... | --- +++ @@ -1,3 +1,8 @@+"""
+Helpers for approval handling within the run loop. Keep only execution-time utilities that
+coordinate approval placeholders and normalization; public APIs should stay in run.py or
+peer modules.
+"""
from __future__ import annotations
@@ -24,6 +29,7 @@ call_id: str | None,
me... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/approvals.py |
Write clean docstrings for readability | from __future__ import annotations
import asyncio
from typing import Any
from ..agent import Agent
from ..exceptions import InputGuardrailTripwireTriggered, OutputGuardrailTripwireTriggered
from ..guardrail import (
InputGuardrail,
InputGuardrailResult,
OutputGuardrail,
OutputGuardrailResult,
)
from .... | --- +++ @@ -59,6 +59,7 @@ streamed_result: RunResultStreaming,
parent_span: Span[Any],
) -> None:
+ """Run guardrails concurrently and stream results into the queue."""
queue = streamed_result._input_guardrail_queue
guardrail_tasks = [
@@ -104,6 +105,7 @@ input: str | list[TResponseInputIte... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/guardrails.py |
Write clean docstrings for readability | import abc
from typing import TYPE_CHECKING, Any
if TYPE_CHECKING:
from .spans import Span
from .traces import Trace
class TracingProcessor(abc.ABC):
@abc.abstractmethod
def on_trace_start(self, trace: "Trace") -> None:
pass
@abc.abstractmethod
def on_trace_end(self, trace: "Trace")... | --- +++ @@ -7,34 +7,136 @@
class TracingProcessor(abc.ABC):
+ """Interface for processing and monitoring traces and spans in the OpenAI Agents system.
+
+ This abstract class defines the interface that all tracing processors must implement.
+ Processors receive notifications when traces and spans start and... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/processor_interface.py |
Write proper docstrings for these functions | from __future__ import annotations
from dataclasses import dataclass
from typing import Any, Callable, Generic, Union
from typing_extensions import TypedDict
from .agent import Agent
from .exceptions import MaxTurnsExceeded
from .items import ModelResponse, RunItem, TResponseInputItem
from .run_context import RunCon... | --- +++ @@ -14,6 +14,7 @@
@dataclass
class RunErrorData:
+ """Snapshot of run data passed to error handlers."""
input: str | list[TResponseInputItem]
new_items: list[RunItem]
@@ -32,6 +33,7 @@
@dataclass
class RunErrorHandlerResult:
+ """Result returned by an error handler."""
final_outpu... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_error_handlers.py |
Write docstrings for backend logic | from __future__ import annotations
from collections.abc import Mapping, Sequence
from typing import TYPE_CHECKING, Any
from ..logger import logger
from .config import TracingConfig
from .setup import get_trace_provider
from .span_data import (
AgentSpanData,
CustomSpanData,
FunctionSpanData,
Generatio... | --- +++ @@ -34,6 +34,30 @@ tracing: TracingConfig | None = None,
disabled: bool = False,
) -> Trace:
+ """
+ Create a new trace. The trace will not be started automatically; you should either use
+ it as a context manager (`with trace(...):`) or call `trace.start()` + `trace.finish()`
+ manually.
... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/create.py |
Document this code for team use |
from __future__ import annotations
from typing import Any, cast
from ..agent import Agent
from ..agent_tool_state import set_agent_tool_state_scope
from ..exceptions import UserError
from ..guardrail import InputGuardrailResult
from ..items import ModelResponse, RunItem, ToolApprovalItem, TResponseInputItem
from ..m... | --- +++ @@ -1,3 +1,4 @@+"""Internal helpers for AgentRunner.run."""
from __future__ import annotations
@@ -53,6 +54,7 @@
def should_cancel_parallel_model_task_on_input_guardrail_trip() -> bool:
+ """Return whether an in-flight model task should be cancelled on guardrail trip."""
try:
from temp... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/agent_runner_helpers.py |
Generate descriptive docstrings automatically |
from __future__ import annotations
import asyncio
import dataclasses
import functools
import inspect
import json
from collections.abc import Awaitable, Callable, Mapping, Sequence
from typing import TYPE_CHECKING, Any, Literal, TypeVar, cast
from openai.types.responses import ResponseFunctionToolCall
from openai.typ... | --- +++ @@ -1,3 +1,7 @@+"""
+Tool execution helpers for the run pipeline. This module hosts execution-time helpers,
+approval plumbing, and payload coercion. Action classes live in tool_actions.py.
+"""
from __future__ import annotations
@@ -158,6 +162,7 @@
@dataclasses.dataclass(frozen=True)
class _FunctionToo... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/tool_execution.py |
Document all endpoints with docstrings | from __future__ import annotations
import ast
import asyncio
import copy
import dataclasses
import inspect
import json
import math
import weakref
from collections.abc import Awaitable, Mapping
from dataclasses import dataclass, field
from types import UnionType
from typing import (
TYPE_CHECKING,
Annotated,
... | --- +++ @@ -85,18 +85,25 @@
class ToolOutputText(BaseModel):
+ """Represents a tool output that should be sent to the model as text."""
type: Literal["text"] = "text"
text: str
class ToolOutputTextDict(TypedDict, total=False):
+ """TypedDict variant for text tool outputs."""
type: Lite... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tool.py |
Replace inline comments with docstrings |
from __future__ import annotations
import asyncio
import dataclasses
import inspect
import json
from typing import TYPE_CHECKING, Any, Literal, cast
from openai.types.responses import ResponseComputerToolCall
from openai.types.responses.response_input_item_param import (
ComputerCallOutputAcknowledgedSafetyCheck... | --- +++ @@ -1,3 +1,7 @@+"""
+Action executors used by the run loop. This module only houses XXXAction classes; helper
+functions and approval plumbing live in tool_execution.py.
+"""
from __future__ import annotations
@@ -67,6 +71,7 @@
def _serialize_trace_payload(payload: Any) -> str:
+ """Serialize tool p... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/tool_actions.py |
Add docstrings following best practices |
from __future__ import annotations
import json
from collections.abc import Sequence
from typing import Any, Literal, cast
from openai.types.responses import ResponseFunctionToolCall
from pydantic import BaseModel
from ..agent_tool_state import drop_agent_tool_run_result
from ..items import ItemHelpers, RunItem, Too... | --- +++ @@ -1,3 +1,7 @@+"""
+Item utilities for the run pipeline. Hosts input normalization helpers and lightweight builders
+for synthetic run items or IDs used during tool execution. Internal use only.
+"""
from __future__ import annotations
@@ -49,6 +53,7 @@
def copy_input_items(value: str | list[TResponseI... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/items.py |
Add docstrings for better understanding | from __future__ import annotations
from dataclasses import dataclass
from typing import Any, Literal, Union
from typing_extensions import TypeAlias
from .agent import Agent
from .items import RunItem, TResponseStreamEvent
@dataclass
class RawResponsesStreamEvent:
data: TResponseStreamEvent
"""The raw resp... | --- +++ @@ -11,6 +11,9 @@
@dataclass
class RawResponsesStreamEvent:
+ """Streaming event from the LLM. These are 'raw' events, i.e. they are directly passed through
+ from the LLM.
+ """
data: TResponseStreamEvent
"""The raw responses streaming event from the LLM."""
@@ -21,6 +24,9 @@
@datacla... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/stream_events.py |
Please document this code using docstrings | from __future__ import annotations
import httpx
from openai import AsyncOpenAI, DefaultAsyncHttpxClient
from ...models import _openai_shared
from ..model import STTModel, TTSModel, VoiceModelProvider
from .openai_stt import OpenAISTTModel
from .openai_tts import OpenAITTSModel
_http_client: httpx.AsyncClient | None ... | --- +++ @@ -25,6 +25,7 @@
class OpenAIVoiceModelProvider(VoiceModelProvider):
+ """A voice model provider that uses OpenAI models."""
def __init__(
self,
@@ -35,6 +36,18 @@ organization: str | None = None,
project: str | None = None,
) -> None:
+ """Create a new OpenA... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/models/openai_model_provider.py |
Document helper functions with docstrings | from __future__ import annotations
import abc
from collections.abc import AsyncIterator
from typing import Any
from ..agent import Agent
from ..items import TResponseInputItem
from ..result import RunResultStreaming
from ..run import Runner
class VoiceWorkflowBase(abc.ABC):
@abc.abstractmethod
def run(self... | --- +++ @@ -11,12 +11,32 @@
class VoiceWorkflowBase(abc.ABC):
+ """
+ A base class for a voice workflow. You must implement the `run` method. A "workflow" is any
+ code you want, that receives a transcription and yields text that will be turned into speech
+ by a text-to-speech model.
+ In most cases... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/workflow.py |
Write reusable docstrings | from __future__ import annotations
import abc
import contextvars
from typing import Any, Generic, TypeVar
from typing_extensions import TypedDict
from ..logger import logger
from . import util
from .processor_interface import TracingProcessor
from .scope import Scope
from .span_data import SpanData
TSpanData = Type... | --- +++ @@ -16,34 +16,104 @@
class SpanError(TypedDict):
+ """Represents an error that occurred during span execution.
+
+ Attributes:
+ message: A human-readable error description
+ data: Optional dictionary containing additional error context
+ """
message: str
data: dict[str, A... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/spans.py |
Improve documentation using docstrings |
from __future__ import annotations
import dataclasses
from dataclasses import dataclass
from typing import Any
from openai.types.responses import ResponseComputerToolCall, ResponseFunctionToolCall
from openai.types.responses.response_output_item import LocalShellCall, McpApprovalRequest
from ..agent import Agent, T... | --- +++ @@ -1,3 +1,7 @@+"""
+Internal step/result data structures used by the run loop orchestration.
+These types are not part of the public SDK surface.
+"""
from __future__ import annotations
@@ -43,6 +47,7 @@
class QueueCompleteSentinel:
+ """Sentinel used to signal completion when streaming run loop re... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/run_steps.py |
Generate docstrings for exported functions |
from __future__ import annotations
import asyncio
import dataclasses as _dc
import json
from collections.abc import Awaitable, Callable, Mapping
from typing import Any, TypeVar, cast
from openai.types.responses import Response, ResponseCompletedEvent, ResponseOutputItemDoneEvent
from openai.types.responses.response_... | --- +++ @@ -1,3 +1,7 @@+"""
+Run-loop orchestration helpers used by the Agent runner. This module coordinates tool execution,
+approvals, and turn processing; all symbols here are internal and not part of the public SDK.
+"""
from __future__ import annotations
@@ -410,6 +414,7 @@ *,
is_resumed_state: bool... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/run_loop.py |
Document helper functions with docstrings | from __future__ import annotations
import asyncio
import base64
import io
import wave
from dataclasses import dataclass
from ..exceptions import UserError
from .imports import np, npt
DEFAULT_SAMPLE_RATE = 24000
def _buffer_to_audio_file(
buffer: npt.NDArray[np.int16 | np.float32 | np.float64],
frame_rate:... | --- +++ @@ -39,6 +39,7 @@
@dataclass
class AudioInput:
+ """Static audio to be used as input for the VoicePipeline."""
buffer: npt.NDArray[np.int16 | np.float32]
"""
@@ -55,9 +56,11 @@ """The number of channels in the audio data. Defaults to 1."""
def to_audio_file(self) -> tuple[str, io.By... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/input.py |
Write docstrings for utility functions |
from __future__ import annotations
from typing import Any, get_args, get_origin
from .._tool_identity import get_function_tool_trace_name
from ..agent import Agent
from ..items import (
HandoffCallItem,
ToolCallItem,
ToolCallItemTypes,
ToolCallOutputItem,
ToolSearchCallItem,
ToolSearchOutputI... | --- +++ @@ -1,3 +1,7 @@+"""
+Tool-use tracking utilities. Hosts AgentToolUseTracker and helpers to serialize/deserialize
+its state plus lightweight tool-call type utilities. Internal use only.
+"""
from __future__ import annotations
@@ -40,6 +44,7 @@
class AgentToolUseTracker:
+ """Track which tools an age... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/tool_use_tracker.py |
Turn comments into proper docstrings | from __future__ import annotations
from collections.abc import Mapping
from dataclasses import field
from typing import Annotated, Any
from openai.types.completion_usage import CompletionTokensDetails, PromptTokensDetails
from openai.types.responses.response_usage import InputTokensDetails, OutputTokensDetails
from p... | --- +++ @@ -11,6 +11,7 @@
def deserialize_usage(usage_data: Mapping[str, Any]) -> Usage:
+ """Rebuild a Usage object from serialized JSON data."""
input_tokens_details_raw = usage_data.get("input_tokens_details")
output_tokens_details_raw = usage_data.get("output_tokens_details")
input_details = _... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/usage.py |
Please document this code using docstrings | from __future__ import annotations
import asyncio
import dataclasses as _dc
import inspect
import json
from collections.abc import Awaitable, Callable, Hashable, Mapping, Sequence
from typing import Any, TypeVar, cast
from openai.types.responses import ResponseFunctionToolCall
from openai.types.responses.response_inp... | --- +++ @@ -64,6 +64,7 @@
def _hashable_identity_value(value: Any) -> Hashable | None:
+ """Convert a tool call field into a stable, hashable representation."""
if value is None:
return None
if isinstance(value, (dict, list, tuple)):
@@ -77,6 +78,7 @@
def _tool_call_identity(raw: Any) -> t... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/tool_planning.py |
Generate docstrings for this script | from __future__ import annotations
import abc
import contextvars
import hashlib
import threading
from collections import OrderedDict
from collections.abc import Mapping
from dataclasses import dataclass, field
from typing import Any
from ..logger import logger
from . import util
from .processor_interface import Traci... | --- +++ @@ -16,6 +16,36 @@
class Trace(abc.ABC):
+ """A complete end-to-end workflow containing related spans and metadata.
+
+ A trace represents a logical workflow or operation (e.g., "Customer Service Query"
+ or "Code Generation") and contains all the spans (individual operations) that occur
+ durin... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/traces.py |
Add docstrings for better understanding | from __future__ import annotations
from dataclasses import dataclass
from typing import Literal, Union
from typing_extensions import TypeAlias
from .imports import np, npt
@dataclass
class VoiceStreamEventAudio:
data: npt.NDArray[np.int16 | np.float32] | None
"""The audio data."""
type: Literal["voic... | --- +++ @@ -10,6 +10,7 @@
@dataclass
class VoiceStreamEventAudio:
+ """Streaming event from the VoicePipeline"""
data: npt.NDArray[np.int16 | np.float32] | None
"""The audio data."""
@@ -20,6 +21,7 @@
@dataclass
class VoiceStreamEventLifecycle:
+ """Streaming event from the VoicePipeline"""
... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/events.py |
Help me add docstrings to my project | import re
from typing import Callable
def get_sentence_based_splitter(
min_sentence_length: int = 20,
) -> Callable[[str], tuple[str, str]]:
def sentence_based_text_splitter(text_buffer: str) -> tuple[str, str]:
sentences = re.split(r"(?<=[.!?])\s+", text_buffer.strip())
if len(sentences) >= ... | --- +++ @@ -5,8 +5,27 @@ def get_sentence_based_splitter(
min_sentence_length: int = 20,
) -> Callable[[str], tuple[str, str]]:
+ """Returns a function that splits text into chunks based on sentence boundaries.
+
+ Args:
+ min_sentence_length: The minimum length of a sentence to be included in a chun... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/utils.py |
Add docstrings explaining edge cases | from __future__ import annotations
import inspect
from collections.abc import Awaitable
from dataclasses import dataclass, field
from typing import TYPE_CHECKING, Any, Callable, Generic, Literal, overload
from typing_extensions import TypedDict, TypeVar
from .exceptions import UserError
from .tool_context import Too... | --- +++ @@ -17,6 +17,7 @@
@dataclass
class ToolInputGuardrailResult:
+ """The result of a tool input guardrail run."""
guardrail: ToolInputGuardrail[Any]
"""The guardrail that was run."""
@@ -27,6 +28,7 @@
@dataclass
class ToolOutputGuardrailResult:
+ """The result of a tool output guardrail run... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tool_guardrails.py |
Generate consistent documentation across files | from __future__ import annotations
import asyncio
import inspect
from typing import Any
from ..agent import Agent
from ..agent_output import AgentOutputSchema, AgentOutputSchemaBase
from ..exceptions import UserError
from ..handoffs import Handoff, handoff
from ..items import TResponseInputItem
from ..lifecycle impor... | --- +++ @@ -30,6 +30,7 @@ def validate_run_hooks(
hooks: RunHooksBase[Any, Agent[Any]] | AgentHooksBase[Any, Agent[Any]] | Any | None,
) -> RunHooks[Any]:
+ """Normalize hooks input and enforce RunHooks type."""
if hooks is None:
return RunHooks[Any]()
input_hook_type = type(hooks).__name__
... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/turn_preparation.py |
Create structured documentation for my script | from __future__ import annotations
import asyncio
from ..exceptions import UserError
from ..logger import logger
from ..tracing import TraceCtxManager
from .input import AudioInput, StreamedAudioInput
from .model import STTModel, TTSModel
from .pipeline_config import VoicePipelineConfig
from .result import StreamedAu... | --- +++ @@ -13,6 +13,11 @@
class VoicePipeline:
+ """An opinionated voice agent pipeline. It works in three steps:
+ 1. Transcribe audio input into text.
+ 2. Run the provided `workflow`, which produces a sequence of text responses.
+ 3. Convert the text responses into streaming audio output.
+ """
... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/pipeline.py |
Write docstrings describing functionality |
from __future__ import annotations
from collections.abc import Sequence
from dataclasses import dataclass, field
from typing import Any, cast
from ..items import (
ItemHelpers,
ModelResponse,
RunItem,
TResponseInputItem,
_output_item_to_input_item,
)
from ..logger import logger
from ..models.fake... | --- +++ @@ -1,3 +1,7 @@+"""
+Conversation-state helpers used during agent runs. This module should only host internal
+tracking and normalization logic for conversation-aware execution, not public-facing APIs.
+"""
from __future__ import annotations
@@ -29,6 +33,7 @@
def _normalize_server_item_id(value: Any) -... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/oai_conversation.py |
Write docstrings for backend logic | from __future__ import annotations
import atexit
import threading
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from .provider import TraceProvider
GLOBAL_TRACE_PROVIDER: TraceProvider | None = None
_GLOBAL_TRACE_PROVIDER_LOCK = threading.Lock()
_SHUTDOWN_HANDLER_REGISTERED = False
def _shutdown_global_tr... | --- +++ @@ -19,6 +19,7 @@
def set_trace_provider(provider: TraceProvider) -> None:
+ """Set the global trace provider used by tracing utilities."""
global GLOBAL_TRACE_PROVIDER
global _SHUTDOWN_HANDLER_REGISTERED
@@ -30,6 +31,11 @@
def get_trace_provider() -> TraceProvider:
+ """Get the global... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/setup.py |
Add docstrings for internal functions | from __future__ import annotations
import abc
from collections.abc import AsyncIterator
from dataclasses import dataclass
from typing import Any, Callable, Literal
from .imports import np, npt
from .input import AudioInput, StreamedAudioInput
from .utils import get_sentence_based_splitter
DEFAULT_TTS_INSTRUCTIONS = ... | --- +++ @@ -20,6 +20,7 @@
@dataclass
class TTSModelSettings:
+ """Settings for a TTS model."""
voice: TTSVoice | None = None
"""
@@ -61,30 +62,47 @@
class TTSModel(abc.ABC):
+ """A text-to-speech model that can convert text into audio output."""
@property
@abc.abstractmethod
de... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/model.py |
Expand my code with proper documentation strings | from __future__ import annotations
import asyncio
import base64
from collections import deque
from collections.abc import AsyncIterator
from typing import Any
from ..exceptions import UserError
from ..logger import logger
from ..tracing import Span, SpeechGroupSpanData, speech_group_span, speech_span
from ..tracing.u... | --- +++ @@ -27,6 +27,7 @@
class StreamedAudioResult:
+ """The output of a `VoicePipeline`. Streams events and audio data as they're generated."""
def __init__(
self,
@@ -34,6 +35,13 @@ tts_settings: TTSModelSettings,
voice_pipeline_config: VoicePipelineConfig,
):
+ ""... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/result.py |
Document this module using docstrings |
from __future__ import annotations
import asyncio
import copy
import inspect
import json
from collections.abc import Sequence
from typing import Any, cast
from ..exceptions import UserError
from ..items import HandoffOutputItem, ItemHelpers, RunItem, ToolCallOutputItem, TResponseInputItem
from ..logger import logger... | --- +++ @@ -1,3 +1,7 @@+"""
+Session persistence helpers for the run pipeline. Only internal persistence/retry helpers
+live here; public session interfaces stay in higher-level modules.
+"""
from __future__ import annotations
@@ -55,6 +59,20 @@ include_history_in_prepared_input: bool = True,
preserve_dro... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/session_persistence.py |
Generate docstrings for this script | from __future__ import annotations
import abc
from collections.abc import Mapping, Sequence
from typing import TYPE_CHECKING, Any
if TYPE_CHECKING:
from openai.types.responses import Response, ResponseInputItemParam
class SpanData(abc.ABC):
@abc.abstractmethod
def export(self) -> dict[str, Any]:
... | --- +++ @@ -9,18 +9,27 @@
class SpanData(abc.ABC):
+ """
+ Represents span data in the trace.
+ """
@abc.abstractmethod
def export(self) -> dict[str, Any]:
+ """Export the span data as a dictionary."""
pass
@property
@abc.abstractmethod
def type(self) -> str:
+ ... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/tracing/span_data.py |
Write docstrings for data processing functions | from __future__ import annotations
import asyncio
from ..items import (
HandoffCallItem,
HandoffOutputItem,
MCPApprovalRequestItem,
MCPApprovalResponseItem,
MCPListToolsItem,
MessageOutputItem,
ReasoningItem,
RunItem,
ToolApprovalItem,
ToolCallItem,
ToolCallOutputItem,
... | --- +++ @@ -28,6 +28,7 @@ new_step_items: list[RunItem],
queue: asyncio.Queue[StreamEvent | QueueCompleteSentinel],
) -> None:
+ """Emit run items as streaming events, skipping approval placeholders."""
for item in new_step_items:
if isinstance(item, MessageOutputItem):
event = R... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/run_internal/streaming.py |
Add verbose docstrings with examples | from collections.abc import AsyncIterator
from typing import Literal
from openai import AsyncOpenAI
from ..model import TTSModel, TTSModelSettings
DEFAULT_VOICE: Literal["ash"] = "ash"
class OpenAITTSModel(TTSModel):
def __init__(
self,
model: str,
openai_client: AsyncOpenAI,
):
... | --- +++ @@ -9,12 +9,19 @@
class OpenAITTSModel(TTSModel):
+ """A text-to-speech model for OpenAI."""
def __init__(
self,
model: str,
openai_client: AsyncOpenAI,
):
+ """Create a new OpenAI text-to-speech model.
+
+ Args:
+ model: The name of the mod... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/models/openai_tts.py |
Add docstrings to improve readability | from __future__ import annotations
import asyncio
import base64
import json
import time
from collections.abc import AsyncIterator
from dataclasses import dataclass
from typing import Any, cast
from openai import AsyncOpenAI
from ... import _debug
from ...exceptions import AgentsException
from ...logger import logger... | --- +++ @@ -52,6 +52,9 @@ async def _wait_for_event(
event_queue: asyncio.Queue[dict[str, Any]], expected_types: list[str], timeout: float
):
+ """
+ Wait for an event from event_queue whose type is in expected_types within the specified timeout.
+ """
start_time = time.time()
while True:
... | https://raw.githubusercontent.com/openai/openai-agents-python/HEAD/src/agents/voice/models/openai_stt.py |
Help me document legacy Python code | # Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -29,6 +29,15 @@
def parquet_opener(data, mode='train'):
+ """ Give url or local file, return file descriptor
+ Inplace operation.
+
+ Args:
+ data(Iterable[str]): url or local file list
+
+ Returns:
+ Iterable[{src, stream}]
+ """
for sample in data:
... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/dataset/processor.py |
Generate documentation strings for clarity | # Copyright (c) 2020 Mobvoi Inc (Binbin Zhang)
# 2024 Alibaba Inc (authors: Xiang Lyu)
# 2025 Alibaba Inc (authors: Xiang Lyu, Bofan Zhou)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a... | --- +++ @@ -14,6 +14,7 @@ # See the License for the specific language governing permissions and
# limitations under the License.
# Modified from ESPnet(https://github.com/espnet/espnet)
+"""Unility functions for Transformer."""
import queue
import random
@@ -53,6 +54,25 @@
def pad_list(xs: List[torch.Tensor],... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/utils/common.py |
Add docstrings for utility scripts |
from __future__ import annotations
from typing import Optional
import math
import torch
from torch import nn
import torch.nn.functional as F
import torchaudio
from x_transformers.x_transformers import apply_rotary_pos_emb
# raw wav to mel spec
class MelSpec(nn.Module):
def __init__(
self,
filt... | --- +++ @@ -1,4 +1,12 @@
+"""
+ein notation:
+b - batch
+n - sequence
+nt - text sequence
+nw - raw wave length
+d - dimension
+"""
from __future__ import annotations
from typing import Optional
@@ -526,6 +534,14 @@
class MMDiTBlock(nn.Module):
+ r"""
+ modified from diffusers/src/diffusers/models/attent... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/flow/DiT/modules.py |
Write Python docstrings for this snippet | # Copyright (c) 2021 Mobvoi Inc. (authors: Binbin Zhang)
# 2024 Alibaba Inc (authors: Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/li... | --- +++ @@ -36,6 +36,9 @@ self.source.set_epoch(epoch)
def __iter__(self):
+ """ Return an iterator over the source dataset processed by the
+ given processor.
+ """
assert self.source is not None
assert callable(self.f)
return self.f(iter(self.source),... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/dataset/dataset.py |
Write docstrings for utility functions | # Copyright (c) 2019 Shigeki Karita
# 2020 Mobvoi Inc (Binbin Zhang)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless ... | --- +++ @@ -12,6 +12,7 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+"""Decoder self-attention layer definition."""
from typing import Optional, Tuple
import torch
@@ -19,6 +20,2... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/transformer/decoder_layer.py |
Create docstrings for API functions | # Copyright (c) 2021 Mobvoi Inc (Binbin Zhang, Di Wu)
# 2022 Xingchen Song (sxc19@mails.tsinghua.edu.cn)
# 2024 Alibaba Inc (Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a co... | --- +++ @@ -14,6 +14,7 @@ # See the License for the specific language governing permissions and
# limitations under the License.
# Modified from ESPnet(https://github.com/espnet/espnet)
+"""Encoder definition."""
from typing import Tuple
import torch
@@ -54,6 +55,37 @@ use_dynamic_left_chunk: bool = False... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/transformer/encoder.py |
Help me write clear docstrings | # Copyright (c) 2020 Mobvoi Inc. (authors: Binbin Zhang, Di Wu)
# 2024 Alibaba Inc (Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/lice... | --- +++ @@ -13,6 +13,7 @@ # See the License for the specific language governing permissions and
# limitations under the License.
# Modified from ESPnet(https://github.com/espnet/espnet)
+"""Positonal Encoding Module."""
import math
from typing import Tuple, Union
@@ -23,12 +24,22 @@
class PositionalEncoding(t... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/transformer/embedding.py |
Write Python docstrings for this snippet | # Copyright (c) 2021 Mobvoi Inc. (authors: Binbin Zhang, Di Wu)
# 2024 Alibaba Inc (Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/lice... | --- +++ @@ -13,6 +13,7 @@ # See the License for the specific language governing permissions and
# limitations under the License.
# Modified from ESPnet(https://github.com/espnet/espnet)
+"""Decoder definition."""
from typing import Tuple, List, Optional
import torch
@@ -30,6 +31,29 @@
class TransformerDecoder... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/transformer/decoder.py |
Add concise docstrings to each method | # Copyright (c) 2019 Shigeki Karita
# 2020 Mobvoi Inc (Binbin Zhang)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless ... | --- +++ @@ -12,18 +12,51 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+"""Label smoothing module."""
import torch
from torch import nn
class LabelSmoothingLoss(nn.Module):
+... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/transformer/label_smoothing_loss.py |
Add docstrings to improve readability | # Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Zhihao Du)
# 2025 Alibaba Inc (authors: Xiang Lyu, Bofan Zhou)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http... | --- +++ @@ -35,6 +35,23 @@
@torch.inference_mode()
def forward(self, mu, mask, n_timesteps, temperature=1.0, spks=None, cond=None, prompt_len=0, cache=torch.zeros(1, 80, 0, 2)):
+ """Forward diffusion
+
+ Args:
+ mu (torch.Tensor): output of encoder
+ shape: (batch_siz... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/flow/flow_matching.py |
Add docstrings to improve collaboration | # Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Zhihao Du)
# 2025 Alibaba Inc (authors: Xiang Lyu, Yabin Li, Qihua, Shengqiang Li)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the Lic... | --- +++ @@ -102,6 +102,13 @@ batch: dict,
device: torch.device,
) -> Dict[str, Optional[torch.Tensor]]:
+ """
+ Args:
+ text: (B, L, D)
+ text_lengths: (B,)
+ audio: (B, T, N) or (B, T)
+ audio_lengths: (B,)
+ """
te... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/llm/llm.py |
Generate helpful docstrings for debugging | # Copyright (c) 2020 Mobvoi Inc. (authors: Binbin Zhang, Di Wu)
# 2024 Alibaba Inc (Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/lice... | --- +++ @@ -13,6 +13,7 @@ # See the License for the specific language governing permissions and
# limitations under the License.
# Modified from ESPnet(https://github.com/espnet/espnet)
+"""ConvolutionModule definition."""
from typing import Tuple
@@ -22,6 +23,7 @@
class ConvolutionModule(nn.Module):
+ ""... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/transformer/convolution.py |
Create documentation strings for testing functions | # Copyright (c) 2019 Shigeki Karita
# 2020 Mobvoi Inc (Binbin Zhang)
# 2024 Alibaba Inc (authors: Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# ... | --- +++ @@ -54,6 +54,31 @@ size: int,
device: torch.device = torch.device("cpu"),
) -> torch.Tensor:
+ """Create mask for subsequent steps (size, size).
+
+ This mask is used only in decoder which works in an auto-regressive mode.
+ This means the current step could only do attention with its... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/utils/mask.py |
Add structured docstrings to improve clarity | # Copyright (c) 2021 Mobvoi Inc (Binbin Zhang, Di Wu)
# 2022 Xingchen Song (sxc19@mails.tsinghua.edu.cn)
# 2024 Alibaba Inc (Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a co... | --- +++ @@ -14,6 +14,7 @@ # See the License for the specific language governing permissions and
# limitations under the License.
# Modified from ESPnet(https://github.com/espnet/espnet)
+"""Encoder definition."""
from typing import Tuple
import torch
@@ -34,6 +35,18 @@
class Upsample1D(nn.Module):
+ """A 1... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/transformer/upsample_encoder.py |
Document this module using docstrings | # Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Kai Hu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applica... | --- +++ @@ -12,6 +12,7 @@ # See the License for the specific language governing permissions and
# limitations under the License.
+"""HIFI-GAN"""
from typing import Dict, Optional, List
import numpy as np
@@ -43,6 +44,7 @@
class ResBlock(torch.nn.Module):
+ """Residual block module in HiFiGAN/BigVGAN."""
... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/hifigan/generator.py |
Add docstrings that explain logic | # Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Zhihao Du)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by appl... | --- +++ @@ -98,6 +98,10 @@ num_heads=4,
act_fn="snake",
):
+ """
+ This decoder requires an input with the same shape of the target. So, if your text content
+ is shorter or longer than the outputs, please re-sampling it before feeding to the decoder.
+ """
sup... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/flow/decoder.py |
Annotate my code with docstrings | # Copyright (c) 2019 Shigeki Karita
# 2020 Mobvoi Inc (Binbin Zhang)
# 2022 Xingchen Song (sxc19@mails.tsinghua.edu.cn)
# 2024 Alibaba Inc (Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the L... | --- +++ @@ -14,6 +14,7 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+"""Multi-Head Attention layer definition."""
import math
from typing import Tuple
@@ -23,12 +24,21 @@
cla... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/transformer/attention.py |
Add docstrings to incomplete code | # Copyright (c) 2020 Mobvoi Inc (Binbin Zhang)
# 2022 Ximalaya Inc (Yuguang Yang)
# 2024 Alibaba Inc (authors: Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the Lice... | --- +++ @@ -25,6 +25,21 @@
class WarmupLR(_LRScheduler):
+ """The WarmupLR scheduler
+
+ This scheduler is almost same as NoamLR Scheduler except for following
+ difference:
+
+ NoamLR:
+ lr = optimizer.lr * model_size ** -0.5
+ * min(step ** -0.5, step * warmup_step ** -1.5)
+ War... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/utils/scheduler.py |
Write proper docstrings for these functions | import torch
import torch.nn as nn
import torch.nn.functional as F
try:
from torch.nn.utils.parametrizations import weight_norm, spectral_norm
except ImportError:
from torch.nn.utils import weight_norm, spectral_norm
from typing import List, Optional, Tuple
from einops import rearrange
from torchaudio.transform... | --- +++ @@ -41,6 +41,15 @@ fft_sizes: Tuple[int, ...] = (2048, 1024, 512),
num_embeddings: Optional[int] = None,
):
+ """
+ Multi-Resolution Discriminator module adapted from https://github.com/descriptinc/descript-audio-codec.
+ Additionally, it allows incorporating condition... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/cosyvoice/hifigan/discriminator.py |
Document functions with clear intent | # Copyright 2022 Xiaomi Corp. (authors: Fangjun Kuang)
# 2023 Nvidia (authors: Yuekai Zhang)
# 2023 Recurrent.ai (authors: Songtao Shi)
# See LICENSE for clarification regarding multiple authors
#
# Licensed under the Apache License, Version 2.0 (the "Lic... | --- +++ @@ -14,6 +14,31 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+"""
+This script supports to load dataset from huggingface and sends it to the server
+for decoding, in parallel... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/client_grpc.py |
Create Google-style docstrings for my code | # Copyright 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of con... | --- +++ @@ -37,8 +37,18 @@
class TritonPythonModel:
+ """Triton Python model for audio tokenization.
+
+ This model takes reference audio input and extracts semantic tokens
+ using s3tokenizer.
+ """
def initialize(self, args):
+ """Initialize the model.
+
+ Args:
+ args:... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo/audio_tokenizer/1/model.py |
Generate helpful docstrings for debugging | # Copyright 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of con... | --- +++ @@ -38,8 +38,18 @@
class TritonPythonModel:
+ """Triton Python model for audio tokenization.
+
+ This model takes reference audio input and extracts semantic tokens
+ using s3tokenizer.
+ """
def initialize(self, args):
+ """Initialize the model.
+
+ Args:
+ args:... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo/speaker_embedding/1/model.py |
Generate consistent documentation across files | import argparse
import json
import os
import time
import asyncio
import torch
import torchaudio
import s3tokenizer
import soundfile as sf
import requests
import httpx
from transformers import AutoTokenizer
from datasets import load_dataset
from torch.utils.data import DataLoader
from functools import partial
from tqdm... | --- +++ @@ -1,3 +1,14 @@+""" Example Usage
+ CUDA_VISIBLE_DEVICES=0 \
+ python3 infer_cosyvoice3_token2wav.py \
+ --output-dir $output_dir \
+ --llm-model-name-or-path $huggingface_model_local_dir \
+ --token2wav-path $token2wav_model_dir \
+ --backend $backend \
+ ... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/infer_cosyvoice3.py |
Add inline docstrings for readability | import torch
import torchaudio
import torchaudio.compliance.kaldi as kaldi
import onnxruntime
import s3tokenizer
import os
import logging
import argparse
import queue
import time
import numpy as np
from functools import partial
from hyperpyyaml import load_hyperpyyaml
from matcha.utils.audio import mel_spectrogram as m... | --- +++ @@ -1,3 +1,7 @@+""" Example Usage
+ CUDA_VISIBLE_DEVICES=0 \
+ python3 token2wav_cosyvoice3.py --enable-trt || exit 1
+"""
import torch
import torchaudio
import torchaudio.compliance.kaldi as kaldi
@@ -300,6 +304,7 @@ def forward_stream(self, generated_speech_tokens, prompt_speech_tokens,
... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/token2wav_cosyvoice3.py |
Add docstrings to incomplete code | # Copyright 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of con... | --- +++ @@ -161,8 +161,18 @@
class TritonPythonModel:
+ """Triton Python model for vocoder.
+
+ This model takes global and semantic tokens as input and generates audio waveforms
+ using the BiCodec vocoder.
+ """
def initialize(self, args):
+ """Initialize the model.
+
+ Args:
+ ... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo/token2wav/1/model.py |
Document functions with clear intent | # Copyright 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of con... | --- +++ @@ -50,6 +50,9 @@
def parse_speech_token_string(response_text: str) -> List[int]:
+ """
+ Parses a string of speech tokens (e.g., "<|s_123|><|s_456|>") into a list of integer IDs.
+ """
speech_tokens = response_text.strip().split('><')
if len(speech_tokens) > 1:
# Add back the mi... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo/cosyvoice2_dit/1/model.py |
Add well-formatted docstrings | # Copyright 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of con... | --- +++ @@ -46,8 +46,18 @@
class TritonPythonModel:
+ """Triton Python model for Spark TTS.
+
+ This model orchestrates the end-to-end TTS pipeline by coordinating
+ between audio tokenizer, LLM, and vocoder components.
+ """
def initialize(self, args):
+ """Initialize the model.
+
+ ... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo/cosyvoice2/1/model.py |
Write docstrings for utility functions | # Copyright 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of con... | --- +++ @@ -54,6 +54,10 @@
def get_spk_id_from_prompt_audio(tensor: torch.Tensor) -> str:
+ """
+ Generates a unique ID for a torch.Tensor.
+ Tensors with the same elements and properties will have the same ID.
+ """
# Convert tensor to a byte string
tensor_bytes = tensor.numpy().tobytes()
@... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo/token2wav_dit/1/model.py |
Auto-generate documentation strings for this file | import json
import re
import time
import asyncio
import numpy as np
import torch
from torch.utils.dlpack import to_dlpack
import triton_python_backend_utils as pb_utils
import httpx
import torchaudio
from functools import partial
from matcha.utils.audio import mel_spectrogram as matcha_mel_spectrogram
torch.set_num... | --- +++ @@ -23,6 +23,7 @@
def parse_speech_token_string(response_text):
+ """Parse speech tokens from string like '<|s_123|><|s_456|>' into list of int IDs."""
speech_tokens = response_text.strip().split('><')
if len(speech_tokens) > 1:
speech_tokens = ['<' + t if not t.startswith('<') else t ... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo_cosyvoice3/cosyvoice3/1/model.py |
Document this module using docstrings | # Copyright 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of con... | --- +++ @@ -37,8 +37,18 @@
class TritonPythonModel:
+ """Triton Python model for audio tokenization.
+
+ This model takes reference audio input and extracts semantic tokens
+ using s3tokenizer.
+ """
def initialize(self, args):
+ """Initialize the model.
+
+ Args:
+ args:... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo_cosyvoice3/audio_tokenizer/1/model.py |
Add docstrings following best practices | # SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved.
# SPDX-License-Identifier: Apache-2.0
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://ww... | --- +++ @@ -12,6 +12,10 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+""" Example Usage
+ CUDA_VISIBLE_DEVICES=0 \
+ python3 token2wav.py --enable-trt || exit 1
+"""
impor... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo/token2wav_dit/1/token2wav_dit.py |
Add docstrings following best practices | # Copyright 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# * Redistributions of source code must retain the above copyright
# notice, this list of con... | --- +++ @@ -38,8 +38,18 @@
class TritonPythonModel:
+ """Triton Python model for audio tokenization.
+
+ This model takes reference audio input and extracts semantic tokens
+ using s3tokenizer.
+ """
def initialize(self, args):
+ """Initialize the model.
+
+ Args:
+ args:... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/model_repo_cosyvoice3/speaker_embedding/1/model.py |
Add professional docstrings to my codebase | #!/usr/bin/env python3
# Copyright 2025 CosyVoice3 TRT-LLM Integration
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required b... | --- +++ @@ -12,6 +12,24 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+"""
+Convert CosyVoice3 LLM to HuggingFace format with merged embeddings.
+
+This script:
+1. Loads CosyVoice3 m... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/scripts/convert_cosyvoice3_to_hf.py |
Provide clean and structured docstrings | # SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved.
# SPDX-License-Identifier: Apache-2.0
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://ww... | --- +++ @@ -12,6 +12,17 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+""" Example Usage
+ CUDA_VISIBLE_DEVICES=0 \
+ python3 offline_inference.py \
+ --output-di... | https://raw.githubusercontent.com/FunAudioLLM/CosyVoice/HEAD/runtime/triton_trtllm/offline_inference.py |
Generate docstrings for exported functions | # Copyright 2024 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -11,6 +11,9 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+"""Diagnose script for checking OS/hardware/python/pip/verl/network.
+The output of this script can be a very goo... | https://raw.githubusercontent.com/verl-project/verl/HEAD/scripts/diagnose.py |
Add structured docstrings to improve clarity | # Copyright 2024 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -47,6 +47,9 @@
class NixlAgent:
+ """This is a wrapper class for nixl_agent, the main purpose is to use ZeroMQ instead of
+ `nixl_agent.send_notif` to send bucket tensor metadata.
+ """
def __init__(self):
self.agent_name = str(uuid.uuid4())
@@ -132,6 +135,17 @@
class Readabl... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/checkpoint_engine/nixl_checkpoint_engine.py |
Fill in missing docstrings in my code | # Copyright 2024 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -41,6 +41,16 @@
class BroadcastOperation:
+ """Async broadcast operation with NCCL in separate thread.
+
+ Args:
+ rank (int): The rank of the current process.
+ group_name (str): The name of the NCCL process group.
+ bucket (cp.ndarray | torch.Tensor): The tensor to broadcast.... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/checkpoint_engine/nccl_checkpoint_engine.py |
Add inline docstrings for readability | # Copyright 2025 Meituan Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or... | --- +++ @@ -25,6 +25,9 @@
@ray.remote(num_cpus=2, max_concurrency=20)
class MessageQueue:
+ """
+ Simplified Ray-based asynchronous message queue for communication between Rollouter and Trainer
+ """
def __init__(self, config: DictConfig, max_queue_size: int = 1000):
self.config = config
@@ ... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/fully_async_policy/message_queue.py |
Write documentation strings for class attributes | # Copyright 2025 Meituan Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or... | --- +++ @@ -26,6 +26,7 @@
@dataclass
class RolloutSample:
+ """Enhanced rollout sample containing both original batch info and AgentLoopOutput"""
# Original batch information
full_batch: Any
@@ -40,12 +41,20 @@
@dataclass
class ValidateMetrics:
+ """Metrics for validation"""
timing_raw: d... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/fully_async_policy/detach_utils.py |
Add docstrings explaining edge cases | # Copyright 2025 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -115,6 +115,20 @@ x_t: torch.Tensor,
timestep: torch.Tensor,
) -> Tensor:
+ """Full forward pass for one diffusion denoising step.
+
+ Args:
+ images: List of image tensors, each shaped (B, C, H, W) after batching.
+ img_masks: List of boolean masks c... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/vla/models/pi0_torch/modeling_pi0_torch.py |
Provide docstrings following PEP 257 | # Copyright 2024 Bytedance Ltd. and/or its affiliates
# Copyright 2023-2024 SGLang Team
# Copyright 2025 ModelBest Inc. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#... | --- +++ @@ -13,6 +13,10 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+"""
+PPO Trainer with Ray-based single controller.
+This trainer supports model-agonistic model initialization w... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/vla/rob_ray_trainer.py |
Write docstrings that follow conventions | # Copyright 2025 Bytedance Ltd. and/or its affiliates
# Copyright 2025 Giga Team. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/... | --- +++ @@ -109,6 +109,17 @@ output_hidden_states: Optional[bool] = None,
interpolate_pos_encoding: Optional[bool] = False,
) -> BaseModelOutputWithPooling:
+ """Forward pass of the SigLIP vision encoder.
+
+ Args:
+ pixel_values: Image tensor expected by SigLIP (B, C, H, ... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/vla/models/pi0_torch/model/paligemma_with_expert.py |
Add docstrings to my Python code | # Copyright 2024 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -25,6 +25,9 @@
@ray.remote(num_cpus=1)
class RewardComputeWorker:
+ """
+ WARNING: This class cannot have async methods.
+ """
def __init__(self, compute_score_fn):
# since the reward function may not be pickleable, we need to init it in the worker
@@ -36,6 +39,13 @@
@register("... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/reward_loop/reward_manager/remote.py |
Document helper functions with docstrings | # Copyright 2025 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law... | --- +++ @@ -15,6 +15,12 @@ # from https://github.com/PRIME-RL/SimpleVLA-RL/blob/main/verl/utils/vla_utils/openvla_oft/
# form https://huggingface.co/Haozhan72/Openvla-oft-SFT-libero10-trajall/blob/main/
+"""
+processing_prismatic.py
+
+HuggingFace-style preprocessor definitions for Prismatic VLMs, inheriting from `P... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/vla/models/openvla_oft/processing_prismatic.py |
Include argument descriptions in docstrings | # Copyright 2024 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -36,6 +36,9 @@
def migrate_legacy_reward_impl(config):
+ """
+ Migrate the legacy reward model implementation to the new one.
+ """
# 1. reward workers migration
# config.reward_model.num_workers -> config.reward.num_workers
if config.reward_model.num_workers is not None:
@@ -87,8... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/reward_loop/reward_loop.py |
Add docstrings to incomplete code | # Copyright 2025 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -50,6 +50,8 @@
class AgentData:
+ """Encapsulates all state variables for the agent loop. AgentData is passed to tool calling in case that
+ tool may need to access full history state. User can store any tool session data in `extra_fields`."""
def __init__(
self,
@@ -199,6 +201,7 @@... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/agent_loop/tool_agent_loop.py |
Write docstrings for algorithm functions | # Copyright 2024 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -11,6 +11,9 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
+"""
+The main entry point to run the PPO algorithm
+"""
import asyncio
import contextlib
@@ -47,6 +50,10 @@
... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/vla/fsdp_workers.py |
Add professional docstrings to my codebase | # Copyright 2025 Bytedance Ltd. and/or its affiliates
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law o... | --- +++ @@ -20,28 +20,67 @@ # BaseConfig class inherits from collections.abc.Mapping, which means it can act like a dictionary
@dataclass
class BaseConfig(collections.abc.Mapping):
+ """The BaseConfig provides dict-like interface for a dataclass config.
+
+ By default all fields in the config is not mutable, un... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/base_config.py |
Write documentation strings for class attributes | # Copyright 2024 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -43,6 +43,16 @@
class BroadcastOperation:
+ """Async broadcast operation with HCCL in separate thread.
+
+ Args:
+ rank (int): The rank of the current process.
+ group_name (str): The name of the HCCL process group.
+ bucket (torch.Tensor): The tensor to broadcast.
+ met... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/checkpoint_engine/hccl_checkpoint_engine.py |
Provide clean and structured docstrings | # Copyright 2025 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law ... | --- +++ @@ -28,8 +28,18 @@
class EnvLoop:
+ """An env loop manages interactions between models and vectorized environments. It's designed for computationally
+ intensive environments, such as robotics simulators."""
def __init__(self, env_wg: RayWorkerGroup, rollout_wg: RayWorkerGroup, config: DictConf... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/vla/env_loop.py |
Help me add docstrings to my project | # Copyright 2025 Meituan Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or... | --- +++ @@ -41,6 +41,11 @@
@ray.remote(num_cpus=10, max_concurrency=100)
class FullyAsyncRollouter(SeparateRayPPOTrainer):
+ """
+ Asynchronous sample generator, responsible for continuously generating training samples
+ and putting them into MessageQueue
+ Based on the mature implementation improvements... | https://raw.githubusercontent.com/verl-project/verl/HEAD/verl/experimental/fully_async_policy/fully_async_rollouter.py |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.