d9d 0.1.0__py3-none-any.whl

d9d/tracker/provider/aim/tracker.py

@@ -0,0 +1,114 @@
+from collections.abc import Generator
+from contextlib import contextmanager
+from typing import Any, Self, TypedDict, cast
+
+import torch
+from aim import Distribution, Run
+
+from d9d.tracker import BaseTracker, BaseTrackerRun, RunConfig
+
+from .config import AimConfig
+
+
+class AimState(TypedDict):
+    """
+    State dictionary format for persisting Aim tracker state.
+    """
+
+    restart_hash: str | None
+
+
+class AimRun(BaseTrackerRun):
+    """
+    Active run implementation for Aim.
+
+    Wraps the underlying `aim.Run` object to adhere to the d9d BaseTrackerRun interface.
+    """
+
+    def __init__(self, run: Run):
+        self._run = run
+        self._step = 0
+        self._context: dict[str, str] = {}
+
+    def set_step(self, step: int):
+        self._step = step
+
+    def set_context(self, context: dict[str, str]):
+        self._context = context
+
+    def scalar(self, name: str, value: float, context: dict[str, str] | None = None):
+        if context is None:
+            track_context = self._context
+        else:
+            track_context = {**self._context, **context}
+
+        self._run.track(
+            name=name,
+            value=value,
+            context=track_context,
+            step=self._step
+        )
+
+    def bins(self, name: str, values: torch.Tensor, context: dict[str, str] | None = None):
+        if context is None:
+            track_context = self._context
+        else:
+            track_context = {**self._context, **context}
+
+        self._run.track(
+            name=name,
+            value=Distribution(
+                hist=values.numpy(),
+                bin_range=(0, values.shape[0])
+            ),
+            context=track_context,
+            step=self._step
+        )
+
+
+class AimTracker(BaseTracker[AimConfig]):
+    """
+    Aim-based tracker implementation.
+
+    Caches the run hash to allow experiment resumption from checkpoints.
+    """
+
+    def __init__(self, config: AimConfig):
+        self._config = config
+
+        self._restart_hash: str | None = None
+        self._run: Run | None = None
+
+    def load_state_dict(self, state_dict: dict[str, Any]) -> None:
+        state = cast(AimState, state_dict)
+        self._restart_hash = state["restart_hash"]
+
+    def state_dict(self) -> dict[str, Any]:
+        return {
+            "restart_hash": self._restart_hash
+        }
+
+    @contextmanager
+    def open(self, properties: RunConfig) -> Generator[BaseTrackerRun, None, None]:
+        run = Run(
+            run_hash=self._restart_hash,
+            repo=self._config.repo,
+            log_system_params=self._config.log_system_params,
+            capture_terminal_logs=self._config.capture_terminal_logs,
+            system_tracking_interval=self._config.system_tracking_interval
+        )
+        run.name = properties.name
+        run.description = properties.description
+        run["hparams"] = properties.hparams
+
+        self._restart_hash = run.hash
+        self._run = run
+
+        yield AimRun(run)
+
+        self._run.close()
+        self._run = None
+
+    @classmethod
+    def from_config(cls, config: AimConfig) -> Self:
+        return cls(config)

d9d/tracker/provider/null.py

@@ -0,0 +1,61 @@
+from collections.abc import Generator
+from contextlib import contextmanager
+from typing import Any, Literal, Self
+
+import torch
+from pydantic import BaseModel
+
+from d9d.tracker import BaseTracker, BaseTrackerRun, RunConfig
+
+
+class NullTrackerConfig(BaseModel):
+    """
+    Configuration for the Null (no-op) tracker.
+
+    Attributes:
+        provider: Discriminator field, must be 'null'.
+    """
+
+    provider: Literal["null"] = "null"
+
+
+class NullRun(BaseTrackerRun):
+    """
+    No-op implementation of a tracking run.
+
+    Discard all inputs; useful for testing or when tracking is disabled.
+    """
+
+    def set_step(self, step: int):
+        pass
+
+    def set_context(self, context: dict[str, str]):
+        pass
+
+    def scalar(self, name: str, value: float, context: dict[str, str] | None = None):
+        pass
+
+    def bins(self, name: str, values: torch.Tensor, context: dict[str, str] | None = None):
+        pass
+
+
+class NullTracker(BaseTracker[NullTrackerConfig]):
+    """
+    No-op tracker factory.
+
+    Does not modify state or perform any IO.
+    """
+
+    @contextmanager
+    def open(self, properties: RunConfig) -> Generator[BaseTrackerRun, None, None]:
+        yield NullRun()
+
+    @classmethod
+    def from_config(cls, config: NullTrackerConfig) -> Self:
+        return cls()
+
+    def state_dict(self) -> dict[str, Any]:
+        return {}
+
+    def load_state_dict(self, state_dict: dict[str, Any]) -> None:
+        pass

d9d-0.1.0.dist-info/METADATA

@@ -0,0 +1,90 @@
+Metadata-Version: 2.4
+Name: d9d
+Version: 0.1.0
+Summary: d9d - d[istribute]d - distributed training framework based on PyTorch that tries to be efficient yet hackable
+License: Apache-2.0
+Author: Maksim Afanasyev
+Author-email: mr.applexz@gmail.com
+Requires-Python: >=3.11,<3.15
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Education
+Classifier: Intended Audience :: Science/Research
+Classifier: Topic :: Scientific/Engineering
+Classifier: Topic :: Scientific/Engineering :: Mathematics
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Provides-Extra: aim
+Provides-Extra: cce
+Provides-Extra: moe
+Provides-Extra: visualization
+Requires-Dist: aim (>=3.0.0,<4.0.0) ; extra == "aim"
+Requires-Dist: cut-cross-entropy (>=25.9.3) ; extra == "cce"
+Requires-Dist: deep-ep (>=1.2.1) ; extra == "moe"
+Requires-Dist: nv-grouped-gemm (>=1.1.4) ; extra == "moe"
+Requires-Dist: plotly (>=6.0.0) ; extra == "visualization"
+Requires-Dist: pydantic (>=2.0.0)
+Requires-Dist: safetensors (>=0.7.0)
+Requires-Dist: setuptools (>=70.0.0) ; extra == "aim"
+Requires-Dist: torch (>=2.10.0)
+Requires-Dist: torchdata (>=0.11.0)
+Requires-Dist: tqdm (>=4.0.0)
+Requires-Dist: triton (>=3.6.0)
+Project-URL: Documentation, https://d9d-project.github.io/d9d
+Project-URL: Homepage, https://d9d-project.github.io/d9d
+Project-URL: Issues, https://github.com/d9d-project/d9d/issues
+Project-URL: Repository, https://github.com/d9d-project/d9d
+Description-Content-Type: text/markdown
+
+# The d9d Project
+
+**d9d** is a distributed training framework built on top of PyTorch 2.0. It aims to be hackable, modular, and efficient, designed to scale from single-GPU debugging to massive clusters running 6D-Parallelism.
+
+[LET'S START TRAINING 🚀](https://d9d-project.github.io/d9d/)
+
+## Why another framework?
+
+Distributed training frameworks such as **Megatron-LM** are monolithic in the way you run a script from the command line to train any of a set of *predefined* models, using *predefined* regimes. While powerful, these systems can be difficult to hack and integrate into novel research workflows. Their focus is often on providing a complete, end-to-end solution, which can limit flexibility for experimentally-driven research.
+
+Conversely, creating your own distributed training solution from scratch is tricky. You have to implement many low-level components (like distributed checkpoints and synchronization) that are identical across setups, and manually tackle common performance bottlenecks.
+
+**d9d** was designed to fill the gap between monolithic frameworks and homebrew setups, providing a modular yet effective solution for distributed training.
+
+## What d9d is and isn't
+
+In terms of **core concept**:
+
+*   **IS** a pluggable framework for implementing distributed training regimes for your deep learning models.
+*   **IS** built on clear interfaces and building blocks that may be composed and implemented in your own way.
+*   **IS NOT** an all-in-one CLI platform for setting up pre-training and post-training like **torchtitan**, **Megatron-LM**, or **torchforge**.
+
+In terms of **codebase & engineering**:
+
+*   **IS** built on a **strong engineering foundation**: We enforce strict type-checking and rigorous linting to catch errors before execution.
+*   **IS** reliable: The framework is backed by a suite of **over 450 tests**, covering unit logic, integration flows, and End-to-End distributed scenarios.
+*   **IS** eager to use performance hacks (like **DeepEp** or custom kernels) if they improve MFU, even if they aren't PyTorch-native.
+*   **IS NOT** for legacy setups: We do not maintain backward compatibility with older PyTorch versions or hardware. We prioritize simplicity and modern APIs (like `DTensor`).
+
+## Key Philosophies
+
+To achieve the balance between hackability and performance, d9d adheres to specific design principles:
+
+*   **Composition over Monoliths**: We avoid "God Classes" like `DistributedDataParallel` or `ParallelDims` that assume ownership of the entire execution loop. Instead, we provide composable and extendable APIs. For instance, specific horizontal parallelism strategies for specific layers (`parallelize_replicate`, `parallelize_expert_parallel`, ...).
+*   **White-Box Modelling**: We encourage standard PyTorch code. Models are not wrapped in obscure metadata specifications; they are standard `nn.Module`s that implement lightweight protocols.
+*   **Pragmatic Efficiency**: While we prefer native PyTorch, we are eager to integrate non-native solutions if they improve MFU. For example, we implement MoE using **DeepEp** communications, reindexing kernels from **Megatron-LM**, and efficient grouped-GEMM implementations.
+*   **Graph-Based State Management**: Our IO system treats model checkpoints as directed acyclic graphs. This allows you to transform architectures (e.g., merging `q`, `k`, `v` into `qkv`) on-the-fly while streaming from disk, without massive memory overhead.
+*   **DTensors**: We mandate that distributed parameters be represented as `torch.distributed.tensor.DTensor`. This simplifies checkpointing by making them topology-aware automatically. We leverage modern PyTorch 2.0 APIs (`DeviceMesh`) as much as possible.
+
+---
+
+## Examples
+
+### Qwen3-MoE Pretraining
+An example showing causal LM pretraing for the Qwen3-MoE model.
+
+WIP: MoE load balancing is currently work in progress.
+
+[Link](https://github.com/d9d-project/d9d/blob/main/example/qwen3_moe/pretrain.py).
+

d9d-0.1.0.dist-info/RECORD

@@ -0,0 +1,238 @@
+d9d/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/core/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/core/autograd/__init__.py,sha256=uQrNtaXUMxXYyt05j4vRtiI5_pOUg0l4TXDXG3BeXyA,167
+d9d/core/autograd/grad_context.py,sha256=nWN3BvM-GEFO4CZ8pEH5pjdYkrymOkQaNQXQF7laRPc,2891
+d9d/core/dist_context/__init__.py,sha256=HqGt6_gplSNYnohGZl2GEmDO7fKloJVNiKBC9GXIQdM,498
+d9d/core/dist_context/configured.py,sha256=wqVRzpOibMSYKULNLU4HynCoLWhRU2z3nNhTNZ3Y3BI,7127
+d9d/core/dist_context/device_mesh_domains.py,sha256=pI7gW2e-tQgx0TcZOAT4aFBYiiRph6VgF03zJL6WlyE,4854
+d9d/core/dist_context/log.py,sha256=mG-3gkZBleBTUJiOlUGDfXaSFRf5q5fGK97HybWe_E8,900
+d9d/core/dist_context/params.py,sha256=SOgu5T5Q1PNSdM_06U-wuRHwSMxIT8_6ua3t8dTmkeg,3645
+d9d/core/dist_ops/__init__.py,sha256=iwatM6f60NWU8oAYoHcFWvBc7HwTfsQpaaRFg9GwIDc,397
+d9d/core/dist_ops/object.py,sha256=cKAL2M5aUeCYggYqeari-FlsMEYvZvDszhXdLTHTDlo,2006
+d9d/core/dist_ops/tensor.py,sha256=6zqHakdQPDXUbjOQ0tu2rLvM5fsf2RP8503tQzhfciU,6165
+d9d/core/protocol/__init__.py,sha256=amTi465uKy2kEEQfYxrfIQ3METESZluwDoBBBi2BjYE,203
+d9d/core/protocol/training.py,sha256=i3DHGy9P9G9X8UPg45zLX-_eOrEI5Y3DjQ4A2Wix5Ec,961
+d9d/core/sharding/__init__.py,sha256=DYkdFk8LVIssTZr-DdnlXUmz4Xd3OpsoqFLzqsFdsRg,385
+d9d/core/sharding/auto_spec.py,sha256=M8fIm_8zqtGsAsPk8UpPZaUHdhLipjHDvFj6L7zTlLA,2283
+d9d/core/sharding/shard.py,sha256=UhcaCgmJNwb8vhRvKi3mwJ-22zwtB4E1o0qPNn69IgE,5243
+d9d/core/sharding/spec.py,sha256=4sA70POEdWzekBzOP0DwdUbfJa6HBr0VrLVqOTpjUWM,668
+d9d/core/sharding/unshard.py,sha256=T1OQRZ8BwCSTf0Dw4_-N79IDDD6Fwii_-idjpurg7hY,3696
+d9d/core/types/__init__.py,sha256=f2fX48bowRCELUEMLICLwLwi-sl_B5f9VQLfpVJCdIY,223
+d9d/core/types/data.py,sha256=IyZtwLSlofHPzc_F-BrCQVgpBISh6wbInuCMVW92IXk,473
+d9d/core/types/pytree.py,sha256=L93BReqv7fiegPVRzZQd63FYv986rJdMbk-bYFv2bYA,842
+d9d/dataset/__init__.py,sha256=UhxpUpSM5gGtXaVSZ1Y5m8eEO61E5ebcYx-hF9z8T7Q,509
+d9d/dataset/buffer_sorted.py,sha256=oiCnoKfhofo-RQszEX9UZGNEi-E9Gq-oQUbpD5-qYe8,4996
+d9d/dataset/padding.py,sha256=MSG7tkVHjAJdUAR1BKkW_s-14mBJ-g-0GZyf-8MA0PE,2333
+d9d/dataset/sharded.py,sha256=D699wxy0TsGsuIKwiRK4fpuJYn5CJoSddyXREdUFIYs,6923
+d9d/internals/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/internals/determinism/__init__.py,sha256=702y3p8NFHExaCQ962JImMLKPbvEZpV3nE4MYfEOWIs,150
+d9d/internals/determinism/seed.py,sha256=daJ85obD8bw580zCeKGqXCRZ0MesunZUTX41yD2ljGU,2271
+d9d/internals/grad_norm/__init__.py,sha256=rS8lXo3vOLxLUY-rujevcGEQHl4gLKLfAse2gGwyoYY,217
+d9d/internals/grad_norm/group.py,sha256=tWybTJTGYt-c_9by3wdfj6SwoTFbRC6G2U2OotFo2nI,2802
+d9d/internals/grad_norm/norm.py,sha256=c_rQNfbErswYhaaEuFeTd-4rW20w2PxcolC4TL2kYaw,5275
+d9d/internals/grad_sync/__init__.py,sha256=tllIQpOxQcz3bRbBFgG5FrHDzyGXWwQ-IPLKYW7cIZU,326
+d9d/internals/grad_sync/bucket.py,sha256=G5asn2dErBvFYCBlhYlafiAay65EfeiEMTQu-ZZ-id0,8935
+d9d/internals/grad_sync/placement_helper.py,sha256=-3Tdkrf745iz9vd68Iil4PQ-oP8eSGLi_XrjWjp0uYc,634
+d9d/internals/grad_sync/synchronizer.py,sha256=-z3PHc9nPw5pHz9Ht5DYOWNQsJWss5e1I82w9Cc_8lE,8012
+d9d/internals/pipeline_state/__init__.py,sha256=WYEWgFgp_8ycn7_T43DTGG_Kj2Iq_xo6lsMmyzS3a7g,359
+d9d/internals/pipeline_state/api.py,sha256=JZCqSZuJPXHvodyrcMvTs_gxO39FMoAlmhWSLAOGDHI,1102
+d9d/internals/pipeline_state/handler.py,sha256=MOUmv7Xvl0d7oo8Hg5G6Ayvhw8XNX-PCB7F1r7YZTtQ,3172
+d9d/internals/pipeline_state/storage.py,sha256=4EScog_SKZfCHlTivFEJnXfMnxhX9rXfbtEGWStEoDs,6522
+d9d/internals/profiling/__init__.py,sha256=8timWWLXgZyJjYACmdgi-SdUC-rpWKJqCcG_Fr1Boy8,110
+d9d/internals/profiling/profile.py,sha256=vm7jgJRyRCzRW7w-ebSbaOtz6OU0ElOgwe9EIR9hp2s,3848
+d9d/internals/state/__init__.py,sha256=8SOtDBRhP9k-WVKLTnmMkcgNUnSQY6z4LaBajD0mO2s,161
+d9d/internals/state/main_process.py,sha256=I5CxCjZ7xoLVtqbFbBY3Y_1T6_7a89BSqlzy9YsDdzg,1533
+d9d/kernel/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/kernel/cce/__init__.py,sha256=ndtKganqw3kgjdYE0WldhbsLecBA9H_c2I-2DyAE1UQ,81
+d9d/kernel/cce/cce.py,sha256=cL4Us9GEBPtjPrSMjOAeODOWVyBsTOQAwg2O6oVvyXE,9755
+d9d/kernel/cce/main.py,sha256=LxheuDLO5gCHtwy6qJ6a2jMGJbXT9Ex26iXy_e0A7fk,9087
+d9d/kernel/general/__init__.py,sha256=Gv-CuWwtHXwgjqhus_eSVJMiOtOnmbWi2uliCnGjr1s,76
+d9d/kernel/general/get_int_dtype.py,sha256=IcRLpo3AiPJ7pJjkRYWaodf8X8XPNJBxEq5-6k-7FMQ,183
+d9d/kernel/gmm/__init__.py,sha256=h9VaTX9wiYLvXUi00zk98CAc35Nx_dAwwny4vaheYeQ,51
+d9d/kernel/gmm/function.py,sha256=rVRwowNsZYlLlsV47-7LQNJBxRZ1-Arj7PPYClWKSoQ,2623
+d9d/kernel/moe/__init__.py,sha256=tLQ9bF4gDNR6MvVpS2No5RyTCl_b4lpRTPimkiEwbHA,237
+d9d/kernel/moe/indices_to_multihot.py,sha256=xdUVsRHLT19R9JR2dnMcPqdjyHSUyt5312sU8pbLhyU,10138
+d9d/kernel/moe/permute_with_probs.py,sha256=c244yZTEojDH7vbrTJhOTTmzQsiAKQM03ODTFgRSGcQ,35713
+d9d/kernel/stochastic/__init__.py,sha256=yXJ5Cf8SJ3Iag_2_uAKUi2BU4xkxdIp-ontyWfEE87I,245
+d9d/kernel/stochastic/adamw_step.py,sha256=dD7Z9tucQR6U203-XU_s3lEw7q1X_5qVkDXvgH8EOEs,6831
+d9d/kernel/stochastic/copy.py,sha256=S5NNOp1Wxzi8KszhkCh_j5pU0dBE-wsRpl9KHWCuvzQ,3211
+d9d/kernel/stochastic/ops/__init__.py,sha256=aOGiKLx82ZzW42fhTabz0rD7OSYyX0JIZkhM4jXNaW4,80
+d9d/kernel/stochastic/ops/round.py,sha256=w511LzP9a_p5oJWjQmD4QHAVzvpLNK8xnsHbLkMgDnI,545
+d9d/kernel/swiglu/__init__.py,sha256=bYHPadWelNxz9NVCYiDvm_igMZ-gnOLutSXQlVYbyLk,61
+d9d/kernel/swiglu/function.py,sha256=KMpDFEIN9CUmbKT1Z3Eqy9hS_2cvdMxmuqD-dzHleBo,918
+d9d/kernel/swiglu/op.py,sha256=r0AS6ckIl0R_0TeW6xDSJWgJqS2_TWa_ZrulwuOTxxo,4454
+d9d/loop/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/loop/auto/__init__.py,sha256=zvpOu934BCSblPK7N03VHEKUzz4EoiEmDEhDpACm7D0,279
+d9d/loop/auto/auto_lr_scheduler.py,sha256=53244OD5ooBpx6lch9SW0KDNVamLhPsCguxuVOKjpeY,1419
+d9d/loop/auto/auto_optimizer.py,sha256=2823wmCdFydERuQKVu5jDMuKAjb0volA3-XNdE6ntZc,6035
+d9d/loop/component/__init__.py,sha256=IgVUupHm-nPEkLx3BNcDClJrxwVqVm62Bo5g47o8QJM,1090
+d9d/loop/component/batch_maths.py,sha256=IYWf952-8aOuoZXVKWPgK7vgPYnn0It3QpXK-tg4cho,3962
+d9d/loop/component/checkpointer.py,sha256=3yhjMutYzeQLxkHYzMe1wpKcNpIeVAv_wtMpQjQAa5Y,5794
+d9d/loop/component/data_loader_factory.py,sha256=XOI9_XLNmqjMkmQguPM4g_-LbhANNS73hGPE_6fCyOQ,8029
+d9d/loop/component/garbage_collector.py,sha256=t1wHN4HvaaDfvhnBPfSsq5_7THd_OuYO8tuUA95-UGQ,2867
+d9d/loop/component/gradient_clipper.py,sha256=0nDGELrwczlqCpHojeC4N3j7dMUhELnod2pWTY1Hv3g,3078
+d9d/loop/component/gradient_manager.py,sha256=rtuDooke4HeoBoMXr_YqxSYecZHSzr-pAbXGmSsl10Q,4976
+d9d/loop/component/job_logger.py,sha256=CZx9EGnILwGQ8mqTaKstjVp_n_5_UO5fQnTEWuJAdsM,5183
+d9d/loop/component/job_profiler.py,sha256=qTK9LUAbLzuvwfkA_UyOVUMsO5s6cy61x7ZeuWQmiCQ,1835
+d9d/loop/component/loss_computer.py,sha256=J1SBzFRjyDZ7QbE7rfMvxyzLjtod5E0NynwNJqJ-qlE,2845
+d9d/loop/component/model_stage_exporter.py,sha256=g1f7WDmsJ6MAN_H8ymh4B3MHIURibUhO_3Mvespwfpc,1332
+d9d/loop/component/model_stage_factory.py,sha256=e1BpmjMwfot9V1pI5drcHCmrAvMJTyb0AwXfu_2MgTM,9831
+d9d/loop/component/optimizer_factory.py,sha256=fjqknSmlYbcxVogO2YgZht6-YhcMvfIh0AExFQ-aDdw,3636
+d9d/loop/component/stepper.py,sha256=5IM-5uvXswfB3---Kk_Dpi2iNBjmOCWqDXGl9v263wU,1804
+d9d/loop/component/timeout_manager.py,sha256=S7bF2iTQVkfvvhzR6wUVXuD2OCTzOMkzFt95eCs-MUE,1679
+d9d/loop/component/train_task_operator.py,sha256=cpeF_16H6Cj5WdLYogr34NI-ko8Q2MO-BSKzsgf_H-o,5438
+d9d/loop/config/__init__.py,sha256=jjikLZrT8zqGcq_LD02T1qd1np0O4j0zaDKI_XUF0_g,824
+d9d/loop/config/config.py,sha256=Pzb2c1f_z66_0KFyc_L-_abeK3rxezg_KNMtCY8BXR8,7039
+d9d/loop/config/types.py,sha256=JZ0A8-pytCV65drry17vmfWWzT9cyzQ6V4uU_ercru0,652
+d9d/loop/control/__init__.py,sha256=x_42CRVmwh1RKfh238QalrycFY7wDa_L7pY8wFnqgI0,1758
+d9d/loop/control/dataset_provider.py,sha256=KELIAbkVthBP78XUE3E3T3M8zUL1l0ra3Ezr6_FDInU,1605
+d9d/loop/control/lr_scheduler_provider.py,sha256=0hxJxIhvlFl-z22rZpeWryQ79LZK8VTU6hx8-oeAQUw,1177
+d9d/loop/control/model_provider.py,sha256=SAIFJwF8rMk4YGEAzFVTmWkIqPBjlrUxvqgK8F_jo-M,4628
+d9d/loop/control/optimizer_provider.py,sha256=M7DL_6298avKGsjto7j4VparfXQcIHvZ2CiYfSlV7T0,1039
+d9d/loop/control/task.py,sha256=YC3Itf_qo7WZmPrX3TVVIaoYVUsf18j56majdO6icjk,8150
+d9d/loop/run/__init__.py,sha256=wQlVo3TayCwoHaONBsfwsAPBu3uue8ckHaM4YDqIkMc,106
+d9d/loop/run/train.py,sha256=XbCa9r7xpVbVPb0dnD9uutPDDfdOZi1JhtD7dQwehOw,12244
+d9d/loop/state.py,sha256=s0Iz6W8iHO9md_3QqbQouqyVz1r7vt8xvcj1AoLqtvU,5017
+d9d/lr_scheduler/__init__.py,sha256=ZQke_m2Zl1adfFGp3wCvYQo-CO_c3bpVePrVGAERhfU,140
+d9d/lr_scheduler/piecewise/__init__.py,sha256=8SPGeMnZSF40pPjfEjkhAd5fb4e8fYjeNdF-pRuC1lM,496
+d9d/lr_scheduler/piecewise/builder.py,sha256=94ufa-aoL90xUIGu_W2XwkWPIkE53KHeUfNvwK61zew,4974
+d9d/lr_scheduler/piecewise/config.py,sha256=lljITKE8gpVyzEIfONzmYbhRSXZSpgieqIFH3yxnoec,4699
+d9d/lr_scheduler/piecewise/curves.py,sha256=LG2GzEYKJhiF9701D8lzxvtO4pzvhSObVEHMiayVCTE,1995
+d9d/lr_scheduler/piecewise/engine.py,sha256=ZV1L4mRDxZABIPhCR8BmwGZ8qzr2z4NsNSJZl9-hcXY,2054
+d9d/lr_scheduler/visualizer.py,sha256=CcaSOewQbM__IGlLNkzRzyvj0X9BAgCEFaWGu55aCmM,2083
+d9d/metric/__init__.py,sha256=VTByKUezSl_xOjpt0FPsRl_zzAXTu3pBHWYrCbocPa8,115
+d9d/metric/abc.py,sha256=Nkd6dhS-KhAf5MWacYCsjmBgqsUeEm-UgA5KPHkv2Rs,2485
+d9d/metric/impl/__init__.py,sha256=6xHkax_7Tglg5EUHnrluo7Pa7kcBp4l4fZ-LRT7x83A,133
+d9d/metric/impl/compose.py,sha256=oQyp3mmku-m_SXqDK5WbjF9Gv0HSsUQVfi9gmLwNuHI,1693
+d9d/metric/impl/mean.py,sha256=Lan6eDNUK617OMk4p99pFA1rGe1f1YwL0e51MB0SoA0,2832
+d9d/model_state/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/model_state/io/__init__.py,sha256=3RXRA5RLHS3hCQlvhRsy9q-KrGSNuh0r6kAppak0EA4,542
+d9d/model_state/io/dto.py,sha256=NY_C8sU9QoeXJyRqiGYjl7W-pfUUYSPe1L_Z_dvctkE,704
+d9d/model_state/io/module_reader.py,sha256=3-xOLPbdm20Ssm-E8quluCmL5Cwc70eWkUOYNe6bkB4,2779
+d9d/model_state/io/module_writer.py,sha256=TG7kVb_YS90U7Hv7oHqqzYx8pp6q6ba1TTu3E5AHJK8,4373
+d9d/model_state/io/reader.py,sha256=WACBaeVrezQUWU-xKqwIo_e08MeyEy-kr4anDTYQFMk,4546
+d9d/model_state/io/writer.py,sha256=5TpK2EMepKbQ6hOy_stgV7j33OzafQWw6czVwCsmkyQ,11621
+d9d/model_state/mapper/__init__.py,sha256=EVjhWQF8WLLtJp_Z7irGJomw5R0TYbA9x4_dttt0rVg,178
+d9d/model_state/mapper/abc.py,sha256=yXI7FDblBWrcSJb-_rPWHRARJSbw2JrojQI_ENluBFg,2634
+d9d/model_state/mapper/adapters/__init__.py,sha256=T5utmHgdzAVb-RcoJMN3U126EJb7CQX98uS5nCrnsrY,362
+d9d/model_state/mapper/adapters/mapper.py,sha256=rqROLlPux0qzU078fvu8nJ67cooGwe0POOtAQSIzNss,1083
+d9d/model_state/mapper/adapters/module.py,sha256=DM6VYVR97OWhfRUE1I7MaobPnF3pZErkkg-i9l7nS5c,775
+d9d/model_state/mapper/compose/__init__.py,sha256=-i0nYcl3iUHTW7d6eErk9VMh33-GBYXVpMdYXC9wa8s,469
+d9d/model_state/mapper/compose/helper.py,sha256=oPDJdJb5brqbWubVEnArE-UNJd572j33lqbyw3TrmlM,637
+d9d/model_state/mapper/compose/parallel.py,sha256=2zU8fEGhOKprM6bQfjbYI0oB8FA2gC0mTwwyvhrur8Y,2234
+d9d/model_state/mapper/compose/sequential.py,sha256=CNLFadwlvoPy0iZsI3Q4GKCq5weGkw5zRUtjqFnoNLI,5548
+d9d/model_state/mapper/compose/shard.py,sha256=ah2ONYdpxmPBJS4OssSnAzb0Wki1YQdrHOUsgwB3UEs,1484
+d9d/model_state/mapper/leaf/__init__.py,sha256=-S2w0jFXsSq5NiMUCO8w6PJAHO_IQpR2IAv10O7OoIQ,570
+d9d/model_state/mapper/leaf/dtensor.py,sha256=uDYltxXprQJfE1e22V06dpVJfoxi5kBHDQpvd-Gzl0U,1861
+d9d/model_state/mapper/leaf/identity.py,sha256=mZZAdurT42V6dy4964vCxEP4i-sXgiZfOBZOIyYKjTo,603
+d9d/model_state/mapper/leaf/rename.py,sha256=obsPpw9Y1SOGW-DZUX2scxFC2mAs8V7ivVEh19YUihw,739
+d9d/model_state/mapper/leaf/select_child.py,sha256=QxBft_g9WfBds_f8FTpFK6pLDMeumgBehhMDH6Kr3SQ,1161
+d9d/model_state/mapper/leaf/stack.py,sha256=TZ3ulpahcJn3qbKErzCJOIEt65mOLooxCmqktZ7F3Qw,1017
+d9d/module/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/module/base/__init__.py,sha256=vl05xyGyB-jQV_S50NvgR--jfThueRSl3vssLXRNh3U,178
+d9d/module/base/late_init.py,sha256=-Re3QCZcJO6vzBxN7FTj5r1H0hzPX6_wyIgEud9s-eo,292
+d9d/module/block/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/module/block/attention/__init__.py,sha256=hT32obgT6bRpUAJIFmDgfqtR5P9zwTS6UiiYL9v2hMk,141
+d9d/module/block/attention/grouped_query.py,sha256=4lqZfQgLNt0GkQ2GueQ9SJTKuV76qT_skJ41coHmlls,4814
+d9d/module/block/attention/sdpa/__init__.py,sha256=YijNzSFZedbMjGYGz0aE8AMriADadYffcjgP1oNbYzg,60
+d9d/module/block/attention/sdpa/flash.py,sha256=0YHd0Mtd7-Hrx-VORNcdZVmMHwE_N8bkScSoxfCNlQY,1925
+d9d/module/block/embedding/__init__.py,sha256=tNRPw0FMryV_SCRlf5YVhyTUGtqlezB1aBB0pwN0ZfM,163
+d9d/module/block/embedding/shard_token_embedding.py,sha256=PeaW_ekbhsYxf03dzlzIoW7K8tsT5JSs4b3pNWtZ_Hc,3519
+d9d/module/block/ffn/__init__.py,sha256=3WbxF0WP1Kp1Z5_qIVuYjmuDiNiiUNX3r3_rMstf8nQ,55
+d9d/module/block/ffn/swiglu.py,sha256=bIAVs3mNOJo_vvCV9p9MkmLkY8UpHAqo68rkHJmYbP0,1598
+d9d/module/block/head/__init__.py,sha256=9fhLQSXXOycz-dN2Pym-r9bzZCqsb1VTFBrS67BHMGc,147
+d9d/module/block/head/language_modelling.py,sha256=TKOvxPIYAQqiVAzRMORI3e1zB0UNKj54hznvAblGfP4,3071
+d9d/module/block/hidden_states_aggregator/__init__.py,sha256=oyi_9ZXn-0gkw6YQ8uzEyNOJupkuaimqbQkQUgKDHYc,305
+d9d/module/block/hidden_states_aggregator/base.py,sha256=Ty1_UlOOV7DuIPbpcjGnzjxrMAWoEfsZQCeiySF7mIE,1261
+d9d/module/block/hidden_states_aggregator/factory.py,sha256=fmar06v-oMwmo-zEdFpFDCGC5ic6x1LGEgyzVUO_e04,1511
+d9d/module/block/hidden_states_aggregator/mean.py,sha256=MN9x2TMaGTHrAkft9Wsh3u8jF9nXgkArPXjcZBStM60,2125
+d9d/module/block/hidden_states_aggregator/noop.py,sha256=dqqOZ-ZCfOYQJXU7rxUE-nLtPca9G_dcd2RACsqW6Dw,654
+d9d/module/block/moe/__init__.py,sha256=oL0Li4raYSufCQY_3llP_RBYXOnQi3_cmoEib8YcqDE,310
+d9d/module/block/moe/communications/__init__.py,sha256=cud1BcwcozsGENkQ70DtIPTXUE1G2o-5E_Okg8ffOC0,330
+d9d/module/block/moe/communications/base.py,sha256=zZq6IBMG86Eo2j0EZZGgkMj3bJ3sj2rH-SszhbcqHeQ,2107
+d9d/module/block/moe/communications/deepep.py,sha256=PRTCCEWnGT3u9XB_3V3cbupDGdhsrZkFk44sIrBOn-I,8809
+d9d/module/block/moe/communications/naive.py,sha256=jQU8u9z8ooFiEN18Wc8uwqq505Fa9c18My5SP03vNAE,2196
+d9d/module/block/moe/grouped_experts.py,sha256=0bMDGPyXINmDrcIbxJeAS8NwI0ZlUFfxRaQ7sg0Lq7Q,2782
+d9d/module/block/moe/grouped_linear.py,sha256=eVMvwnholSL-m5lDrGfsK0kYcGYbbPqkuH80JAOao9k,2529
+d9d/module/block/moe/layer.py,sha256=UhKuRLAFLtDyv9o7-9SxjEi_N_TaG9AkWITuPR6qqgI,4331
+d9d/module/block/moe/router.py,sha256=7l7C2Hr3AINlOvyfVRA---fEXMT6IHptPX-w0eS_fxw,3291
+d9d/module/block/positional/__init__.py,sha256=RsPCbuRgl3UW7Fow5dRf014MuwntakJuQiW-XezWYxw,236
+d9d/module/block/positional/rope.py,sha256=SB9cCcUTQXQOjdebGcn3v7nWDgMU8vfyNiZQY4NyAbE,4954
+d9d/module/model/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/module/model/qwen3_moe/__init__.py,sha256=dQbZpQYSFtoq0cSLnX4OHNTiEBDxEmmY9_BMUB9iXtQ,384
+d9d/module/model/qwen3_moe/decoder_layer.py,sha256=Pu8J8qi3ZdYLQ5UG1LmkcBDRceSPJLFbQn9wIjiMAlI,3334
+d9d/module/model/qwen3_moe/model.py,sha256=KHDX1XApNRcm-0hiNRhIKPwS-D5E2J9biiBVnVkRso4,14025
+d9d/module/model/qwen3_moe/params.py,sha256=OlnDLVeUWCydkyCnrbzdGhQiuVY9YgQWSTmFTLoQjsg,2480
+d9d/module/parallelism/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/module/parallelism/api/__init__.py,sha256=KCiPpSlL2yy0-KN1B-zex6v0oQo-ep7mXupKBKEkvtQ,558
+d9d/module/parallelism/api/expert_parallel.py,sha256=tid3TXadzeUhPRfVcYMP7eynSq2oMSkzrA3gQ3LxfAw,1399
+d9d/module/parallelism/api/fully_sharded.py,sha256=K0ttq70F6IF0dmvFGxR4PwcIb9Qpw7LKx5DkQQR2dw8,1643
+d9d/module/parallelism/api/hybrid_sharded.py,sha256=230Rg8KG9SfOF9Iv17oW1GVsdajc5UEuJztmWvyNsKQ,1715
+d9d/module/parallelism/api/replicate_parallel.py,sha256=HHMxLybsxtX7yaTx6ek1u8hU3Yo0p8yK7TGf4YBqpgk,1246
+d9d/module/parallelism/model/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/module/parallelism/model/qwen3_moe.py,sha256=eKC9DkCMrnzemvcdryhkkZ3l5Q0Do3i2A9PhwvZqRhQ,3575
+d9d/module/parallelism/style/__init__.py,sha256=ICUJJs6V-W3AWQ9gTY4fRDQHKzG18EzfaisSHEgHL1w,169
+d9d/module/parallelism/style/shard_experts.py,sha256=wm8x7OVfRJl7oe8quQ3no3qu0VLyJsPUwXMkQohzJa0,2158
+d9d/module/parallelism/style/to_local.py,sha256=xCwBKhbnybElPXhivf3qb92YddxWa5X55pGFfkrh4YE,3273
+d9d/optim/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/optim/stochastic/__init__.py,sha256=H4sJvNTHs27gIJ8lR9h8AM0j6BrnPlVlgOi8LiDid4c,72
+d9d/optim/stochastic/adamw.py,sha256=XNTUtm3Z4YAPqZFKJsD09cTWmKd0i5EnrJ-sC-1WnQc,5530
+d9d/peft/__init__.py,sha256=XTgJcW9v4sQTkur7lcMZ0BkJaqWU9BruotImwK1fCss,322
+d9d/peft/all/__init__.py,sha256=ID2lcBOD5-4W4sL7lsSM37psqxV4-EfsU2NXO3D6nWY,243
+d9d/peft/all/config.py,sha256=YvCsWGDU2TRO0eIRk5-kTbX9V9DArob5KxxXK7HxGoc,765
+d9d/peft/all/method.py,sha256=2IztonETLROCOTxt61nJ4knEwXRVcQxAi1ptgRn80BA,2117
+d9d/peft/applicator.py,sha256=jrqEwKY9EbLwo_qZHaQhTubVDgv8okQmZdcd6PBS1BI,1393
+d9d/peft/base.py,sha256=kPImZaZETnIhew3_MhbViYrXvZJG16uraqEDZpr26XI,1729
+d9d/peft/full_tune/__init__.py,sha256=Z4U8Nbqt6kZJZqNlDV8kX-m2YW9QjLJ5dc0BxvpLru8,195
+d9d/peft/full_tune/config.py,sha256=ZdYDFn1DVmKdibk502WvU9rdAmJj0NKopPzOzmhAJYM,494
+d9d/peft/full_tune/method.py,sha256=Hut8VHMfX05RnmSS6lJzkSv49uB4BpwagQEus5KX4d4,1265
+d9d/peft/lora/__init__.py,sha256=nr3R8uJlCkt2mmZYpFtWiBfRDdpUtaU9zx2r5Nl8NYo,294
+d9d/peft/lora/config.py,sha256=KElokzzMJeh91SwRrxj-ZTzSLHF8ZVSb8ehNfYKFFpg,801
+d9d/peft/lora/layer.py,sha256=7i3QGR4QveJ_Tk_yNl_QZuLzK-l1y9xp5O0N6DRtOrE,4990
+d9d/peft/lora/method.py,sha256=psYYkSvL0telO2ULM8dtDTlf4H87lTtn_4hlB07M-iw,4123
+d9d/pipelining/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/pipelining/api/__init__.py,sha256=6vwCq9PiJWkREzoM68yO-UySf2YEy6BK14oph3IOuHo,439
+d9d/pipelining/api/module.py,sha256=WhDa4GTGoaL00XMoPAQB-XdjcmGiL_U0T9mTwRSMBN8,5280
+d9d/pipelining/api/schedule.py,sha256=m6FCe1Yh2cOYGxShirq3HakDlRnQyAIwePA5-UU_xxE,1629
+d9d/pipelining/api/sharding.py,sha256=XhaRarZOHf6nDqwnzLvz0SPh9VHi6kDiQkkx_gRNoHY,204
+d9d/pipelining/factory/__init__.py,sha256=_tN9unIv9bHojC26MKt34W9ra-zMdGrfS0i0a4JyurU,602
+d9d/pipelining/factory/config.py,sha256=MW8ARW0gPZGN9jsPPq-nTSFaGYMKPDwIBO06juy4DTE,2565
+d9d/pipelining/factory/factory.py,sha256=E_eUyvYP6YXMzy8QdFUoBfpFUkA39eFk7zeYsidqGzU,3920
+d9d/pipelining/factory/registry.py,sha256=k8nBSM2vlYE_fAo4pV7zzX9iqOopLJ0YugeieYwwU10,3163
+d9d/pipelining/infra/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/pipelining/infra/schedule/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/pipelining/infra/schedule/component/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/pipelining/infra/schedule/component/program/__init__.py,sha256=c_FBMR7y2fo8Hbzxu23fBXgnfJ0PYXoFHnvS78tuPKo,572
+d9d/pipelining/infra/schedule/component/program/base.py,sha256=tVJNJr6IWSg-qT72Nci39Z4eCX0ggO5oJcJocwJFOOk,1010
+d9d/pipelining/infra/schedule/component/program/communications.py,sha256=2Q-4LCVWgB1TAdf1-GcxRTo8GZuh4ljlHyVusTl25Qo,6893
+d9d/pipelining/infra/schedule/component/program/topology.py,sha256=4c47yx_9-tx37ySCzdK-DXpJi0e42xC_Fkdg2Z2cwQI,2364
+d9d/pipelining/infra/schedule/component/runtime/__init__.py,sha256=XouJZVjEfekcoeRWyzjGHD0D8yRg90JaXVeqojRnJPI,633
+d9d/pipelining/infra/schedule/component/runtime/action.py,sha256=8Rgtk5pZjXCbsjoJxhArkEFIeNzY1ZWvESCgLnA8Hyg,10733
+d9d/pipelining/infra/schedule/component/runtime/communications.py,sha256=X1vSYnBHY4uBB0TvAApiSbdibJxxwNV-EZ3l_LXRnC4,3344
+d9d/pipelining/infra/schedule/component/runtime/executor.py,sha256=QAbvdA6vfS909jririg_Rw_FK2H6aZUYyAILGOpGHU8,4392
+d9d/pipelining/infra/schedule/component/runtime/loss.py,sha256=DGATscTXF2fKG7M2UiagpnVYvUI3DSmKKOMH1QLvH0k,1646
+d9d/pipelining/infra/schedule/program/__init__.py,sha256=991C1CTiAoC93hxsZxTJoBLBFUa-9phcNIIH8SUw4vA,447
+d9d/pipelining/infra/schedule/program/bfs.py,sha256=wiTGaFaUJ2O9nAB3qTtgH1AM8NZrZbF3v5bqDY5VQvw,3092
+d9d/pipelining/infra/schedule/program/dualpipev.py,sha256=GDuNHmqjOjTHgee18EE4py_9p97pBg91BJDcYh5HlSo,7875
+d9d/pipelining/infra/schedule/program/interleaved.py,sha256=6aGQsFTY_Jv3NU19lNenax_U4ji5pAO0dxXR_uRKfq0,8655
+d9d/pipelining/infra/schedule/program/zerobubblev.py,sha256=eWEeUcp-XItGmFKAP4-TmVVXpLTkfXAz4Mfsm2NvJ7g,7219
+d9d/pipelining/infra/stage/__init__.py,sha256=q_6-TTpeg55mSIhzM0aux5FSsB31SQfanDy6pau5Kwk,68
+d9d/pipelining/infra/stage/communications.py,sha256=h72UrnTK15AaJC-rb00cKrTu1JL6ny1TycH9tmuAzlc,9865
+d9d/pipelining/infra/stage/computations.py,sha256=F1LCCMAdqU7iXHnLGjbHn6WsU0aKAq7ctZb70W1GUXc,10426
+d9d/pipelining/infra/stage/splitgrad.py,sha256=xJkbhy6uKDxVAUyW-LUQhJQ55DRRKe5_8_V0GC0eroc,13166
+d9d/pipelining/infra/stage/stage.py,sha256=Ac7lAERDIUvnYNaY5UTGOOLMrVUd5-KSbxMSNRpShuE,11521
+d9d/pipelining/infra/stage/struct_helper.py,sha256=y9gTbKUmmsDHO6YCXPSKAj_Xq-vg6yoqHEmRkLVpekE,1361
+d9d/pipelining/training/__init__.py,sha256=mjxKtgXNU59QfgaxQrcahEG-kv6xwsNJBkBdq6FWJT0,154
+d9d/pipelining/training/optimizer.py,sha256=7VZ4b8I8CC0DnW4h-QYEhrf9jZm5LFsYUpei1Q2lkR4,1322
+d9d/pipelining/training/scheduler.py,sha256=QBDu4Z9P7Vqgt1R8AOJvhEISH4BChysjwtFRCZWyEfw,1147
+d9d/tracker/__init__.py,sha256=KShifEoIeGYksmAMMxm8jcuyaw9WsX4OY0wuU0DlhdM,323
+d9d/tracker/base.py,sha256=keS0UsgbffYT4mf8_LfZc-FueOsujePHdlBF4OIfPA4,3492
+d9d/tracker/factory.py,sha256=RVE_NZu9Tn9sI_1QEl3b-DPRvikg-nxksOZ_AcxK6Yw,1660
+d9d/tracker/provider/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/tracker/provider/aim/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+d9d/tracker/provider/aim/config.py,sha256=7-Kgt2XhdNRhgxLYEWpB27uhgB-RBiPrGvH6S2L3rIM,672
+d9d/tracker/provider/aim/tracker.py,sha256=g34BdjIYEftEwuK2oTKKzHLN3dpG7-i14zhXIF_onIo,3110
+d9d/tracker/provider/null.py,sha256=c1nvUaOz8RbRY8XzwSPTi7t0lSsmdlwGAYfYgprwaf8,1440
+d9d-0.1.0.dist-info/METADATA,sha256=NyAv5QdE_TtWzvNWcWormljU9O3DzQ8yAEp_8XZJv9s,5681
+d9d-0.1.0.dist-info/WHEEL,sha256=zp0Cn7JsFoX2ATtOhtaFYIiE2rmFAD4OcMhtUki8W3U,88
+d9d-0.1.0.dist-info/RECORD,,

d9d-0.1.0.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: poetry-core 2.2.1
+Root-Is-Purelib: true
+Tag: py3-none-any