lightning 2.2.1__tar.gz → 2.2.5__tar.gz

{lightning-2.2.1 → lightning-2.2.5}/.actions/assistant.py

@@ -44,11 +44,7 @@ REQUIREMENT_FILES = {
         "requirements/fabric/base.txt",
         "requirements/fabric/strategies.txt",
     ),
-    "data": (
-        "requirements/data/data.txt",
-        "requirements/data/cloud.txt",
-        "requirements/data/examples.txt",
-    ),
+    "data": ("requirements/data/data.txt",),
 }
 REQUIREMENT_FILES_ALL = list(chain(*REQUIREMENT_FILES.values()))
 

{lightning-2.2.1/src/lightning.egg-info → lightning-2.2.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: lightning
-Version: 2.2.1
+Version: 2.2.5
 Summary: The Deep Learning framework to train, deploy, and ship AI products Lightning fast.
 Home-page: https://github.com/Lightning-AI/lightning
 Author: Lightning AI et al.
@@ -27,21 +27,18 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Requires-Python: >=3.8
 Description-Content-Type: text/markdown
-Provides-Extra: app-test
-Provides-Extra: app-components
-Provides-Extra: app-cloud
-Provides-Extra: app-ui
-Provides-Extra: data-test
-Provides-Extra: data-cloud
-Provides-Extra: data-examples
-Provides-Extra: fabric-test
-Provides-Extra: fabric-strategies
 Provides-Extra: fabric-examples
+Provides-Extra: fabric-strategies
+Provides-Extra: fabric-test
 Provides-Extra: store-test
-Provides-Extra: pytorch-test
+Provides-Extra: app-ui
+Provides-Extra: app-cloud
+Provides-Extra: app-test
+Provides-Extra: app-components
+Provides-Extra: pytorch-examples
 Provides-Extra: pytorch-strategies
 Provides-Extra: pytorch-extra
-Provides-Extra: pytorch-examples
+Provides-Extra: pytorch-test
 Provides-Extra: fabric-all
 Provides-Extra: fabric-dev
 Provides-Extra: pytorch-all
@@ -49,16 +46,14 @@ Provides-Extra: pytorch-dev
 Provides-Extra: app-extra
 Provides-Extra: app-all
 Provides-Extra: app-dev
-Provides-Extra: data-all
-Provides-Extra: data-dev
-Provides-Extra: test
-Provides-Extra: components
-Provides-Extra: app
-Provides-Extra: cloud
-Provides-Extra: ui
 Provides-Extra: data
 Provides-Extra: examples
 Provides-Extra: strategies
+Provides-Extra: test
+Provides-Extra: ui
+Provides-Extra: cloud
+Provides-Extra: app
+Provides-Extra: components
 Provides-Extra: extra
 Provides-Extra: all
 Provides-Extra: dev
@@ -83,7 +78,7 @@ ______________________________________________________________________
   <a href="https://lightning.ai/docs/pytorch/stable/">PyTorch Lightning</a> •
   <a href="https://lightning.ai/docs/fabric/stable/">Fabric</a> •
   <a href="https://lightning.ai/docs/app/stable/">Lightning Apps</a> •
-  <a href="https://pytorch-lightning.readthedocs.io/en/2.2.1">Docs</a> •
+  <a href="https://pytorch-lightning.readthedocs.io/en/2.2.5">Docs</a> •
   <a href="#community">Community</a> •
   <a href="https://lightning.ai/docs/pytorch/stable/generated/CONTRIBUTING.html">Contribute</a> •
 </p>
@@ -626,7 +621,7 @@ Lightning is rigorously tested across multiple CPUs, GPUs and TPUs and against m
 
 |       System / PyTorch ver.        | 1.13                                                                                                                                                                                                                            | 2.0                                                                                                                                                                                                                             |                                                                                                               2.1                                                                                                               |
 | :--------------------------------: | :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|
-|        Linux py3.9 \[GPUs\]        |  |  | [![Build Status](https://dev.azure.com/Lightning-AI/lightning/_apis/build/status%2Fpytorch-lightning%20%28GPUs%29?branchName=refs%2Ftags%2F2.2.1)](https://dev.azure.com/Lightning-AI/lightning/_build/latest?definitionId=24&branchName=master) |
+|        Linux py3.9 \[GPUs\]        |  |  | [![Build Status](https://dev.azure.com/Lightning-AI/lightning/_apis/build/status%2Fpytorch-lightning%20%28GPUs%29?branchName=refs%2Ftags%2F2.2.5)](https://dev.azure.com/Lightning-AI/lightning/_build/latest?definitionId=24&branchName=master) |
 |        Linux py3.9 \[TPUs\]        |                                                                                                                                                                                                                                 |  [![Test PyTorch - TPU](https://github.com/Lightning-AI/lightning/actions/workflows/tpu-tests.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/tpu-tests.yml)     |      |
 |  Linux (multiple Python versions)  | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 |                 [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                 |
 |   OSX (multiple Python versions)   | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 |                 [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                 |

{lightning-2.2.1 → lightning-2.2.5}/pyproject.toml

@@ -243,6 +243,7 @@ module = [
     "lightning.app.utilities.state",
     "lightning.app.utilities.tracer",
     "lightning.app.utilities.tree",
+    "lightning.data.streaming.serializers",
     "lightning.store.utils",
 ]
 ignore_errors = "True"

{lightning-2.2.1 → lightning-2.2.5}/requirements/app/app.txt

@@ -1,4 +1,4 @@
-lightning-cloud == 0.5.64  # Must be pinned to ensure compatibility
+lightning-cloud == 0.5.69  # Must be pinned to ensure compatibility
 packaging
 typing-extensions >=4.4.0, <4.10.0
 deepdiff >=5.7.0, <6.6.0
@@ -6,7 +6,7 @@ fsspec[http] >=2022.5.0, <2023.11.0
 croniter >=1.3.0, <1.5.0  # strict; TODO: for now until we find something more robust.
 traitlets >=5.3.0, <5.12.0
 arrow >=1.2.0, <1.3.0
-lightning-utilities >=0.8.0, <0.10.0
+lightning-utilities >=0.8.0, <0.11.0
 beautifulsoup4 >=4.8.0, <4.13.0
 inquirer >=2.10.0, <3.2.0
 psutil <5.9.6

{lightning-2.2.1 → lightning-2.2.5}/requirements/fabric/base.txt

@@ -2,8 +2,8 @@
 #  in case you want to preserve/enforce restrictions on the latest compatible version, add "strict" as an in-line comment
 
 numpy >=1.17.2, <1.27.0
-torch >=1.13.0, <2.2.0
+torch >=1.13.0, <2.3.0
 fsspec[http] >=2022.5.0, <2023.11.0
 packaging >=20.0, <=23.1
 typing-extensions >=4.4.0, <4.10.0
-lightning-utilities >=0.8.0, <0.10.0
+lightning-utilities >=0.8.0, <0.11.0

{lightning-2.2.1 → lightning-2.2.5}/requirements/fabric/examples.txt

@@ -3,4 +3,4 @@
 
 torchvision >=0.14.0, <0.17.0
 torchmetrics >=0.10.0, <1.3.0
-lightning-utilities >=0.8.0, <0.10.0
+lightning-utilities >=0.8.0, <0.11.0

{lightning-2.2.1 → lightning-2.2.5}/requirements/pytorch/base.txt

@@ -2,11 +2,11 @@
 #  in case you want to preserve/enforce restrictions on the latest compatible version, add "strict" as an in-line comment
 
 numpy >=1.17.2, <1.27.0
-torch >=1.13.0, <2.2.0
+torch >=1.13.0, <2.3.0
 tqdm >=4.57.0, <4.67.0
 PyYAML >=5.4, <6.1.0
 fsspec[http] >=2022.5.0, <2023.11.0
 torchmetrics >=0.7.0, <1.3.0  # needed for using fixed compare_version
 packaging >=20.0, <=23.1
 typing-extensions >=4.4.0, <4.10.0
-lightning-utilities >=0.8.0, <0.10.0
+lightning-utilities >=0.8.0, <0.11.0

{lightning-2.2.1 → lightning-2.2.5}/requirements/pytorch/examples.txt

@@ -6,4 +6,4 @@ torchvision >=0.14.0, <0.17.0
 gym[classic_control] >=0.17.0, <0.27.0
 ipython[all] <8.15.0
 torchmetrics >=0.10.0, <1.3.0
-lightning-utilities >=0.8.0, <0.10.0
+lightning-utilities >=0.8.0, <0.11.0

{lightning-2.2.1 → lightning-2.2.5}/requirements/pytorch/extra.txt

@@ -5,7 +5,7 @@
 matplotlib>3.1, <3.9.0
 omegaconf >=2.0.5, <2.4.0
 hydra-core >=1.0.5, <1.4.0
-jsonargparse[signatures] >=4.26.1, <4.28.0
+jsonargparse[signatures] >=4.27.7, <4.28.0
 rich >=12.3.0, <13.6.0
 tensorboardX >=2.2, <2.7.0  # min version is set by torch.onnx missing attribute
 bitsandbytes ==0.41.0  # strict

lightning-2.2.1/src/lightning/e0b28c77f6af750da91bf2c61915cbf17eafa5878ba3331a → lightning-2.2.5/src/lightning/4d5c5591d7f5f4c27eb297f887022e3bf38682a3273da157

@@ -19,7 +19,6 @@ include src/lightning/pytorch/version.info
 include src/lightning/version.info
 include src/version.info
 recursive-include requirements/app *.txt
-recursive-include requirements/data *.txt
 recursive-include requirements/fabric *.txt
 recursive-include requirements/pytorch *.txt
 recursive-include src/lightning/app/cli/*-template *

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/__init__.py

@@ -39,7 +39,7 @@ def _cli_entry_point() -> None:
     if not (
         ModuleAvailableCache("lightning.app")
         if RequirementCache("lightning-utilities<0.10.0")
-        else RequirementCache(module="lightning.app")  # type: ignore[call-arg]
+        else RequirementCache(module="lightning.app")
     ):
         print("The `lightning` command requires additional dependencies: `pip install lightning[app]`")
         sys.exit(1)

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/__setup__.py

@@ -48,8 +48,6 @@ def _prepare_extras() -> Dict[str, Any]:
     extras["app-extra"] = extras["app-app"] + extras["app-cloud"] + extras["app-ui"] + extras["app-components"]
     extras["app-all"] = extras["app-extra"]
     extras["app-dev"] = extras["app-all"] + extras["app-test"]
-    extras["data-all"] = extras["data-data"] + extras["data-cloud"] + extras["data-examples"]
-    extras["data-dev"] = extras["data-all"] + extras["data-test"]
     extras["store-store"] = extras["app-app"]  # todo: consider cutting/leaning this dependency
 
     # merge per-project extras of the same category, e.g. `app-test` + `fabric-test`

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/app/CHANGELOG.md

@@ -4,6 +4,14 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
+
+## [2.2.4] - 2024-05-02
+
+## Fixed
+
+- Fixed HTTPClient retry for flow/work queue ([#19837](https://github.com/Lightning-AI/pytorch-lightning/pull/19837))
+
+
 ## [2.2.0] - 2024-02-09
 
 ## Changed

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/app/cli/react-ui-template/ui/package.json

@@ -24,7 +24,7 @@
     "@vitejs/plugin-react": "^1.0.7",
     "prettier": "^2.5.1",
     "typescript": "^4.5.4",
-    "vite": "^2.9.16"
+    "vite": "^2.9.17"
   },
   "main": "index.js",
   "license": "MIT"

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/app/cli/react-ui-template/ui/yarn.lock

@@ -918,9 +918,9 @@ find-root@^1.1.0:
   integrity sha512-NKfW6bec6GfKc0SGx1e07QZY9PE99u0Bft/0rzSD5k3sO/vwkVUpDUKVm5Gpp5Ue3YfShPFTX2070tDs5kB9Ng==
 
 follow-redirects@^1.15.0:
-  version "1.15.4"
-  resolved "https://registry.yarnpkg.com/follow-redirects/-/follow-redirects-1.15.4.tgz#cdc7d308bf6493126b17ea2191ea0ccf3e535adf"
-  integrity sha512-Cr4D/5wlrb0z9dgERpUL3LrmPKVDsETIJhaCMeDfuFYcqa5bldGV6wBsAN6X/vxlXQtFBMrXdXxdL8CbDTGniw==
+  version "1.15.6"
+  resolved "https://registry.yarnpkg.com/follow-redirects/-/follow-redirects-1.15.6.tgz#7f815c0cda4249c74ff09e95ef97c23b5fd0399b"
+  integrity sha512-wWN62YITEaOpSK584EZXJafH1AGpO8RVgElfkuXbTOrPX4fIfOyEpW/CsiNd8JdYrAoOvafRTOEnvsO++qCqFA==
 
 form-data@^4.0.0:
   version "4.0.0"
@@ -1260,10 +1260,10 @@ update-browserslist-db@^1.0.4:
     escalade "^3.1.1"
     picocolors "^1.0.0"
 
-vite@^2.9.16:
-  version "2.9.16"
-  resolved "https://registry.yarnpkg.com/vite/-/vite-2.9.16.tgz#daf7ba50f5cc37a7bf51b118ba06bc36e97898e9"
-  integrity sha512-X+6q8KPyeuBvTQV8AVSnKDvXoBMnTx8zxh54sOwmmuOdxkjMmEJXH2UEchA+vTMps1xw9vL64uwJOWryULg7nA==
+vite@^2.9.17:
+  version "2.9.17"
+  resolved "https://registry.yarnpkg.com/vite/-/vite-2.9.17.tgz#6b770525e12fa2a2e3a0fa0d028d304f4f7dc7d4"
+  integrity sha512-XxcRzra6d7xrKXH66jZUgb+srThoPu+TLJc06GifUyKq9JmjHkc1Numc8ra0h56rju2jfVWw3B3fs5l3OFMvUw==
   dependencies:
     esbuild "^0.14.27"
     postcss "^8.4.13"

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/app/core/app.py

@@ -30,6 +30,7 @@ from lightning.app import _console
 from lightning.app.api.request_types import _APIRequest, _CommandRequest, _DeltaRequest
 from lightning.app.core.constants import (
     BATCH_DELTA_COUNT,
+    CHECK_ERROR_QUEUE_INTERVAL,
     DEBUG_ENABLED,
     FLOW_DURATION_SAMPLES,
     FLOW_DURATION_THRESHOLD,
@@ -165,6 +166,7 @@ class LightningApp:
 
         self._last_run_time: float = 0.0
         self._run_times: list = []
+        self._last_check_error_queue: float = 0.0
 
         # Path attributes can't get properly attached during the initialization, because the full name
         # is only available after all Flows and Works have been instantiated.
@@ -318,10 +320,12 @@ class LightningApp:
             return []
 
     def check_error_queue(self) -> None:
-        exception: Exception = self.get_state_changed_from_queue(self.error_queue)  # type: ignore[assignment,arg-type]
-        if isinstance(exception, Exception):
-            self.exception = exception
-            self.stage = AppStage.FAILED
+        if (time() - self._last_check_error_queue) > CHECK_ERROR_QUEUE_INTERVAL:
+            exception: Exception = self.get_state_changed_from_queue(self.error_queue)  # type: ignore[assignment,arg-type]
+            if isinstance(exception, Exception):
+                self.exception = exception
+                self.stage = AppStage.FAILED
+            self._last_check_error_queue = time()
 
     @property
     def flows(self) -> List[Union[LightningWork, "LightningFlow"]]:

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/app/core/constants.py

@@ -70,6 +70,7 @@ DOT_IGNORE_FILENAME = ".lightningignore"
 LIGHTNING_COMPONENT_PUBLIC_REGISTRY = "https://lightning.ai/v1/components"
 LIGHTNING_APPS_PUBLIC_REGISTRY = "https://lightning.ai/v1/apps"
 LIGHTNING_MODELS_PUBLIC_REGISTRY = "https://lightning.ai/v1/models"
+ENABLE_ORCHESTRATOR = bool(int(os.getenv("ENABLE_ORCHESTRATOR", "1")))
 
 LIGHTNING_CLOUDSPACE_HOST = os.getenv("LIGHTNING_CLOUDSPACE_HOST")
 LIGHTNING_CLOUDSPACE_EXPOSED_PORT_COUNT = int(os.getenv("LIGHTNING_CLOUDSPACE_EXPOSED_PORT_COUNT", "0"))
@@ -99,6 +100,7 @@ SYS_CUSTOMIZATIONS_SYNC_ROOT = "/tmp/sys-customizations-sync"  # noqa: S108 # to
 SYS_CUSTOMIZATIONS_SYNC_PATH = ".sys-customizations-sync"
 
 BATCH_DELTA_COUNT = int(os.getenv("BATCH_DELTA_COUNT", "128"))
+CHECK_ERROR_QUEUE_INTERVAL = float(os.getenv("CHECK_ERROR_QUEUE_INTERVAL", "30"))
 
 
 def enable_multiple_works_in_default_container() -> bool:

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/app/plugin/plugin.py

@@ -46,6 +46,7 @@ class LightningPlugin:
         self.cloudspace_id = ""
         self.cluster_id = ""
         self.source_app = ""
+        self.keep_machines_after_stop = False
 
     def run(self, *args: str, **kwargs: str) -> Externalv1LightningappInstance:
         """Override with the logic to execute on the cloudspace."""
@@ -95,6 +96,7 @@ class LightningPlugin:
             name=name,
             cluster_id=self.cluster_id,
             source_app=self.source_app,
+            keep_machines_after_stop=self.keep_machines_after_stop,
         )
 
     def _setup(
@@ -103,11 +105,13 @@ class LightningPlugin:
         cloudspace_id: str,
         cluster_id: str,
         source_app: str,
+        keep_machines_after_stop: bool,
     ) -> None:
         self.source_app = source_app
         self.project_id = project_id
         self.cloudspace_id = cloudspace_id
         self.cluster_id = cluster_id
+        self.keep_machines_after_stop = keep_machines_after_stop
 
 
 class _Run(BaseModel):
@@ -118,6 +122,7 @@ class _Run(BaseModel):
     cluster_id: str
     plugin_arguments: Dict[str, str]
     source_app: str
+    keep_machines_after_stop: bool
 
 
 def _run_plugin(run: _Run) -> Dict[str, Any]:
@@ -191,6 +196,7 @@ def _run_plugin(run: _Run) -> Dict[str, Any]:
                 cloudspace_id=run.cloudspace_id,
                 cluster_id=run.cluster_id,
                 source_app=run.source_app,
+                keep_machines_after_stop=run.keep_machines_after_stop,
             )
             app_instance = plugin.run(**run.plugin_arguments)
             return _to_clean_dict(app_instance, True)

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/app/runners/cloud.py

@@ -34,7 +34,7 @@ from lightning_cloud.openapi import (
     CloudspaceIdRunsBody,
     Externalv1LightningappInstance,
     Gridv1ImageSpec,
-    IdGetBody1,
+    IdGetBody,
     ProjectIdCloudspacesBody,
     V1BuildSpec,
     V1CloudSpace,
@@ -197,6 +197,7 @@ class CloudRuntime(Runtime):
         name: str,
         cluster_id: str,
         source_app: Optional[str] = None,
+        keep_machines_after_stop: Optional[bool] = None,
     ) -> Externalv1LightningappInstance:
         """Slim dispatch for creating runs from a cloudspace. This dispatch avoids resolution of some properties such
         as the project and cluster IDs that are instead passed directly.
@@ -206,6 +207,8 @@ class CloudRuntime(Runtime):
             cloudspace_id: The ID of the cloudspace.
             name: The name for the run.
             cluster_id: The ID of the cluster to run on.
+            source_app: Name of the source app that triggered the run.
+            keep_machines_after_stop: If true, machines will be left running after the run is finished and reused after
 
         Raises:
             ApiException: If there was an issue in the backend.
@@ -278,6 +281,7 @@ class CloudRuntime(Runtime):
             queue_server_type,
             env_vars,
             source_app=source_app,
+            keep_machines_after_stop=keep_machines_after_stop,
         )
 
     def dispatch(
@@ -1016,13 +1020,14 @@ class CloudRuntime(Runtime):
         env_vars: Optional[List[V1EnvVar]] = None,
         auth: Optional[V1LightningAuth] = None,
         source_app: Optional[str] = None,
+        keep_machines_after_stop: Optional[bool] = None,
     ) -> Externalv1LightningappInstance:
         """Create a new instance of the given run with the given specification."""
         return self.backend.client.cloud_space_service_create_lightning_run_instance(
             project_id=project_id,
             cloudspace_id=cloudspace_id,
             id=run_id,
-            body=IdGetBody1(
+            body=IdGetBody(
                 cluster_id=cluster_id,
                 name=run_name,
                 desired_state=desired_state,
@@ -1030,6 +1035,7 @@ class CloudRuntime(Runtime):
                 env=env_vars,
                 auth=auth,
                 source_app=source_app,
+                keep_machines_after_stop=keep_machines_after_stop,
             ),
         )
 

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/app/runners/multiprocess.py

@@ -81,16 +81,17 @@ class MultiProcessRuntime(Runtime):
 
             _set_flow_context()
 
-            storage_orchestrator = StorageOrchestrator(
-                self.app,
-                self.app.request_queues,
-                self.app.response_queues,
-                self.app.copy_request_queues,
-                self.app.copy_response_queues,
-            )
-            self.threads.append(storage_orchestrator)
-            storage_orchestrator.setDaemon(True)
-            storage_orchestrator.start()
+            if constants.ENABLE_ORCHESTRATOR:
+                storage_orchestrator = StorageOrchestrator(
+                    self.app,
+                    self.app.request_queues,
+                    self.app.response_queues,
+                    self.app.copy_request_queues,
+                    self.app.copy_response_queues,
+                )
+                self.threads.append(storage_orchestrator)
+                storage_orchestrator.setDaemon(True)
+                storage_orchestrator.start()
 
             if self.start_server:
                 self.app.should_publish_changes_to_api = True

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/app/utilities/network.py

@@ -89,18 +89,36 @@ _CONNECTION_RETRY_BACKOFF_FACTOR = 0.5
 _DEFAULT_REQUEST_TIMEOUT = 30  # seconds
 
 
+def create_retry_strategy():
+    return Retry(
+        # wait time between retries increases exponentially according to: backoff_factor * (2 ** (retry - 1))
+        # but the the maximum wait time is 120 secs. By setting a large value (2880), we'll make sure clients
+        # are going to be alive for a very long time (~ 4 days) but retries every 120 seconds
+        total=_CONNECTION_RETRY_TOTAL,
+        backoff_factor=_CONNECTION_RETRY_BACKOFF_FACTOR,
+        # Any 4xx and 5xx statuses except
+        # 400 Bad Request
+        # 401 Unauthorized
+        # 403 Forbidden
+        # 404 Not Found
+        status_forcelist={
+            402,
+            *range(405, 600),
+        },
+        allowed_methods={
+            "POST",  # Default methods are idempotent, add POST here
+            *Retry.DEFAULT_ALLOWED_METHODS,
+        },
+    )
+
+
 def _configure_session() -> Session:
     """Configures the session for GET and POST requests.
 
     It enables a generous retrial strategy that waits for the application server to connect.
 
     """
-    retry_strategy = Retry(
-        # wait time between retries increases exponentially according to: backoff_factor * (2 ** (retry - 1))
-        total=_CONNECTION_RETRY_TOTAL,
-        backoff_factor=_CONNECTION_RETRY_BACKOFF_FACTOR,
-        status_forcelist=[429, 500, 502, 503, 504],
-    )
+    retry_strategy = create_retry_strategy()
     adapter = HTTPAdapter(max_retries=retry_strategy)
     http = requests.Session()
     http.mount("https://", adapter)
@@ -157,21 +175,7 @@ class HTTPClient:
         self, base_url: str, auth_token: Optional[str] = None, log_callback: Optional[Callable] = None
     ) -> None:
         self.base_url = base_url
-        retry_strategy = Retry(
-            # wait time between retries increases exponentially according to: backoff_factor * (2 ** (retry - 1))
-            # but the the maximum wait time is 120 secs. By setting a large value (2880), we'll make sure clients
-            # are going to be alive for a very long time (~ 4 days) but retries every 120 seconds
-            total=_CONNECTION_RETRY_TOTAL,
-            backoff_factor=_CONNECTION_RETRY_BACKOFF_FACTOR,
-            status_forcelist=[
-                408,  # Request Timeout
-                429,  # Too Many Requests
-                500,  # Internal Server Error
-                502,  # Bad Gateway
-                503,  # Service Unavailable
-                504,  # Gateway Timeout
-            ],
-        )
+        retry_strategy = create_retry_strategy()
         adapter = CustomRetryAdapter(max_retries=retry_strategy, timeout=_DEFAULT_REQUEST_TIMEOUT)
         self.session = requests.Session()
 

lightning-2.2.5/src/lightning/data/__init__.py

@@ -0,0 +1,41 @@
+import sys
+
+from lightning_utilities.core.imports import RequirementCache
+
+from lightning.data.processing.functions import map, optimize, walk
+from lightning.data.streaming.combined import CombinedStreamingDataset
+from lightning.data.streaming.dataloader import StreamingDataLoader
+from lightning.data.streaming.dataset import StreamingDataset
+
+__all__ = []
+
+if not RequirementCache("litdata"):
+    raise ModuleNotFoundError("Please, run `pip install litdata`")  # E111
+
+else:
+    import litdata
+
+    # Enable resolution at least for lower data namespace
+    sys.modules["lightning.data"] = litdata
+
+    from litdata.processing.functions import map, optimize, walk
+    from litdata.streaming.combined import CombinedStreamingDataset
+    from litdata.streaming.dataloader import StreamingDataLoader
+    from litdata.streaming.dataset import StreamingDataset
+
+    __all__ += [
+        "LightningDataset",
+        "StreamingDataset",
+        "CombinedStreamingDataset",
+        "StreamingDataLoader",
+        "LightningIterableDataset",
+        "map",
+        "optimize",
+        "walk",
+    ]
+
+# TODO: Move this to litdata
+if RequirementCache("lightning_sdk"):
+    from lightning_sdk import Machine
+
+    __all__ += ["Machine"]

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/fabric/CHANGELOG.md

@@ -5,6 +5,22 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 
+## [2.2.5] - 2024-05-23
+
+### Fixed
+
+- Fixed a matrix shape mismatch issue when running a model loaded from a quantized checkpoint (bitsandbytes) ([#19886](https://github.com/Lightning-AI/lightning/pull/19886))
+
+
+## [2.2.2] - 2024-04-11
+
+### Fixed
+
+- Fixed a KeyError when saving a FSDP sharded checkpoint and setting `save_weights_only=True` ([#19524](https://github.com/Lightning-AI/pytorch-lightning/pull/19524))
+- Fixed an issue causing a TypeError when using `torch.compile` as a decorator ([#19627](https://github.com/Lightning-AI/pytorch-lightning/pull/19627))
+- Fixed issue where some model methods couldn't be monkeypatched after being Fabric wrapped ([#19705](https://github.com/Lightning-AI/pytorch-lightning/pull/19705))
+- Fixed an issue causing weights to be reset in `Fabric.setup()` when using FSDP ([#19755](https://github.com/Lightning-AI/pytorch-lightning/pull/19755))
+
 ## [2.2.1] - 2024-03-04
 
 ### Fixed

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/fabric/plugins/precision/bitsandbytes.py

@@ -234,9 +234,9 @@ def _import_bitsandbytes() -> ModuleType:
             """Inplace quantize."""
             if weight is None:
                 weight = self.weight.data
-                if weight.data.type == torch.int8:
-                    # already quantized
-                    return
+            if weight.data.dtype == torch.int8:
+                # already quantized
+                return
             assert isinstance(self.weight, bnb.nn.Int8Params)
             self.weight = self.quantize(self.weight, weight, device)
 
@@ -318,9 +318,9 @@ def _import_bitsandbytes() -> ModuleType:
             """Inplace quantize."""
             if weight is None:
                 weight = self.weight.data
-                if weight.data.type == torch.uint8:
-                    # already quantized
-                    return
+            if weight.data.dtype == torch.uint8:
+                # already quantized
+                return
             assert isinstance(self.weight, bnb.nn.Params4bit)
             self.weight = self.quantize(self.weight, weight, device)
 

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/fabric/utilities/device_dtype_mixin.py

@@ -109,14 +109,12 @@ class _DeviceDtypeModuleMixin(Module):
 def _update_properties(
     root: torch.nn.Module, device: Optional[torch.device] = None, dtype: Optional[Union[str, torch.dtype]] = None
 ) -> None:
-    def apply_fn(module: Union[_DeviceDtypeModuleMixin, Module]) -> None:
+    for module in root.modules():
         if not isinstance(module, _DeviceDtypeModuleMixin):
-            return
+            continue
         # cannot use `module.to()` because we don't actually want to move the model in case there are multiple
         # devices types (such as partial meta parameters)
         if device is not None:
             module._device = device
         if dtype is not None:
             module._dtype = dtype
-
-    root.apply(apply_fn)

{lightning-2.2.1 → lightning-2.2.5}/src/lightning/fabric/utilities/logger.py

@@ -11,7 +11,7 @@
 # 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.
-
+import json
 from argparse import Namespace
 from dataclasses import asdict, is_dataclass
 from typing import Any, Dict, Mapping, MutableMapping, Optional, Union
@@ -132,6 +132,23 @@ def _sanitize_params(params: Dict[str, Any]) -> Dict[str, Any]:
     return params
 
 
+def _convert_json_serializable(params: Dict[str, Any]) -> Dict[str, Any]:
+    """Convert non-serializable objects in params to string."""
+    return {k: str(v) if not _is_json_serializable(v) else v for k, v in params.items()}
+
+
+def _is_json_serializable(value: Any) -> bool:
+    """Test whether a variable can be encoded as json."""
+    if value is None or isinstance(value, (bool, int, float, str, list, dict)):  # fast path
+        return True
+    try:
+        json.dumps(value)
+        return True
+    except (TypeError, OverflowError):
+        # OverflowError is raised if number is too large to encode
+        return False
+
+
 def _add_prefix(
     metrics: Mapping[str, Union[Tensor, float]], prefix: str, separator: str
 ) -> Mapping[str, Union[Tensor, float]]: