skypilot-nightly 1.0.0.dev20250802__tar.gz → 1.0.0.dev20250806__tar.gz

{skypilot_nightly-1.0.0.dev20250802/skypilot_nightly.egg-info → skypilot_nightly-1.0.0.dev20250806}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: skypilot-nightly
-Version: 1.0.0.dev20250802
+Version: 1.0.0.dev20250806
 Summary: SkyPilot: Run AI on Any Infra — Unified, Faster, Cheaper.
 Author: SkyPilot Team
 License: Apache 2.0
@@ -57,6 +57,7 @@ Requires-Dist: pyjwt
 Requires-Dist: gitpython
 Requires-Dist: types-paramiko
 Requires-Dist: alembic
+Requires-Dist: aiohttp
 Provides-Extra: aws
 Requires-Dist: awscli>=1.27.10; extra == "aws"
 Requires-Dist: botocore>=1.29.10; extra == "aws"
@@ -132,6 +133,7 @@ Requires-Dist: casbin; extra == "server"
 Requires-Dist: sqlalchemy_adapter; extra == "server"
 Requires-Dist: passlib; extra == "server"
 Requires-Dist: pyjwt; extra == "server"
+Requires-Dist: aiohttp; extra == "server"
 Provides-Extra: all
 Requires-Dist: awscli>=1.27.10; extra == "all"
 Requires-Dist: botocore>=1.29.10; extra == "all"
@@ -186,6 +188,7 @@ Requires-Dist: casbin; extra == "all"
 Requires-Dist: sqlalchemy_adapter; extra == "all"
 Requires-Dist: passlib; extra == "all"
 Requires-Dist: pyjwt; extra == "all"
+Requires-Dist: aiohttp; extra == "all"
 Dynamic: author
 Dynamic: classifier
 Dynamic: description
@@ -235,6 +238,7 @@ Dynamic: summary
 ----
 
 :fire: *News* :fire:
+- [Aug 2025] Run and serve **OpenAI GPT-OSS models** (gpt-oss-120b, gpt-oss-20b) with one command on any infra: [**example**](./llm/gpt-oss/)
 - [Jul 2025] Run distributed **RL training for LLMs** with Verl (PPO, GRPO) on any cloud: [**example**](./llm/verl/)
 - [Jul 2025] 🎉 SkyPilot v0.10.0 released! [**blog post**](https://blog.skypilot.co/announcing-skypilot-0.10.0/), [**release notes**](https://github.com/skypilot-org/skypilot/releases/tag/v0.10.0)
 - [Jul 2025] Finetune **Llama4** on any distributed cluster/cloud: [**example**](./llm/llama-4-finetuning/)

{skypilot_nightly-1.0.0.dev20250802 → skypilot_nightly-1.0.0.dev20250806}/README.md

@@ -39,6 +39,7 @@
 ----
 
 :fire: *News* :fire:
+- [Aug 2025] Run and serve **OpenAI GPT-OSS models** (gpt-oss-120b, gpt-oss-20b) with one command on any infra: [**example**](./llm/gpt-oss/)
 - [Jul 2025] Run distributed **RL training for LLMs** with Verl (PPO, GRPO) on any cloud: [**example**](./llm/verl/)
 - [Jul 2025] 🎉 SkyPilot v0.10.0 released! [**blog post**](https://blog.skypilot.co/announcing-skypilot-0.10.0/), [**release notes**](https://github.com/skypilot-org/skypilot/releases/tag/v0.10.0)
 - [Jul 2025] Finetune **Llama4** on any distributed cluster/cloud: [**example**](./llm/llama-4-finetuning/)

{skypilot_nightly-1.0.0.dev20250802 → skypilot_nightly-1.0.0.dev20250806}/sky/__init__.py

@@ -5,7 +5,7 @@ from typing import Optional
 import urllib.request
 
 # Replaced with the current commit when building the wheels.
-_SKYPILOT_COMMIT_SHA = '6ddfb3cd7d476b51f9309c547338407d6eca092c'
+_SKYPILOT_COMMIT_SHA = '57784d97d8d63f3c87f4d5d22f3e820b10154241'
 
 
 def _get_git_commit():
@@ -35,7 +35,7 @@ def _get_git_commit():
 
 
 __commit__ = _get_git_commit()
-__version__ = '1.0.0.dev20250802'
+__version__ = '1.0.0.dev20250806'
 __root_dir__ = os.path.dirname(os.path.abspath(__file__))
 
 

{skypilot_nightly-1.0.0.dev20250802 → skypilot_nightly-1.0.0.dev20250806}/sky/backends/backend_utils.py

@@ -923,7 +923,10 @@ def write_cluster_config(
             cluster_config_overrides=cluster_config_overrides,
             cloud=cloud,
             context=region.name)
-        kubernetes_utils.combine_metadata_fields(tmp_yaml_path, region.name)
+        kubernetes_utils.combine_metadata_fields(
+            tmp_yaml_path,
+            cluster_config_overrides=cluster_config_overrides,
+            context=region.name)
         yaml_obj = common_utils.read_yaml(tmp_yaml_path)
         pod_config: Dict[str, Any] = yaml_obj['available_node_types'][
             'ray_head_default']['node_config']

{skypilot_nightly-1.0.0.dev20250802 → skypilot_nightly-1.0.0.dev20250806}/sky/catalog/kubernetes_catalog.py

@@ -255,6 +255,14 @@ def _list_accelerators(
                     # Get all the pods running on the node
                     if (pod.spec.node_name == node.metadata.name and
                             pod.status.phase in ['Running', 'Pending']):
+                        # Skip pods that should not count against GPU count
+                        if (kubernetes_utils.
+                                should_exclude_pod_from_gpu_allocation(pod)):
+                            logger.debug(
+                                f'Excluding pod '
+                                f'{pod.metadata.name} from GPU count '
+                                f'calculations on node {node.metadata.name}')
+                            continue
                         # Iterate over all the containers in the pod and sum
                         # the GPU requests
                         for container in pod.spec.containers:

{skypilot_nightly-1.0.0.dev20250802 → skypilot_nightly-1.0.0.dev20250806}/sky/catalog/nebius_catalog.py

@@ -38,7 +38,6 @@ def get_hourly_cost(instance_type: str,
                     region: Optional[str] = None,
                     zone: Optional[str] = None) -> float:
     """Returns the cost, or the cheapest cost among all zones for spot."""
-    assert not use_spot, 'Nebius does not support spot.'
     if zone is not None:
         with ux_utils.print_exception_no_traceback():
             raise ValueError('Nebius does not support zones.')

{skypilot_nightly-1.0.0.dev20250802 → skypilot_nightly-1.0.0.dev20250806}/sky/client/cli/command.py

@@ -62,7 +62,6 @@ from sky.provision.kubernetes import constants as kubernetes_constants
 from sky.provision.kubernetes import utils as kubernetes_utils
 from sky.server import common as server_common
 from sky.server import constants as server_constants
-from sky.server import versions
 from sky.server.requests import requests
 from sky.skylet import autostop_lib
 from sky.skylet import constants
@@ -557,17 +556,17 @@ def _parse_override_params(
     """Parses the override parameters into a dictionary."""
     override_params: Dict[str, Any] = {}
     if cloud is not None:
-        if cloud.lower() == 'none':
+        if cloud.lower() == 'none' or cloud == '*':
             override_params['cloud'] = None
         else:
             override_params['cloud'] = registry.CLOUD_REGISTRY.from_str(cloud)
     if region is not None:
-        if region.lower() == 'none':
+        if region.lower() == 'none' or region == '*':
             override_params['region'] = None
         else:
             override_params['region'] = region
     if zone is not None:
-        if zone.lower() == 'none':
+        if zone.lower() == 'none' or zone == '*':
             override_params['zone'] = None
         else:
             override_params['zone'] = zone
@@ -965,9 +964,10 @@ def _handle_infra_cloud_region_zone_options(infra: Optional[str],
 
     if infra is not None:
         infra_info = infra_utils.InfraInfo.from_str(infra)
-        cloud = infra_info.cloud
-        region = infra_info.region
-        zone = infra_info.zone
+        # Convert None to '*' to ensure proper override behavior
+        cloud = infra_info.cloud if infra_info.cloud is not None else '*'
+        region = infra_info.region if infra_info.region is not None else '*'
+        zone = infra_info.zone if infra_info.zone is not None else '*'
     return cloud, region, zone
 
 
@@ -1832,9 +1832,6 @@ def status(verbose: bool, refresh: bool, ip: bool, endpoints: bool,
     show_endpoints = endpoints or endpoint is not None
     show_single_endpoint = endpoint is not None
     show_services = show_services and not any([clusters, ip, endpoints])
-    remote_api_version = versions.get_remote_api_version()
-    if remote_api_version is None or remote_api_version < 12:
-        show_pools = False
 
     query_clusters: Optional[List[str]] = None if not clusters else clusters
     refresh_mode = common.StatusRefreshMode.NONE
@@ -1886,7 +1883,11 @@ def status(verbose: bool, refresh: bool, ip: bool, endpoints: bool,
         return serve_lib.status(service_names=None)
 
     def submit_pools() -> Optional[str]:
-        return managed_jobs.pool_status(pool_names=None)
+        try:
+            return managed_jobs.pool_status(pool_names=None)
+        except exceptions.APINotSupportedError as e:
+            logger.debug(f'Pools are not supported in the remote server: {e}')
+            return None
 
     def submit_workspace() -> Optional[str]:
         try:
@@ -2009,7 +2010,7 @@ def status(verbose: bool, refresh: bool, ip: bool, endpoints: bool,
                 controller_utils.Controllers.JOBS_CONTROLLER.value.
                 in_progress_hint(False).format(job_info=job_info))
 
-    if show_pools:
+    if show_pools and pool_status_request_id:
         num_pools = None
         if managed_jobs_query_interrupted:
             msg = 'KeyboardInterrupt'
@@ -3462,6 +3463,17 @@ def show_gpus(
                                                                region,
                                                                zone=None)
 
+    # cloud and region could be '*' from _handle_infra_cloud_region_zone_options
+    # which normally indicates to
+    # _make_task_or_dag_from_entrypoint_with_overrides -> _parse_override_params
+    # to disregard the cloud and region from the YAML.
+    # In show_gpus, there is no YAML, so we need to handle the '*' value
+    # directly here. We should use None instead to indicate "any".
+    if cloud == '*':
+        cloud = None
+    if region == '*':
+        region = None
+
     # validation for the --region flag
     if region is not None and cloud is None:
         raise click.UsageError(
@@ -3502,7 +3514,14 @@ def show_gpus(
                               (cloud_name is None or cloud_is_ssh))
 
     def _list_to_str(lst):
-        return ', '.join([str(e) for e in lst])
+
+        def format_number(n):
+            # If it's a float that's a whole number, display as int
+            if isinstance(n, float) and n.is_integer():
+                return str(int(n))
+            return str(n)
+
+        return ', '.join([format_number(n) for n in lst])
 
     # TODO(zhwu,romilb): We should move most of these kubernetes related
     # queries into the backend, especially behind the server.

{skypilot_nightly-1.0.0.dev20250802 → skypilot_nightly-1.0.0.dev20250806}/sky/client/sdk.py

@@ -88,14 +88,17 @@ def reload_config() -> None:
     skypilot_config.safe_reload_config()
 
 
-def stream_response(request_id: Optional[str],
+def stream_response(request_id: Optional[server_common.RequestId],
                     response: 'requests.Response',
                     output_stream: Optional['io.TextIOBase'] = None,
                     resumable: bool = False) -> Any:
     """Streams the response to the console.
 
     Args:
-        request_id: The request ID.
+        request_id: The request ID of the request to stream. May be a full
+            request ID or a prefix.
+            If None, the latest request submitted to the API server is streamed.
+            Using None request_id is not recommended in multi-user environments.
         response: The HTTP response.
         output_stream: The output stream to write to. If None, print to the
             console.
@@ -1756,7 +1759,7 @@ def status_kubernetes() -> server_common.RequestId:
 # === API request APIs ===
 @usage_lib.entrypoint
 @annotations.client_api
-def get(request_id: str) -> Any:
+def get(request_id: server_common.RequestId) -> Any:
     """Waits for and gets the result of a request.
 
     This function will not check the server health since /api/get is typically
@@ -1764,7 +1767,8 @@ def get(request_id: str) -> Any:
     may cause GET /api/get being sent to a restarted API server.
 
     Args:
-        request_id: The request ID of the request to get.
+        request_id: The request ID of the request to get. May be a full request
+            ID or a prefix.
 
     Returns:
         The ``Request Returns`` of the specified request. See the documentation
@@ -1818,7 +1822,7 @@ def get(request_id: str) -> Any:
 @server_common.check_server_healthy_or_start
 @annotations.client_api
 def stream_and_get(
-    request_id: Optional[str] = None,
+    request_id: Optional[server_common.RequestId] = None,
     log_path: Optional[str] = None,
     tail: Optional[int] = None,
     follow: bool = True,
@@ -1830,7 +1834,10 @@ def stream_and_get(
     prefix of the full request id.
 
     Args:
-        request_id: The prefix of the request ID of the request to stream.
+        request_id: The request ID of the request to stream. May be a full
+            request ID or a prefix.
+            If None, the latest request submitted to the API server is streamed.
+            Using None request_id is not recommended in multi-user environments.
         log_path: The path to the log file to stream.
         tail: The number of lines to show from the end of the logs.
             If None, show all logs.
@@ -1873,7 +1880,8 @@ def stream_and_get(
 
 @usage_lib.entrypoint
 @annotations.client_api
-def api_cancel(request_ids: Optional[Union[str, List[str]]] = None,
+def api_cancel(request_ids: Optional[Union[
+    server_common.RequestId, List[server_common.RequestId]]] = None,
                all_users: bool = False,
                silent: bool = False) -> server_common.RequestId:
     """Aborts a request or all requests.
@@ -1938,7 +1946,7 @@ def _local_api_server_running(kill: bool = False) -> bool:
 @usage_lib.entrypoint
 @annotations.client_api
 def api_status(
-    request_ids: Optional[List[str]] = None,
+    request_ids: Optional[List[server_common.RequestId]] = None,
     # pylint: disable=redefined-builtin
     all_status: bool = False
 ) -> List[payloads.RequestPayload]:

{skypilot_nightly-1.0.0.dev20250802 → skypilot_nightly-1.0.0.dev20250806}/sky/client/sdk.pyi

@@ -38,7 +38,7 @@ def reload_config() -> None:
     ...
 
 
-def stream_response(request_id: Optional[str],
+def stream_response(request_id: Optional[server_common.RequestId],
                     response: requests.Response,
                     output_stream: Optional['io.TextIOBase'] = ...,
                     resumable: bool = ...) -> Any:
@@ -249,11 +249,11 @@ def status_kubernetes() -> server_common.RequestId:
     ...
 
 
-def get(request_id: str) -> Any:
+def get(request_id: server_common.RequestId) -> Any:
     ...
 
 
-def stream_and_get(request_id: Optional[str] = ...,
+def stream_and_get(request_id: Optional[server_common.RequestId] = ...,
                    log_path: Optional[str] = ...,
                    tail: Optional[int] = ...,
                    follow: bool = ...,
@@ -261,13 +261,14 @@ def stream_and_get(request_id: Optional[str] = ...,
     ...
 
 
-def api_cancel(request_ids: Optional[Union[str, List[str]]] = ...,
+def api_cancel(request_ids: Optional[Union[
+    server_common.RequestId, List[server_common.RequestId]]] = ...,
                all_users: bool = ...,
                silent: bool = ...) -> server_common.RequestId:
     ...
 
 
-def api_status(request_ids: Optional[List[str]] = ...,
+def api_status(request_ids: Optional[List[server_common.RequestId]] = ...,
                all_status: bool = ...) -> List[payloads.RequestPayload]:
     ...