format

Author: V2arK

examples/sdk/create_cserve.py

@@ -36,9 +36,7 @@ def get_default_cserve_config(cclient, name, model):
 def main():
     with get_centml_client() as cclient:
         ### Get the configurations for the Qwen model
-        qwen_config = get_fastest_cserve_config(
-            cclient, name="qwen-fastest", model="Qwen/Qwen2-VL-7B-Instruct"
-        )
+        qwen_config = get_fastest_cserve_config(cclient, name="qwen-fastest", model="Qwen/Qwen2-VL-7B-Instruct")
         # qwen_config = get_default_cserve_config(cclient, name="qwen-default", model="Qwen/Qwen2-VL-7B-Instruct")
 
         ### Modify the recipe if necessary

examples/sdk/create_inference.py

@@ -38,5 +38,6 @@ def main():
         cclient.delete(deployment.id)
         '''
 
+
 if __name__ == "__main__":
     main()

examples/sdk/get_deployment_usage.py

@@ -3,7 +3,6 @@
 from centml.sdk.api import get_centml_client
 from centml.sdk import Metric
 
-
 HOUR_IN_SECONDS = 60 * 60
 DAY_IN_SECONDS = 24 * HOUR_IN_SECONDS
 MAX_DATA_POINTS = 10_000
@@ -17,22 +16,23 @@ def get_step_size(start_time_in_seconds: int, end_time_in_seconds: int) -> int:
     # 2 days to 7 days: 5m
     elif time_delta_in_seconds <= 7 * DAY_IN_SECONDS:
         return 5 * 60
-	#  7 days to 14 days: 10m
+    #  7 days to 14 days: 10m
     elif time_delta_in_seconds <= 14 * DAY_IN_SECONDS:
         return 10 * 60
-	# 14 days to 30 days: 30m
+    # 14 days to 30 days: 30m
     elif time_delta_in_seconds <= 30 * DAY_IN_SECONDS:
         return 30 * 60
-	# 30 days to 60 days: 1 hour
+    # 30 days to 60 days: 1 hour
     elif time_delta_in_seconds <= 60 * DAY_IN_SECONDS:
         return HOUR_IN_SECONDS
-	# 60 days to 90 days: 2 hours
+    # 60 days to 90 days: 2 hours
     elif time_delta_in_seconds <= 90 * DAY_IN_SECONDS:
         return 2 * HOUR_IN_SECONDS
-	# 90+ days: 3 hours (This catches all ranges greater than 90 days)
+    # 90+ days: 3 hours (This catches all ranges greater than 90 days)
     else:
         return 3 * HOUR_IN_SECONDS
 
+
 def main():
     with get_centml_client() as cclient:
         start_time = 1752084581

scripts/data_collection.py

@@ -7,11 +7,10 @@
     AutoModelForCausalLM,
     AutoTokenizer,
     AutoModelForImageClassification,
-    AutoModelForObjectDetection
+    AutoModelForObjectDetection,
 )
 
 
-
 from centml.compiler.prediction.kdtree import KDTreeWithValues
 from centml.compiler.prediction.profiler import Profiler
 from scripts.timer import timed
@@ -26,8 +25,8 @@
 # Different HuggingFace Models + Different Input Sizes
 llm_tests = [
     ("google/gemma-7b", (1, 128)),
-    ("microsoft/phi-2", (1,512)),
-    ("microsoft/phi-2", (2,512)),
+    ("microsoft/phi-2", (1, 512)),
+    ("microsoft/phi-2", (2, 512)),
     ("facebook/bart-large", (1, 1024)),
     ("facebook/bart-large", (2, 512)),
     ("gpt2-xl", (1, 1024)),
@@ -212,6 +211,7 @@ def image_classification_test(model_name, batch_size, custom_backend):
     gc.collect()
     torch.cuda.empty_cache()
 
+
 def object_detection_test(model_name, batch_size, custom_backend):
     global cuda_kernel_time
     global actual_time
@@ -240,6 +240,7 @@ def object_detection_test(model_name, batch_size, custom_backend):
     gc.collect()
     torch.cuda.empty_cache()
 
+
 # for model_name, input_size in large_llm_tests:
 #     llm_test(model_name, input_size, custom_backend)
 

scripts/timer.py

@@ -1,5 +1,6 @@
 import torch
 
+
 def timed(fn):
     start = torch.cuda.Event(enable_timing=True)
     end = torch.cuda.Event(enable_timing=True)

scripts/user_vault/get_vault_items.py

@@ -15,15 +15,11 @@
 from platform_api_python_client import UserVaultType
 
 
-def get_vault_items(
-    vault_type: Optional[UserVaultType] = None,
-    search_query: Optional[str] = None,
-):
+def get_vault_items(vault_type: Optional[UserVaultType] = None, search_query: Optional[str] = None):
     """Retrieve items from user's vault."""
     with get_centml_client() as client:
         response = client._api.get_all_user_vault_items_endpoint_user_vault_get(
-            type=vault_type,
-            search_query=search_query,
+            type=vault_type, search_query=search_query
         )
         return response.results
 
@@ -65,18 +61,8 @@ def display_vault_items(items, show_values: bool = False):
     type=click.Choice([t.value for t in UserVaultType], case_sensitive=False),
     help="Filter by vault type (env_vars, ssh_keys, bearer_tokens, access_tokens, certificates)",
 )
-@click.option(
-    "--search",
-    "search_query",
-    type=str,
-    help="Search query to filter items by key",
-)
-@click.option(
-    "--show-values",
-    is_flag=True,
-    default=False,
-    help="Show vault item values",
-)
+@click.option("--search", "search_query", type=str, help="Search query to filter items by key")
+@click.option("--show-values", is_flag=True, default=False, help="Show vault item values")
 def main(vault_type: Optional[str], search_query: Optional[str], show_values: bool):
     """Retrieve all items from user's vault (secrets).
 
@@ -99,10 +85,7 @@ def main(vault_type: Optional[str], search_query: Optional[str], show_values: bo
     """
     type_enum = UserVaultType(vault_type) if vault_type else None
 
-    items = get_vault_items(
-        vault_type=type_enum,
-        search_query=search_query,
-    )
+    items = get_vault_items(vault_type=type_enum, search_query=search_query)
 
     display_vault_items(items, show_values=show_values)