fix: make sure to add cpu to number

Signed-off-by: paperspace <29749331+aarnphm@users.noreply.github.com>

openllm-core/src/openllm_core/_configuration.py

@@ -3,7 +3,17 @@ import abc, inspect, logging, os, typing as t
 import inflection, orjson, pydantic
 from deepmerge.merger import Merger
 
-from ._typing_compat import DictStrAny, ListStr, LiteralSerialisation, NotRequired, Required, Self, TypedDict, overload
+from ._typing_compat import (
+  DictStrAny,
+  ListStr,
+  LiteralSerialisation,
+  NotRequired,
+  Required,
+  Self,
+  TypedDict,
+  overload,
+  Annotated,
+)
 from .exceptions import ForbiddenAttributeError, MissingDependencyError
 from .utils import field_env_key, first_not_none, is_vllm_available, is_transformers_available
 
@@ -223,6 +233,9 @@ class GenerationConfig(pydantic.BaseModel):
     None, description='Number of log probabilities to return per output token.'
   )
   detokenize: bool = pydantic.Field(True, description='Whether to detokenize the output.')
+  truncate_prompt_tokens: t.Optional[Annotated[int, pydantic.Field(ge=1)]] = pydantic.Field(
+    None, description='Truncate the prompt tokens.'
+  )
   prompt_logprobs: t.Optional[int] = pydantic.Field(
     None, description='Number of log probabilities to return per input token.'
   )

openllm-python/src/_openllm_tiny/_entrypoint.py

@@ -426,7 +426,7 @@ def build_command(
     labels = {'library': 'vllm'}
     service_config = dict(
       resources={
-        'gpu' if device else 'cpu': len(device) if device else 'cpu_count',
+        'gpu' if device else 'cpu': len(device) if device else '1',
         'gpu_type': recommended_instance_type(model_id, bentomodel),
       },
       traffic=dict(timeout=timeout),