squall/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

[Bugfix][TPU] Add prompt adapter methods to TPUExecutor (#6279)

Browse Source
This commit is contained in:
Woosuk Kwon 2024-07-09 19:30:56 -07:00 committed by GitHub
parent da78caecfa
commit 5ed3505d82
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 25 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
vllm/executor/tpu_executor.py
View File

@ -81,8 +81,7 @@ class TPUExecutor(ExecutorBase):
def determine_num_available_blocks(self) -> Tuple[int, int]: def determine_num_available_blocks(self) -> Tuple[int, int]:
"""Determine the number of available KV blocks by invoking the """Determine the number of available KV blocks by invoking the
underlying worker. underlying worker."""
"""
return self.driver_worker.determine_num_available_blocks() return self.driver_worker.determine_num_available_blocks()
def execute_model( def execute_model(
@ -93,16 +92,36 @@ class TPUExecutor(ExecutorBase):
return output return output
def add_lora(self, lora_request: LoRARequest) -> bool: def add_lora(self, lora_request: LoRARequest) -> bool:
raise NotImplementedError("LoRA is not implemented for TPU backend.") raise NotImplementedError(
"LoRA is currently not supported by the TPU backend.")
def remove_lora(self, lora_id: int) -> bool: def remove_lora(self, lora_id: int) -> bool:
raise NotImplementedError("LoRA is not implemented for TPU backend.") raise NotImplementedError(
"LoRA is currently not supported by the TPU backend.")
def pin_lora(self, lora_id: int) -> bool: def pin_lora(self, lora_id: int) -> bool:
raise NotImplementedError("LoRA is not implemented for TPU backend.") raise NotImplementedError(
"LoRA is currently not supported by the TPU backend.")
def list_loras(self) -> Set[int]: def list_loras(self) -> Set[int]:
raise NotImplementedError("LoRA is not implemented for TPU backend.") raise NotImplementedError(
"LoRA is currently not supported by the TPU backend.")
def add_prompt_adapter(self, prompt_adapter_request) -> bool:
raise NotImplementedError(
"Soft prompt is currently not supported by the TPU backend.")
def remove_prompt_adapter(self, prompt_adapter_id: int) -> bool:
raise NotImplementedError(
"Soft prompt is currently not supported by the TPU backend.")
def pin_prompt_adapter(self, prompt_adapter_id: int) -> bool:
raise NotImplementedError(
"Soft prompt is currently not supported by the TPU backend.")
def list_prompt_adapters(self) -> Set[int]:
raise NotImplementedError(
"Soft prompt is currently not supported by the TPU backend.")
def check_health(self) -> None: def check_health(self) -> None:
# TPUExecutor will always be healthy as long as it's running. # TPUExecutor will always be healthy as long as it's running.