mirror of
https://github.com/huggingface/text-generation-inference.git
synced 2025-04-19 22:02:06 +00:00
a6a0c97ed9
* wip * rollback * refactor to use prefix/postfix namming + fix all_input_ids_tensor * maybe patching vlms? * fix filter and concat * wip, no filter, no concat * current * add prepare_for_prefill * working * load tested * re-create slots * re-create slots * fix slot_filtering_indices * feedback loop * remove log * fix benchmarker * fix vlm and seq2seq * rename to cache and input lengths * fix prefill logprobs * fix launcher * fix logprobs? * idk at this point * max input length * omfg * remove debugging lines * fix tests * fix mllama * fix cargo tests * remove support chunking for paged * Fixing non blocked attentions * Fixing dtype + AMD, Ipex targets. * lint fix. * rename * Fix prefix_caching variable, remove defaults in server (confusing a lot of the times). * Add simple resolution when user specifies ATTENTION=paged. * Put back non default simple tests. * Fix env name --------- Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
41 lines
1.2 KiB
Python
41 lines
1.2 KiB
Python
import torch
|
|
import grpc
|
|
|
|
from google.rpc import status_pb2, code_pb2
|
|
from grpc_status import rpc_status
|
|
from grpc_interceptor.server import AsyncServerInterceptor
|
|
from loguru import logger
|
|
from typing import Callable, Any
|
|
|
|
|
|
class ExceptionInterceptor(AsyncServerInterceptor):
|
|
def __init__(self, shutdown_callback):
|
|
self.shutdown_callback = shutdown_callback
|
|
|
|
async def intercept(
|
|
self,
|
|
method: Callable,
|
|
request_or_iterator: Any,
|
|
context: grpc.ServicerContext,
|
|
method_name: str,
|
|
) -> Any:
|
|
try:
|
|
response = method(request_or_iterator, context)
|
|
return await response
|
|
except Exception as err:
|
|
method_name = method_name.split("/")[-1]
|
|
logger.exception(f"Method {method_name} encountered an error.")
|
|
|
|
# Runtime Error cannot be recovered from
|
|
if isinstance(err, RuntimeError):
|
|
self.shutdown_callback()
|
|
|
|
if torch.cuda.is_available():
|
|
torch.cuda.empty_cache()
|
|
|
|
await context.abort_with_status(
|
|
rpc_status.to_status(
|
|
status_pb2.Status(code=code_pb2.INTERNAL, message=str(err))
|
|
)
|
|
)
|