20231088/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Minor

Browse Source
This commit is contained in:
Woosuk Kwon 2023-03-26 08:00:39 +00:00
parent 2f49f15585
commit d359cda5fa
3 changed files with 5 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
cacheflow/master/block_manager.py
View File

@ -7,7 +7,7 @@ from cacheflow.sequence import SequenceStatus
from cacheflow.utils import Device
class BlockManager:
class BlockAllocator:
def __init__(
self,
@ -65,8 +65,8 @@ class BlockSpaceManager:
self.num_total_gpu_blocks = num_gpu_blocks
self.num_total_cpu_blocks = num_cpu_blocks
self.gpu_allocator = BlockManager(Device.GPU, block_size, num_gpu_blocks)
self.cpu_allocator = BlockManager(Device.CPU, block_size, num_cpu_blocks)
self.gpu_allocator = BlockAllocator(Device.GPU, block_size, num_gpu_blocks)
self.cpu_allocator = BlockAllocator(Device.CPU, block_size, num_cpu_blocks)
# Mapping: seq_id -> BlockTable.
self.block_tables: Dict[int, BlockTable] = {}

1
cacheflow/models/sample.py
View File

@ -8,6 +8,7 @@ from cacheflow.sampling_params import SamplingParams
from cacheflow.sequence import SequenceOutputs
from cacheflow.parallel_utils.tensor_parallel import gather_from_tensor_model_parallel_region
class Sampler(nn.Module):
def __init__(self) -> None:

2
cacheflow/sequence.py
View File

@ -30,7 +30,7 @@ class Sequence:
self.status = SequenceStatus.PENDING
self.output_logprobs: List[Dict[int, float]] = []
self.cumulative_logprobs = 1.0
self.cumulative_logprobs = 0.0
def add_block(self) -> None:
block = LogicalTokenBlock(