move files

Signed-off-by: Sage Moore <[email protected]>

Author: SageMoore

vllm/forward_context.py

@@ -13,7 +13,7 @@
 import vllm.envs as envs
 from vllm.config import CUDAGraphMode, ParallelConfig, VllmConfig
 from vllm.logger import init_logger
-from vllm.compilation.ubatch_utils import (UbatchSlice, UBatchSlices)
+from vllm.v1.worker.ubatch_utils import (UbatchSlice, UBatchSlices)
 
 if TYPE_CHECKING:
     from vllm.attention.backends.abstract import AttentionMetadata

vllm/v1/attention/backends/utils.py

@@ -29,7 +29,7 @@
 from vllm.logger import init_logger
 from vllm.v1.kv_cache_interface import AttentionSpec
 
-from vllm.compilation.ubatch_utils import UbatchSlice
+from vllm.v1.worker.ubatch_utils import UbatchSlice
 
 logger = init_logger(__name__)
 _KV_CACHE_LAYOUT_OVERRIDE = None

vllm/v1/worker/gpu_model_runner.py

@@ -23,8 +23,8 @@
 from vllm.compilation.counter import compilation_counter
 from vllm.compilation.cuda_graph import CUDAGraphWrapper
 from vllm.compilation.monitor import set_cudagraph_capturing_enabled
-from vllm.compilation.ubatch_utils import UbatchSlice, UBatchSlices
-from vllm.compilation.ubatch_wrapper import UBatchWrapper
+from vllm.v1.worker.ubatch_utils import UbatchSlice, UBatchSlices
+from vllm.v1.worker.gpu_ubatch_wrapper import UBatchWrapper
 from vllm.config import (CompilationLevel, CUDAGraphMode, VllmConfig,
                          get_layers_from_vllm_config, update_config)
 from vllm.distributed.eplb.eplb_state import EplbState