[1 / N] Clean up logprob utils (#15509)

python/sglang/srt/speculative/eagle_worker.py

@@ -14,7 +14,7 @@ from sglang.srt.layers.moe.utils import (
     speculative_moe_a2a_backend_context,
     speculative_moe_backend_context,
 )
-from sglang.srt.layers.sampler import get_token_ids_logprobs, get_top_logprobs
+from sglang.srt.layers.utils.logprob import get_token_ids_logprobs, get_top_logprobs
 from sglang.srt.managers.io_struct import UpdateWeightsFromTensorReqInput
 from sglang.srt.managers.schedule_batch import ScheduleBatch
 from sglang.srt.managers.scheduler import GenerationBatchResult

python/sglang/srt/speculative/mtp_worker.py

@@ -22,7 +22,7 @@ from sglang.srt.distributed import get_tp_group
 from sglang.srt.layers.dp_attention import get_attention_tp_group
 from sglang.srt.layers.logits_processor import LogitsProcessorOutput
 from sglang.srt.layers.moe.utils import speculative_moe_backend_context
-from sglang.srt.layers.sampler import get_token_ids_logprobs, get_top_logprobs
+from sglang.srt.layers.utils.logprob import get_token_ids_logprobs, get_top_logprobs
 from sglang.srt.managers.schedule_batch import ScheduleBatch
 from sglang.srt.managers.scheduler import GenerationBatchResult
 from sglang.srt.managers.tp_worker import TpModelWorker

python/sglang/srt/speculative/ngram_worker.py

@@ -7,7 +7,7 @@ from sgl_kernel.speculative import reconstruct_indices_from_tree_mask
 
 from sglang.srt.environ import envs
 from sglang.srt.layers.logits_processor import LogitsProcessorOutput
-from sglang.srt.layers.sampler import get_token_ids_logprobs, get_top_logprobs
+from sglang.srt.layers.utils.logprob import get_token_ids_logprobs, get_top_logprobs
 from sglang.srt.managers.schedule_batch import ScheduleBatch
 from sglang.srt.managers.scheduler import GenerationBatchResult
 from sglang.srt.managers.tp_worker import TpModelWorker