MCPcopy
hub / github.com/NVIDIA/TensorRT-LLM / current_stream

Method current_stream

tensorrt_llm/python_plugin.py:319–323  ·  view source on GitHub ↗
(self)

Source from the content-addressed store, hash-verified

317
318 @property
319 def current_stream(self):
320 if self._current_stream == -1:
321 return torch.cuda.current_stream().cuda_stream
322 else:
323 return self._current_stream
324
325 @current_stream.setter
326 def current_stream(self, stream: int):

Callers 15

fmha_bmark.pyFile · 0.80
_construct_executionMethod · 0.80
test_plugin_no_cacheMethod · 0.80
test_fp4_gemmMethod · 0.80
test_mamba_conv1dMethod · 0.80
test_nvfp4_qdqMethod · 0.80
test_nvfp4_gemm_ootbMethod · 0.80
test_rg_lruMethod · 0.80
test_selective_scanMethod · 0.80

Calls

no outgoing calls

Tested by 15

_construct_executionMethod · 0.64
test_plugin_no_cacheMethod · 0.64
test_fp4_gemmMethod · 0.64
test_mamba_conv1dMethod · 0.64
test_nvfp4_qdqMethod · 0.64
test_nvfp4_gemm_ootbMethod · 0.64
test_rg_lruMethod · 0.64
test_selective_scanMethod · 0.64