MCPcopy
hub / github.com/NVIDIA/TensorRT-LLM / has_nvfp4

Method has_nvfp4

tensorrt_llm/quantization/mode.py:183–184  ·  view source on GitHub ↗
(self)

Source from the content-addressed store, hash-verified

181 return self._any(self.FP8_ROWWISE)
182
183 def has_nvfp4(self):
184 return self._any(self.NVFP4)
185
186 def has_w4a8_nvfp4_fp8(self):
187 return self._any(self.W4A8_NVFP4_FP8)

Callers 15

test_mlp_comparisonMethod · 0.95
test_ootb_comparisonMethod · 0.95
to_dictMethod · 0.95
MLPMethod · 0.45
set_weight_layerMethod · 0.45
create_trt_sessionMethod · 0.45
buildFunction · 0.45
_load_model_from_hfMethod · 0.45
fp4_quantizeFunction · 0.45
quantizeFunction · 0.45
_moe_pluginFunction · 0.45
__init__Method · 0.45

Calls 1

_anyMethod · 0.95

Tested by 5

test_mlp_comparisonMethod · 0.76
test_ootb_comparisonMethod · 0.76
MLPMethod · 0.36
set_weight_layerMethod · 0.36
create_trt_sessionMethod · 0.36