Code
Hub
Workspaces
Following
Digest
Agents
Trending
Connect
MCP
copy
Index your code
hub
/
github.com/NVIDIA/TensorRT-LLM
/ max
Method
max
tensorrt_llm/functional.py:438–442 ·
view source on GitHub ↗
See functional.max.
(self, dim, keepdim=False)
Source
from the content-addressed store, hash-verified
436
return
mean(self, dim, keepdim)
437
438
def
max(self, dim, keepdim=False):
439
''
&
#x27;
440
See functional.max.
441
''
&
#x27;
442
return
max(self, dim, keepdim)
443
444
def
abs(self):
445
''
&
#x27;
Callers
15
assert_close
Function · 0.45
compute_actual_scale
Function · 0.45
nan_propagating_absmax_reduce
Function · 0.45
downcast_to_mxfp_torch
Function · 0.45
_compute_quant_and_scale
Function · 0.45
thread_local_absmax
Function · 0.45
ref.py
File · 0.45
is_close
Function · 0.45
compute
Function · 0.45
mha_ref
Function · 0.45
extract_stress_test_metrics
Function · 0.45
dump_stats
Function · 0.45
Calls
1
max
Function · 0.85
Tested by
15
assert_close
Function · 0.36
compute_actual_scale
Function · 0.36
extract_stress_test_metrics
Function · 0.36
get_kv_quant_scale
Method · 0.36
test_gpt_attention_IFB
Method · 0.36
test_smooth_quant_rms_norm
Method · 0.36
test_smooth_quant_layer_norm_layer
Method · 0.36
test_quantization_activation_scales
Method · 0.36
test_quantization_weight_scales
Method · 0.36
test_quantization_per_tensor_scales
Method · 0.36
test_quantization_dequantization_activation
Method · 0.36
test_quantization_dequantization_weight
Method · 0.36