speedup_experiment_fx2trt() — pytorch Function Reference

Architecture documentation for the speedup_experiment_fx2trt() function in common.py from the pytorch codebase.

Function python DynamoBenchmarks Configuration calls 1

Entity Profile

DynamoBenchmarks→ Configuration→ speedup_experiment_fx2trt() — pytorch Function Reference

Dependency Diagram

graph TD
  19aa986b_7c81_6518_3e84_45819cd8e90d["speedup_experiment_fx2trt()"]
  04a3a4a6_8db3_854d_a893_02c9542bf9dd["speedup_experiment()"]
  19aa986b_7c81_6518_3e84_45819cd8e90d -->|calls| 04a3a4a6_8db3_854d_a893_02c9542bf9dd
  style 19aa986b_7c81_6518_3e84_45819cd8e90d fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

benchmarks/dynamo/common.py lines 842–848

def speedup_experiment_fx2trt(args, model_iter_fn, model, example_inputs):
    """
    Measure speedups over eager using the trt inference backend. TRT backend is based fx graph
    generated by torch._dynamo.
    Writes to ./speedups_fx2trt.csv
    """
    return speedup_experiment(args, model_iter_fn, model, example_inputs)