Add inline evaluation results for summary evaluators #725
baskaryanAnnotations
1 warning and 2 notices
|
benchmark
ubuntu-latest pipelines will use ubuntu-24.04 soon. For more details, see https://github.com/actions/runner-images/issues/10636
|
|
Benchmark results:
python/langsmith/evaluation/_arunner.py#L1
...........
WARNING: the benchmark result may be unstable
* the standard deviation (111 ms) is 16% of the mean (700 ms)
Try to rerun the benchmark with more runs, values and/or loops.
Run 'python -m pyperf system tune' command to reduce the system jitter.
Use pyperf stats, pyperf dump and pyperf hist to analyze results.
Use --quiet option to hide these warnings.
create_5_000_run_trees: Mean +- std dev: 700 ms +- 111 ms
...........
WARNING: the benchmark result may be unstable
* the standard deviation (202 ms) is 14% of the mean (1.40 sec)
Try to rerun the benchmark with more runs, values and/or loops.
Run 'python -m pyperf system tune' command to reduce the system jitter.
Use pyperf stats, pyperf dump and pyperf hist to analyze results.
Use --quiet option to hide these warnings.
create_10_000_run_trees: Mean +- std dev: 1.40 sec +- 0.20 sec
...........
WARNING: the benchmark result may be unstable
* the standard deviation (199 ms) is 14% of the mean (1.45 sec)
Try to rerun the benchmark with more runs, values and/or loops.
Run 'python -m pyperf system tune' command to reduce the system jitter.
Use pyperf stats, pyperf dump and pyperf hist to analyze results.
Use --quiet option to hide these warnings.
create_20_000_run_trees: Mean +- std dev: 1.45 sec +- 0.20 sec
...........
dumps_class_nested_py_branch_and_leaf_200x400: Mean +- std dev: 691 us +- 11 us
...........
dumps_class_nested_py_leaf_50x100: Mean +- std dev: 24.9 ms +- 0.2 ms
...........
dumps_class_nested_py_leaf_100x200: Mean +- std dev: 104 ms +- 2 ms
...........
dumps_dataclass_nested_50x100: Mean +- std dev: 25.2 ms +- 0.2 ms
...........
WARNING: the benchmark result may be unstable
* the standard deviation (15.5 ms) is 22% of the mean (70.3 ms)
Try to rerun the benchmark with more runs, values and/or loops.
Run 'python -m pyperf system tune' command to reduce the system jitter.
Use pyperf stats, pyperf dump and pyperf hist to analyze results.
Use --quiet option to hide these warnings.
dumps_pydantic_nested_50x100: Mean +- std dev: 70.3 ms +- 15.5 ms
...........
dumps_pydanticv1_nested_50x100: Mean +- std dev: 194 ms +- 3 ms
|
|
Comparison against main:
python/langsmith/evaluation/_arunner.py#L1
+-----------------------------------------------+----------+------------------------+
| Benchmark | main | changes |
+===============================================+==========+========================+
| dumps_pydanticv1_nested_50x100 | 217 ms | 194 ms: 1.12x faster |
+-----------------------------------------------+----------+------------------------+
| create_5_000_run_trees | 717 ms | 700 ms: 1.02x faster |
+-----------------------------------------------+----------+------------------------+
| dumps_dataclass_nested_50x100 | 25.5 ms | 25.2 ms: 1.01x faster |
+-----------------------------------------------+----------+------------------------+
| dumps_class_nested_py_leaf_50x100 | 25.2 ms | 24.9 ms: 1.01x faster |
+-----------------------------------------------+----------+------------------------+
| dumps_class_nested_py_leaf_100x200 | 104 ms | 104 ms: 1.01x faster |
+-----------------------------------------------+----------+------------------------+
| dumps_class_nested_py_branch_and_leaf_200x400 | 691 us | 691 us: 1.00x faster |
+-----------------------------------------------+----------+------------------------+
| create_10_000_run_trees | 1.38 sec | 1.40 sec: 1.01x slower |
+-----------------------------------------------+----------+------------------------+
| create_20_000_run_trees | 1.40 sec | 1.45 sec: 1.03x slower |
+-----------------------------------------------+----------+------------------------+
| dumps_pydantic_nested_50x100 | 66.1 ms | 70.3 ms: 1.06x slower |
+-----------------------------------------------+----------+------------------------+
| Geometric mean | (ref) | 1.01x faster |
+-----------------------------------------------+----------+------------------------+
|