fix(core): apply cap when scaling approximate token counts#35005
fix(core): apply cap when scaling approximate token counts#35005ccurme (ccurme) merged 1 commit intomasterfrom
Conversation
Merging this PR will degrade performance by 23.71%
|
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ❌ | WallTime | test_import_time[CallbackManager] |
424.2 ms | 475.8 ms | -10.85% |
| ❌ | WallTime | test_async_callbacks_in_sync |
18.1 ms | 23.7 ms | -23.71% |
| ❌ | WallTime | test_import_time[Runnable] |
447.9 ms | 519.8 ms | -13.83% |
| ❌ | WallTime | test_import_time[tool] |
465.1 ms | 539.2 ms | -13.74% |
| ❌ | WallTime | test_import_time[Document] |
169 ms | 189.8 ms | -10.94% |
| ❌ | WallTime | test_import_time[HumanMessage] |
237.8 ms | 268.1 ms | -11.32% |
| ❌ | WallTime | test_import_time[LangChainTracer] |
404.3 ms | 453.1 ms | -10.77% |
| ❌ | WallTime | test_import_time[ChatPromptTemplate] |
548.6 ms | 641.4 ms | -14.48% |
| ❌ | WallTime | test_import_time[BaseChatModel] |
488.8 ms | 555.2 ms | -11.96% |
Comparing cc/token_scaling (a4d3120) with master (09654f4)2
Footnotes
-
21 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩
-
No successful run was found on
master(643355f) during the generation of this report, so 09654f4 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩
No description provided.