v0.58.1

roboflow/inferencev0.58.1Oct 3, 2025by PawelPeczek-Roboflow

AI Summary

Introduced a critical fix for potential memory (OOM) errors caused by setting low confidence thresholds, adding a new environment variable to control the minimum allowed confidence.

Key Highlights

Fixed potential OOM errors caused by setting confidence to 0.0.
Added CONFIDENCE_LOWER_BOUND_OOM_PREVENTION environment variable (default 0.01).
Added option to disable model monitoring for legacy requests.
Made triggering anchor optional in line_counter block.

New Features

OOM prevention environment variable
Model monitoring toggle
Line counter anchor option

Full Release Notes

# 🔧 Fixed

> [!WARNING] 
> Potential OOM errors in `inference` server as a result of bug introduced in `v0.55.0`

We've detected a bug that was in inference since `0.55.0` release - setting confidence to 0.0 was historically prevented globally in inference server. Since we wanted to remove that constraint, in `0.55.0` release, we make it possible to set such confidence threshold, but that turned out to lead to OOM issues with some models - namely instance segmentation ones, which in some cases may yield large amount of segmentation masks using wast amount of RAM. To provide quick-fix, we've added `CONFIDENCE_LOWER_BOUND_OOM_PREVENTION` env variable flag to control the minimal allowed confidence, set by default to `0.01`.

Review the change here: https://github.com/roboflow/inference/pull/1611

# 🏡 Maintenance
* Add disable model monitoring on legacy requests by @rvirani1 in https://github.com/roboflow/inference/pull/1607
* Make triggering anchor optional in line_counter block by @grzegorz-roboflow in https://github.com/roboflow/inference/pull/1608


**Full Changelog**: https://github.com/roboflow/inference/compare/v0.58.0...v0.58.1