# Offline Reinforcement Learning on Robomimic Can-Paired

Source: [Offline Reinforcement Learning on Robomimic Can-Paired](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-robomimic-can-paired>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:36:27.605Z

Evaluation of Offline Reinforcement Learning performance using the Robomimic Can-Paired dataset, typically measured by return metrics.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Offline Reinforcement Learning |
| Dataset (including any stated split) | Robomimic Can-Paired |
| Metric | Avg True Return |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Avg True Return](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-robomimic-can-paired.md?metric=Avg%20True%20Return>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Avg True Return | Result date | Report paper |
| --- | --- | --- | --- |
| Optimal Policy | 75 | Not recorded | 68fa6e5b1e209131cea3f53f |
| SSR-RRS | 73 | Not recorded | 68fa6e5b1e209131cea3f53f |
| CV | 72 | Not recorded | 68fa6e5b1e209131cea3f53f |
| CV | 72 | Not recorded | 68fa6e5b1e209131cea3f53f |
| BVFT-FQE | 71 | Not recorded | 68fa6e5b1e209131cea3f53f |
| SSR-RRS | 71 | Not recorded | 68fa6e5b1e209131cea3f53f |
| BVFT-FQE | 65 | Not recorded | 68fa6e5b1e209131cea3f53f |

### Optimal Policy

Avg True Return: 75. Result ID: 69902dd13a6c314849feb04a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e5b1e209131cea3f53f.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### SSR-RRS

Avg True Return: 73. Result ID: 69902dd13a6c314849feb049.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e5b1e209131cea3f53f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e5b1e209131cea3f53f.md>)

[Implementation](<https://github.com/StanfordAI4HI/Split-select-retrain>)

Recorded evaluation configuration: {"repeats":5,"re-trained\_on\_full\_dataset":true}. This record does not assert independent reproduction.

### CV

Avg True Return: 72. Result ID: 69902dd13a6c314849feb046.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e5b1e209131cea3f53f.md>)

Recorded evaluation configuration: {"folds":2,"re-trained\_on\_full\_dataset":true}. This record does not assert independent reproduction.

### CV

Avg True Return: 72. Result ID: 69902dd13a6c314849feb047.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e5b1e209131cea3f53f.md>)

Recorded evaluation configuration: {"folds":5,"re-trained\_on\_full\_dataset":true}. This record does not assert independent reproduction.

### BVFT-FQE

Avg True Return: 71. Result ID: 69902dd13a6c314849feb045.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e5b1e209131cea3f53f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b291e209131cea2c68b.md>)

Recorded evaluation configuration: {"variant":"pi + FQE","re-trained\_on\_full\_dataset":true}. This record does not assert independent reproduction.

### SSR-RRS

Avg True Return: 71. Result ID: 69902dd13a6c314849feb048.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e5b1e209131cea3f53f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e5b1e209131cea3f53f.md>)

[Implementation](<https://github.com/StanfordAI4HI/Split-select-retrain>)

Recorded evaluation configuration: {"repeats":2,"re-trained\_on\_full\_dataset":true}. This record does not assert independent reproduction.

### BVFT-FQE

Avg True Return: 65. Result ID: 69902dd13a6c314849feb044.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e5b1e209131cea3f53f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b291e209131cea2c68b.md>)

Recorded evaluation configuration: {"variant":"pi x FQE","re-trained\_on\_full\_dataset":true}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.