# Offline Reinforcement Learning on D4RL Adroit v1

Source: [Offline Reinforcement Learning on D4RL Adroit v1](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-adroit-v1>)

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

Data updated: 2026-05-12T07:20:16.229Z

Evaluation of Offline Reinforcement Learning performance on the D4RL Adroit dataset, version 1, reporting scores for specific tasks (e.g., pen-human, pen-cloned) and the overall average.

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) | D4RL Adroit v1 |
| Metric | Score (pen-human) |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [Score (pen-human)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-adroit-v1.md?metric=Score%20(pen-human)>): Higher is better; unit: SCALAR.

- [Score (pen-cloned)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-adroit-v1.md?metric=Score%20(pen-cloned)>): Higher is better; unit: SCALAR.

- [Average Score (Adroit v1)](<https://www.sota2.com/research/sota/offline-reinforcement-learning-on-d4rl-adroit-v1.md?metric=Average%20Score%20(Adroit%20v1)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Score (pen-human) | Result date | Report paper |
| --- | --- | --- | --- |
| DOSER | 87.8 | Not recorded | 6a02b3ebdbc836b87d824520 |
| QGPO | 73.9 | Not recorded | 6a02b3ebdbc836b87d824520 |
| SVR | 73.1 | Not recorded | 6a02b3ebdbc836b87d824520 |
| DQL | 72.8 | Not recorded | 6a02b3ebdbc836b87d824520 |
| IQL | 71.5 | Not recorded | 6a02b3ebdbc836b87d824520 |
| DTQL | 64.1 | Not recorded | 6a02b3ebdbc836b87d824520 |
| TD3+BC | 54.9 | Not recorded | 6a02b3ebdbc836b87d824520 |
| CQL | 35.2 | Not recorded | 6a02b3ebdbc836b87d824520 |

### DOSER

Score (pen-human): 87.8. Result ID: 6a02d46147f34ce924e53123.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3ebdbc836b87d824520.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a02b3ebdbc836b87d824520.md>)

Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction.

### QGPO

Score (pen-human): 73.9. Result ID: 6a02d46147f34ce924e53121.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3ebdbc836b87d824520.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f931e209131cea465bc.md>)

Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction.

### SVR

Score (pen-human): 73.1. Result ID: 6a02d46147f34ce924e5311f.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3ebdbc836b87d824520.md>)

Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction.

### DQL

Score (pen-human): 72.8. Result ID: 6a02d46147f34ce924e53120.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3ebdbc836b87d824520.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70791e209131cea4b56e.md>)

Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction.

### IQL

Score (pen-human): 71.5. Result ID: 6a02d46147f34ce924e5311d.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3ebdbc836b87d824520.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6af71e209131cea2b67f.md>)

Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction.

### DTQL

Score (pen-human): 64.1. Result ID: 6a02d46147f34ce924e53122.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3ebdbc836b87d824520.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75751e209131cea680ec.md>)

Recorded evaluation configuration: {"Method category":"Diffusion-based methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction.

### TD3+BC

Score (pen-human): 54.9. Result ID: 6a02d46147f34ce924e5311c.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3ebdbc836b87d824520.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b631e209131cea2df4f.md>)

[Implementation](<https://github.com/sfujim/TD3_BC>)

Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction.

### CQL

Score (pen-human): 35.2. Result ID: 6a02d46147f34ce924e5311e.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b3ebdbc836b87d824520.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa682a1e209131cea1a6c2.md>)

Recorded evaluation configuration: {"Method category":"Conventional methods","Evaluation protocol":"Average normalized scores at the last training iteration over 4 random seeds"}. This record does not assert independent reproduction.

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