# Image Classification on MNIST (Transfer/Average/Last Scores)

Source: [Image Classification on MNIST (Transfer/Average/Last Scores)](<https://www.sota2.com/research/sota/image-classification-on-mnist-transfer-average-last-scores>)

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

Data updated: 2026-03-12T07:37:05.948Z

Evaluation of Image Classification performance on the MNIST dataset, reporting specific scores related to transfer learning or sequential evaluation phases (Transfer score, Average score, Last score).

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 | Image Classification |
| Dataset (including any stated split) | MNIST |
| Metric | Average Score |
| Unit | PERCENT |
| 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

- [Average Score](<https://www.sota2.com/research/sota/image-classification-on-mnist-transfer-average-last-scores.md?metric=Average%20Score>): Higher is better; unit: PERCENT.

- [Transfer Score](<https://www.sota2.com/research/sota/image-classification-on-mnist-transfer-average-last-scores.md?metric=Transfer%20Score>): Higher is better; unit: PERCENT.

- [Last Score](<https://www.sota2.com/research/sota/image-classification-on-mnist-transfer-average-last-scores.md?metric=Last%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Average Score | Result date | Report paper |
| --- | --- | --- | --- |
| Individual FT | 96.6 | Not recorded | 68fa77941e209131cea741ae |
| MoE | 76.5 | Not recorded | 68fa77941e209131cea741ae |
| ConDU (LoRA) | 74 | Not recorded | 68fa77941e209131cea741ae |
| ConDU (FT) | 73.4 | Not recorded | 68fa77941e209131cea741ae |
| WiSE-FT | 71.5 | Not recorded | 68fa77941e209131cea741ae |
| Continual FT | 64.7 | Not recorded | 68fa77941e209131cea741ae |
| Zero-shot | 59.6 | Not recorded | 68fa77941e209131cea741ae |
| ZSCL | 56.4 | Not recorded | 68fa77941e209131cea741ae |

### Individual FT

Average Score: 96.6. Result ID: 69b26d106dbb7075b95cb412.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

Recorded evaluation configuration: {"Adaptation strategy":"Individual Fine-tuning"}. This record does not assert independent reproduction.

### MoE

Average Score: 76.5. Result ID: 69b26d106dbb7075b95cb416.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73491e209131cea5b8b2.md>)

Recorded evaluation configuration: {"Continual learning method":"MoE"}. This record does not assert independent reproduction.

### ConDU (LoRA)

Average Score: 74. Result ID: 69b26d106dbb7075b95cb418.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

Recorded evaluation configuration: {"Adaptation strategy":"LoRA"}. This record does not assert independent reproduction.

### ConDU (FT)

Average Score: 73.4. Result ID: 69b26d106dbb7075b95cb417.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

Recorded evaluation configuration: {"Adaptation strategy":"Fine-tuning"}. This record does not assert independent reproduction.

### WiSE-FT

Average Score: 71.5. Result ID: 69b26d106dbb7075b95cb414.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cd31e209131cea3601a.md>)

[Implementation](<https://github.com/mlfoundations/wise-ft>)

Recorded evaluation configuration: {"Adaptation strategy":"WiSE-FT"}. This record does not assert independent reproduction.

### Continual FT

Average Score: 64.7. Result ID: 69b26d106dbb7075b95cb413.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

Recorded evaluation configuration: {"Adaptation strategy":"Continual Fine-tuning"}. This record does not assert independent reproduction.

### Zero-shot

Average Score: 59.6. Result ID: 69b26d106dbb7075b95cb411.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Adaptation strategy":"Zero-shot"}. This record does not assert independent reproduction.

### ZSCL

Average Score: 56.4. Result ID: 69b26d106dbb7075b95cb415.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77941e209131cea741ae.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70421e209131cea49f8b.md>)

Recorded evaluation configuration: {"Continual learning method":"ZSCL"}. This record does not assert independent reproduction.

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