# Autoregressive Image Generation on MS-COCO 2017 (val)

Source: [Autoregressive Image Generation on MS-COCO 2017 (val)](<https://www.sota2.com/research/sota/autoregressive-image-generation-on-ms-coco-2017-val>)

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

Data updated: 2026-02-26T05:38:30.449Z

Evaluation of autoregressive models for image generation on the validation split of the MS-COCO 2017 dataset, assessed using perceptual metrics (FID, CLIP), generation quality indicators (IR, Acc. Len.), and inference performance timings.

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 | Autoregressive Image Generation |
| Dataset (including any stated split) | MS-COCO 2017 (val) |
| Metric | CLIP Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| 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

- [CLIP Score](<https://www.sota2.com/research/sota/autoregressive-image-generation-on-ms-coco-2017-val.md?metric=CLIP%20Score>): Higher is better; unit: SCALAR.

- [FID](<https://www.sota2.com/research/sota/autoregressive-image-generation-on-ms-coco-2017-val.md?metric=FID>): Lower is better; unit: SCALAR.

- [IR](<https://www.sota2.com/research/sota/autoregressive-image-generation-on-ms-coco-2017-val.md?metric=IR>): Higher is better; unit: SCALAR.

- [Acc. Len.](<https://www.sota2.com/research/sota/autoregressive-image-generation-on-ms-coco-2017-val.md?metric=Acc.%20Len.>): Higher is better; unit: SCALAR.

- [Latency (s)](<https://www.sota2.com/research/sota/autoregressive-image-generation-on-ms-coco-2017-val.md?metric=Latency%20(s)>): Lower is better; unit: TIME.

- [Speed-up (x)](<https://www.sota2.com/research/sota/autoregressive-image-generation-on-ms-coco-2017-val.md?metric=Speed-up%20(x)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | CLIP Score | Result date | Report paper |
| --- | --- | --- | --- |
| Target Model | 0.333 | Not recorded | 69687499e25ad2fa9d1ec5d0 |
| Eagle-1 | 0.333 | Not recorded | 69687499e25ad2fa9d1ec5d0 |
| LANTERN++ | 0.3328 | Not recorded | 69687499e25ad2fa9d1ec5d0 |
| COOL-SD | 0.3325 | Not recorded | 69687499e25ad2fa9d1ec5d0 |
| COOL-SD | 0.3167 | Not recorded | 69687499e25ad2fa9d1ec5d0 |
| Target Model | 0.3162 | Not recorded | 69687499e25ad2fa9d1ec5d0 |
| Eagle-1 | 0.3157 | Not recorded | 69687499e25ad2fa9d1ec5d0 |
| LANTERN++ | 0.3157 | Not recorded | 69687499e25ad2fa9d1ec5d0 |
| COOL-SD | 0.3154 | Not recorded | 69687499e25ad2fa9d1ec5d0 |

### Target Model

CLIP Score: 0.333. Result ID: 6997c0c45698665ec4178d6f.

[Reporting paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

Recorded evaluation configuration: {"Target Model":"Lumina-mGPT"}. This record does not assert independent reproduction.

### Eagle-1

CLIP Score: 0.333. Result ID: 6997c0c45698665ec4178d70.

[Reporting paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77391e209131cea71c7d.md>)

Recorded evaluation configuration: {"Target Model":"Lumina-mGPT"}. This record does not assert independent reproduction.

### LANTERN++

CLIP Score: 0.3328. Result ID: 6997c0c45698665ec4178d71.

[Reporting paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77a51e209131cea74974.md>)

Recorded evaluation configuration: {"Target Model":"Lumina-mGPT","lambda":2,"k":10}. This record does not assert independent reproduction.

### COOL-SD

CLIP Score: 0.3325. Result ID: 6997c0c45698665ec4178d72.

[Reporting paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

Recorded evaluation configuration: {"Target Model":"Lumina-mGPT","delta":1.1}. This record does not assert independent reproduction.

### COOL-SD

CLIP Score: 0.3167. Result ID: 6997c0c45698665ec4178d76.

[Reporting paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

Recorded evaluation configuration: {"Target Model":"LlamaGen-XL","delta":1.1}. This record does not assert independent reproduction.

### Target Model

CLIP Score: 0.3162. Result ID: 6997c0c45698665ec4178d73.

[Reporting paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

Recorded evaluation configuration: {"Target Model":"LlamaGen-XL"}. This record does not assert independent reproduction.

### Eagle-1

CLIP Score: 0.3157. Result ID: 6997c0c45698665ec4178d74.

[Reporting paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77391e209131cea71c7d.md>)

Recorded evaluation configuration: {"Target Model":"LlamaGen-XL"}. This record does not assert independent reproduction.

### LANTERN++

CLIP Score: 0.3157. Result ID: 6997c0c45698665ec4178d75.

[Reporting paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77a51e209131cea74974.md>)

Recorded evaluation configuration: {"Target Model":"LlamaGen-XL","lambda":2,"k":10}. This record does not assert independent reproduction.

### COOL-SD

CLIP Score: 0.3154. Result ID: 6997c0c45698665ec4178d77.

[Reporting paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69687499e25ad2fa9d1ec5d0.md>)

Recorded evaluation configuration: {"Target Model":"LlamaGen-XL","delta":2}. This record does not assert independent reproduction.

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