# Image Captioning on ImageNet 200 images (test) Source: [Image Captioning on ImageNet 200 images (test)]() Collection index: [llms.txt]() Data updated: 2026-03-31T09:10:40.824Z Evaluation of Image Captioning performance on a specific 200-image test subset of ImageNet, focusing on efficiency metrics such as token count, energy consumption, and inference latency. 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 Captioning | | Dataset (including any stated split) | ImageNet 200 images (test) | | Metric | Token Count | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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 - [Token Count](): Higher is better; unit: SCALAR. - [Energy Consumption (J)](): Lower is better; unit: OTHER. - [Latency (s)](): Lower is better; unit: TIME. ## Recorded results | Method | Token Count | Result date | Report paper | | --- | --- | --- | --- | | LingoLoop | 1014.62 | Not recorded | 68fa79891e209131cea7e18a | | LingoLoop | 984.65 | Not recorded | 68fa79891e209131cea7e18a | | LingoLoop | 825.23 | Not recorded | 68fa79891e209131cea7e18a | | LingoLoop | 613.35 | Not recorded | 68fa79891e209131cea7e18a | | Verbose images | 525.7 | Not recorded | 68fa79891e209131cea7e18a | | Verbose images | 451.85 | Not recorded | 68fa79891e209131cea7e18a | | Verbose images | 384.62 | Not recorded | 68fa79891e209131cea7e18a | | Verbose images | 329.02 | Not recorded | 68fa79891e209131cea7e18a | | None | 82.35 | Not recorded | 68fa79891e209131cea7e18a | | Noise | 79.29 | Not recorded | 68fa79891e209131cea7e18a | | Noise | 74.19 | Not recorded | 68fa79891e209131cea7e18a | | None | 73.03 | Not recorded | 68fa79891e209131cea7e18a | | Noise | 67.1 | Not recorded | 68fa79891e209131cea7e18a | | Noise | 65.21 | Not recorded | 68fa79891e209131cea7e18a | | None | 65.04 | Not recorded | 68fa79891e209131cea7e18a | | None | 64.09 | Not recorded | 68fa79891e209131cea7e18a | ### LingoLoop Token Count: 1014.62. Result ID: 69cb8ddd47f34ce924b40a42. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MLLM":"Qwen2.5-VL-3B"}. This record does not assert independent reproduction. ### LingoLoop Token Count: 984.65. Result ID: 69cb8ddd47f34ce924b40a3e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MLLM":"InstructBLIP"}. This record does not assert independent reproduction. ### LingoLoop Token Count: 825.23. Result ID: 69cb8ddd47f34ce924b40a46. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MLLM":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction. ### LingoLoop Token Count: 613.35. Result ID: 69cb8ddd47f34ce924b40a4a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MLLM":"InternVL3-8B"}. This record does not assert independent reproduction. ### Verbose images Token Count: 525.7. Result ID: 69cb8ddd47f34ce924b40a41. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MLLM":"Qwen2.5-VL-3B"}. This record does not assert independent reproduction. ### Verbose images Token Count: 451.85. Result ID: 69cb8ddd47f34ce924b40a3d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MLLM":"InstructBLIP"}. This record does not assert independent reproduction. ### Verbose images Token Count: 384.62. Result ID: 69cb8ddd47f34ce924b40a45. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MLLM":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction. ### Verbose images Token Count: 329.02. Result ID: 69cb8ddd47f34ce924b40a49. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MLLM":"InternVL3-8B"}. This record does not assert independent reproduction. ### None Token Count: 82.35. Result ID: 69cb8ddd47f34ce924b40a43. [Reporting paper]() Recorded evaluation configuration: {"MLLM":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction. ### Noise Token Count: 79.29. Result ID: 69cb8ddd47f34ce924b40a44. [Reporting paper]() Recorded evaluation configuration: {"MLLM":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction. ### Noise Token Count: 74.19. Result ID: 69cb8ddd47f34ce924b40a3c. [Reporting paper]() Recorded evaluation configuration: {"MLLM":"InstructBLIP"}. This record does not assert independent reproduction. ### None Token Count: 73.03. Result ID: 69cb8ddd47f34ce924b40a3b. [Reporting paper]() Recorded evaluation configuration: {"MLLM":"InstructBLIP"}. This record does not assert independent reproduction. ### Noise Token Count: 67.1. Result ID: 69cb8ddd47f34ce924b40a48. [Reporting paper]() Recorded evaluation configuration: {"MLLM":"InternVL3-8B"}. This record does not assert independent reproduction. ### Noise Token Count: 65.21. Result ID: 69cb8ddd47f34ce924b40a40. [Reporting paper]() Recorded evaluation configuration: {"MLLM":"Qwen2.5-VL-3B"}. This record does not assert independent reproduction. ### None Token Count: 65.04. Result ID: 69cb8ddd47f34ce924b40a47. [Reporting paper]() Recorded evaluation configuration: {"MLLM":"InternVL3-8B"}. This record does not assert independent reproduction. ### None Token Count: 64.09. Result ID: 69cb8ddd47f34ce924b40a3f. [Reporting paper]() Recorded evaluation configuration: {"MLLM":"Qwen2.5-VL-3B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.