SAVRN Model Hub · Comparisons
nsfw_image_detection vs rorshark-vit-base
Nsfw_image_detection has 86M parameters and rorshark-vit-base has 86M parameters; both are released under Apache License 2.0; at 16-bit, nsfw_image_detection needs about 0.2 GB (1x MI300X from $1.85 an hour) and rorshark-vit-base about 0.2 GB (1x MI300X from $1.85 an hour).
| Field | nsfw_image_detection Falconsai/nsfw_image_detection | rorshark-vit-base amunchet/rorshark-vit-base |
|---|---|---|
| Publisher | Falcons.ai | Chester Enright |
| Task | Image classification | Image classification |
| Modality | Image | Image |
| Parameters, as reported | 86M parameters | 86M parameters |
| Architecture | ViTForImageClassification | ViTForImageClassification |
| Library | transformers | transformers |
| Context length | Not stated | Not stated |
| Repository size | 1.5 GB | 343.3 MB |
| Artifact formats | safetensors, pytorch | safetensors, tensorboard |
| License | apache-2.0 | apache-2.0 |
| Access | Open weights, no gate | Open weights, no gate |
| Memory at 16-bit (weights and margin) | 0.2 GB | 0.2 GB |
| Cheapest GPUs at 16-bit, per hour | 1x MI300X, $1.85 | 1x MI300X, $1.85 |
| Memory at 4-bit (weights and margin) | 0.1 GB | 0.1 GB |
| Cheapest GPUs at 4-bit, per hour | 1x MI300X, $1.85 | 1x MI300X, $1.85 |
| Revision viewed | 96cb0d0342c7 | 85b973ba04fa |
| Downloads reported by the hub | 3.4M | 788.9k |
| Last observed | 2026-09-18 | 2026-09-18 |
An evaluation row appears only where at least two of these models report the same benchmark with the same stated configuration, metric, unit and setup. Different evaluators stay named in each cell. Values are shown as reported: no unit conversion, no ranking.
Other Reported Results
These results are listed for each model on its own, because the conditions needed to compare them are not stated or do not match. Two results that leave a condition blank are not assumed to share it.
rorshark-vit-base
| Benchmark | Conditions | Result | Reported by | Revision | Date |
|---|---|---|---|---|---|
| imagefolder | Configuration defaultTask Image ClassificationMetric AccuracyComparison conditions not established | 0.992293 | amunchet Publisher reported |
Evaluated revision not stated | — |
SAVRN's Notes on nsfw_image_detection
A moderation gate does not get its own accelerator; it rides along with the pipeline it protects. Falcons.ai fine-tuned the google/vit-base-patch16-224-in21k Vision Transformer, pre-trained on ImageNet-21k, to detect NSFW images at 224 by 224 pixels. With 86M parameters it needs 0.2 GB at 16-bit and 0.1 GB at 8-bit or 4-bit. The table's one MI300X with 192 GB at $1.85 an hour on-demand is there because the table has to name something; in practice it shares whatever accelerator already handles the images.
Under Apache 2.0 you can embed it in a commercial product, modify it and redistribute the result, provided the notices travel with it and you state significant changes. Check the files against the table: the stored precision is float32 and the nine files total 1.46 GB, well above the 0.2 GB row. Released in October 2023, it was last updated in September 2026.
Questions
Which is larger, nsfw_image_detection or rorshark-vit-base?
nsfw_image_detection (86M parameters) is larger than rorshark-vit-base (86M parameters), by the parameter counts their publishers report.
Which is cheaper to run, nsfw_image_detection or rorshark-vit-base?
At 4-bit, nsfw_image_detection fits on 1x MI300X from $1.85 an hour and rorshark-vit-base on 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use nsfw_image_detection commercially?
Yes. nsfw_image_detection is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
Can I use rorshark-vit-base commercially?
Yes. rorshark-vit-base is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.