formray/prism-detection-fakespot-onnx

By formray · ↓ 12 · ♥ 0 · Updated 2026-02-21

What's in this repo

This page lists the files in formray/prism-detection-fakespot-onnx, including model.onnx.

1 ONNX models (.onnx files)
5 files, including 2 .json, 1 .md, 1 .onnx
Last updated 2026-02-21

ONNX files

File	Size	Download
model.onnx	—	Download

How to use

These ONNX files can be loaded with ONNX Runtime. To inspect a model graph, upload the file to the Netron viewer. Download links point to the original files on Hugging Face.

README

PRISM Detection — FakeSpot RoBERTa (ONNX)

ONNX export of fakespot-ai/roberta-base-ai-text-detection-v1 for use as a detection preset in PRISM.

Status: Candidate (Experimental)

This model is available as an experimental preset in PRISM. It has not yet been promoted to production default.

Benchmark Results (2026-02-21)

Metric	Value
Accuracy	77.5%
F1	0.775
FPR	22.5%
Non-EN FPR	24.2%
Mean Latency	~690 ms
Dataset	320 samples (stratified, multilingual)

Compared to legacy presets (Light/Balanced/Accurate) which score 53-55% accuracy and 67-75% FPR on the same dataset.

Files

File	Size	SHA256
`model.onnx`	499 MB	`1f6940bf3a527da52c86407ebfb4f1f2b66574fa28318b3f151caee4e2235ed3`
`tokenizer.json`	3.6 MB	`ebea7a057f31d0c355148f865d22538337d0209d3abf1b21a0336879bf927308`
`config.json`	796 B	—

Attribution

Original model by FakeSpot AI. Exported to ONNX using optimum-cli for local inference in PRISM.

License

MIT — same as the original model.

Model info

onnx roberta region:us

↓ 12 ♥ 0

Created: 2026-02-21

Updated: 2026-02-21

View on Hugging Face

Files (5)

.gitattributes

README.md

config.json

model.onnx ONNX

tokenizer.json

Related models

More from formray

No other models by this author.

More models

No task type declared.