laya-coreml / reports /benchmark-reference.json
alexwengg's picture
Add accuracy benchmark reports
791c64e verified
Raw
History Blame Contribute Delete
2.14 kB
{
"source_repo": "convaiinnovations/laya",
"source_revision": "1c5edc17a7acd8701df6fc341c0d179f1c62c982",
"variant": "multilingual",
"device": "cpu",
"torch": "2.7.0",
"threads": 4,
"seed": 13,
"n_per_task": 400,
"max_len": 1024,
"seconds": 240.0,
"ms_per_question": 61.55,
"suites": {
"app.email_spam": {
"n": 400,
"dropped": 0,
"accuracy": 0.9925,
"max_tokens": 1024,
"mean_tokens": 334.1,
"upstream_published": 0.993
},
"app.guardrails_jailbreak": {
"n": 400,
"dropped": 0,
"accuracy": 0.805,
"max_tokens": 588,
"mean_tokens": 124.9,
"upstream_published": 0.755
},
"app.model_routing_domain": {
"n": 399,
"dropped": 0,
"accuracy": 0.4411,
"max_tokens": 261,
"mean_tokens": 146.7,
"upstream_published": 0.123
},
"app.moderation_toxicity": {
"n": 400,
"dropped": 0,
"accuracy": 0.535,
"max_tokens": 587,
"mean_tokens": 111.2,
"upstream_published": 0.525
},
"app.phishing": {
"n": 400,
"dropped": 0,
"accuracy": 0.9925,
"max_tokens": 1024,
"mean_tokens": 363.9,
"upstream_published": 0.993
},
"app.rag_relevance": {
"n": 400,
"dropped": 0,
"accuracy": 0.6725,
"max_tokens": 269,
"mean_tokens": 147.5,
"upstream_published": 0.657
},
"app.support_triage": {
"n": 400,
"dropped": 0,
"accuracy": 0.54,
"max_tokens": 344,
"mean_tokens": 238.5,
"upstream_published": 0.522
},
"jev.ag_news": {
"n": 400,
"dropped": 0,
"accuracy": 0.935,
"max_tokens": 244,
"mean_tokens": 102.6,
"upstream_published": 0.93
},
"jev.emotion": {
"n": 400,
"dropped": 0,
"accuracy": 0.5375,
"max_tokens": 95,
"mean_tokens": 53.5,
"upstream_published": 0.53
},
"massive_intent.en": {
"n": 300,
"dropped": 0,
"accuracy": 0.6567,
"max_tokens": 236,
"mean_tokens": 199.7,
"upstream_published": 0.657
}
}
}