LIVE

Publisher profile

convaiinnovations AI models

Explore active AI models attributed to convaiinnovations across indexed source catalogs. Open any profile for capabilities, context, pricing and verified source links.

Models
3
Categories
3
Sources
1
typed-decisionsagentscybersecurity

Available catalog

Published models

3 active

Hugging Face↗

laya

laya is a text-classification model intended for typed decisions rather than general-purpose text generation. The available record lists 421,293,830 parameters and identifies support for Transformers and safetensors. Its indexed capabilities and tags include calibrated decisions, RLCD, classification, routing, scoring, guardrails, moderation, and reinforcement learning, suggesting potential use in decision routing, score assignment, moderation workflows, or other applications that require structured classification outputs. The listed modality is text classification, so developers should not assume support for vision, audio, image generation, or conversational generation. The source record marks the model as endpoint-compatible, available in the US region, and licensed under Apache-2.0; developers should still review the Hugging Face repository and license terms before deployment, especially for model-specific conditions or applicable compliance requirements. Context-window size, input and output pricing, and reported inference speed are not disclosed. The available source is the Hugging Face model page, but the supplied page content exposes limited model-card details, so evaluation results, label definitions, training data, safety testing, and production reliability are not established here. Validate classification quality and calibration on representative data before relying on its decisions.

typed-decisionstext
Size
421M parameters
Context
Not listed
Hugging Face↗

laya-multilingual

laya-multilingual is a multilingual text-classification model intended for typed decisions rather than text generation. The model record lists 321,908,998 parameters and identifies Transformers and safetensors support, with capabilities covering classification, routing, guardrails, moderation, calibrated decisions, RLCD, and the Laya/MMBERT/System-One tags. Its multilingual tags include English, German, French, Spanish, Portuguese, Italian, Dutch, Swedish, Danish, Norwegian, Russian, Polish, Turkish, Arabic, Hebrew, Persian, Urdu, Hindi, Bengali, Tamil, Telugu, Kannada, Malayalam, Thai, Vietnamese, Indonesian, Malay, Tagalog, Japanese, Korean, Chinese, Greek, Hungarian, Finnish, Romanian, Albanian, Slovenian, Swahili, Afrikaans, Welsh, Amharic, Armenian, Georgian, Khmer, Burmese, Mongolian, Latvian, Icelandic, Azerbaijani, and Javanese. This makes it potentially useful for multilingual classification pipelines, decision routing, moderation, and guardrail workflows. The supplied source does not disclose a context window, inference-speed benchmark, pricing, evaluation results, or detailed usage examples, so production suitability and performance should be validated independently. The model is tagged Apache-2.0 and commercial-use, but users should still review the repository’s complete license and model-card terms. The source is the Hugging Face model page; no publication or update date is disclosed, so freshness beyond the current page snapshot cannot be verified.

typed-decisionstext
Size
322M parameters
Context
Not listed
Hugging Face↗

laya-typed-decisions

laya-typed-decisions is a 421M-parameter Hugging Face repository intended for text classification and typed decision workflows. Its listed modality is text, and its documented practical use cases include agents, typed decisions, and cybersecurity. This makes it potentially relevant for systems that need to classify inputs into structured decision categories or support agent workflows, but the available information does not disclose specific labels, output schemas, training data, evaluation results, latency, accuracy, context window, or deployment requirements. No inference speed benchmark or pricing is reported. The model record identifies the license as Apache-2.0, but users should review the repository’s licensing and model-card materials for any additional terms or attribution requirements before commercial or security-sensitive deployment. The supplied Hugging Face page is the current source reference, with a canonical repository URL and model title visible; however, the captured page content does not provide further technical documentation beyond the repository identity. Validate behavior on representative cybersecurity and agent tasks, especially where typed outputs or reliable classifications are operationally important.

agentstyped-decisionscybersecuritytext
Size
421M parameters
Context
Not listed

Continue exploring

Related guides and tools