Dohyeon1/ERNIE-M-SMoE-ngroups48

TEXT GENERATIONPricing:Input $0.32 / Cached $0.016 / Output $1.6Concurrent Unit Cost:1Model Size:21BQuant:FP8Context Size:32kPublished:Sep 19, 2026Architecture:Transformer Featherless Exclusive Cold

Dohyeon1/ERNIE-M-SMoE-ngroups48 is a 21 billion parameter model. The model card indicates it is a Hugging Face Transformers model, but specific architectural details, language support, and fine-tuning information are not provided. Its primary differentiators and intended use cases are not specified in the available documentation.

Loading preview...

Model Overview

This model, Dohyeon1/ERNIE-M-SMoE-ngroups48, is a 21 billion parameter model hosted on the Hugging Face Hub. The provided model card is a basic, automatically generated template with placeholders for most key details.

Key Characteristics

  • Parameter Count: 21 billion parameters.
  • Context Length: 32768 tokens.
  • Model Type: Identified as a Hugging Face Transformers model.

Information Not Available

Due to the nature of the provided model card, several critical pieces of information are currently missing:

  • Developed By: The original developer or organization behind the model is not specified.
  • Model Architecture: Specifics about its architecture (e.g., ERNIE-M, SMoE) are mentioned in the name but not detailed in the card.
  • Language(s): The languages it supports are not listed.
  • License: The licensing terms for its use are not provided.
  • Training Details: Information regarding training data, procedure, hyperparameters, or evaluation results is absent.
  • Intended Use Cases: Direct or downstream use cases, as well as out-of-scope uses, are not defined.
  • Bias, Risks, and Limitations: No specific details are provided regarding potential biases, risks, or limitations.

Recommendations

Users are advised that significant information is missing from the model card. Before using this model, it is recommended to seek further documentation from the model developer or maintainer to understand its capabilities, limitations, and appropriate use cases.