February 10th.iFLYTEKOpen Platform announced today DeepSeek-V3, DeepSeek-R1 The models are now online, supporting public cloud API calls and one-click deployment of exclusive models. At the same time, the first DeepSeek large model fine-tuning service, allowing developers to deeply customize the "smarter, more compatible" exclusive DeepSeek large model based on business scenario data.
Officially, the DeepSeek R1 Distillation model fine-tuning serviceAchieve the first, based on high-quality open source small model distillation optimization, to achieve the inference speed increase of more than 3 times, the deployment cost straight down 80%, while maintaining strong logical reasoning ability, the model volume is more lightweight.
Now through March 10, the DeepSeek Inference API free and openIn addition, Xunfei Open Platform will also provide 99.97% SLA cloud services from model training to deployment debugging to scene optimization. In addition, Xunfei Open Platform will also provide 99.97% SLA cloud services from model training, deployment debugging to scene optimization.
1AI Attachment Call Method:
- Log in to the Star MaaS platform, go to the Model Mart, select a model and click "Fine-tune";
- Fill out the application form, after review and approval, click "Create Model", select the corresponding base model, upload the dataset, and submit the training task;