Model Type | |
Use Cases |
Areas: | |
Applications: | Mobile Phone Applications |
|
Limitations: | Hallucinatory issues due to model size, Identity information outputs from ShareGPT corpus |
|
Considerations: | Combining RAG method to enhance memory. |
|
|
Additional Notes | All model parameters for research and limited commercial use. |
|
Supported Languages | languages_supported (Chinese, Mathematics, Coding), proficiency_level (High) |
|
Training Details |
Data Sources: | ShareGPT open-source corpus |
|
Methodology: | |
|
Safety Evaluation |
Risk Categories: | hallucinatory issues, model identity bias |
|
|
Input Output |
Input Format: | |
Accepted Modalities: | |
Output Format: | Language and Multimodal generation |
|
Performance Tips: | Specify model datatype in 'from_pretrained' to avoid calculation errors. |
|
|