Model Type | |
Use Cases |
Areas: | Research, Chatbot development |
|
Applications: | Multilingual chat applications |
|
Primary Use Cases: | Multilingual text generation, Conversational agents |
|
Limitations: | Cannot be used in critical or high-risk situations, May produce undesirable outputs |
|
Considerations: | Users should avoid using this model in scenarios where errors could lead to significant harm. |
|
|
Additional Notes | Prompt format is defined in tokenizer_config.json. Can be used to deploy OpenAI-like API service using vllm. |
|
Supported Languages | zh (Full proficiency), en (Full proficiency), fr (Full proficiency), de (Full proficiency), ja (Full proficiency), ko (Full proficiency), it (Full proficiency), ru (Full proficiency), fi (Full proficiency) |
|
Input Output |
Input Format: | Prompt format using special tokens like <|role|>, <|says|>, <|end|> |
|
Accepted Modalities: | |
Output Format: | |
Performance Tips: | Ensure usage of the fast tokenizer from transformers. |
|
|