OpenLemur/lemur-70b-v1
OpenLemur/lemur-70b-v1 is a 69 billion parameter causal language model developed by OpenLemur, a collaborative research effort between XLang Lab and Salesforce Research. This model is specifically designed for code generation, demonstrating its capabilities through examples like Python factorial function completion. It is not instruction-tuned, meaning it performs best when prompted with code snippets rather than natural language questions.
Loading preview...
OpenLemur/lemur-70b-v1 Overview
OpenLemur/lemur-70b-v1 is a 69 billion parameter causal language model developed through a collaboration between XLang Lab and Salesforce Research. This model is notable for its focus on code generation tasks, as highlighted by its performance on code completion examples. Unlike many instruction-tuned models, lemur-70b-v1 is not trained on instruction-following corpora, meaning it is best utilized by providing direct code prompts rather than conversational questions.
Key Capabilities
- Code Generation: Excels at completing and generating code snippets, as demonstrated with Python examples.
- Causal Language Modeling: Functions as a base model for predicting the next token in a sequence.
Intended Use Cases
- Code Completion: Ideal for scenarios where users provide partial code and require the model to complete it.
- Research and Development: Suitable for researchers exploring base model capabilities in code-centric applications.
Limitations
- Not Instruction-Tuned: Will not respond effectively to natural language questions or complex instruction-following prompts. Users should avoid asking questions like "What is the Python code to do quick sort?" and instead provide direct code context.
Licensing
The model is licensed under the Llama-2 community license agreement.