Model Type | auto-regressive language model, transformer decoder |
|
Use Cases |
Areas: | Research, Commercial applications |
|
Applications: | Code generation, Code completion |
|
Primary Use Cases: | Generate code snippets given some context |
|
Limitations: | No assurance of code functionality, Might produce suboptimal, buggy, or exploitable code |
|
|
Supported Languages | Python (Complete), Java (Complete), JavaScript (Complete), Rust (Complete), C++ (Complete), C (Complete), C# (Complete), Go (Complete) |
|
Training Details |
Data Sources: | |
Methodology: | Fill-in-the-Middle training objective |
|
Context Length: | |
Model Architecture: | variable Grouped Query Attention, Rotary Position Embeddings |
|
|
Input Output |
Input Format: | |
Accepted Modalities: | |
Output Format: | |
Performance Tips: | Measured for maximal batch size on the device |
|
|