| Model Type | | autoregressive, language model |
|
| Use Cases |
| Areas: | | Research, Code generation |
|
| Applications: | |
| Primary Use Cases: | | Generating executable code from English prompts |
|
| Limitations: | | The model is primarily intended for program synthesis., It may not be suitable for tasks outside code generation. |
|
| Considerations: | | Prompts should preferably be in the form of a comment string. |
|
|
| Additional Notes | | Benchmarks used for evaluation include HumanEval and MTPB. |
|
| Supported Languages | | English (proficient), C (supported), C++ (supported), Go (supported), Java (supported), JavaScript (supported), Python (supported) |
|
| Training Details |
| Data Sources: | | BigQuery, GitHub repositories |
|
| Data Volume: | |
| Methodology: | | Pre-trained on multiple programming languages |
|
| Hardware Used: | |
| Model Architecture: | |
|
| Input Output |
| Input Format: | |
| Accepted Modalities: | |
| Output Format: | |
| Performance Tips: | | Use int8_float16 for CUDA device to optimize performance. |
|
|
| Release Notes |
| Version: | |
| Date: | |
| Notes: | | Quantized version of CodeGen-Multi 6B. |
|
|
|