| license: apache-2.0 | |
| tags: | |
| - code-generation | |
| - code-completion | |
| - programming | |
| - lokitrix | |
| language: | |
| - en | |
| - code | |
| pipeline_tag: text-generation | |
| # Lokitrix - AI Code Agent | |
| Lokitrix is a fine-tuned code generation model based on CodeLlama, specialized for: | |
| - Multi-language code generation | |
| - Code completion and refactoring | |
| - Debugging assistance | |
| - Algorithm implementation | |
| ## Model Details | |
| - **Base Model:** codellama/CodeLlama-7b-Instruct-hf | |
| - **Fine-tuning Method:** LoRA (Low-Rank Adaptation) | |
| - **Training Data:** Custom dataset + public coding datasets | |
| - **Parameters:** 7B (with LoRA adapters) | |
| - **License:** Apache 2.0 | |
| ## Training Details | |
| - LoRA Rank: 16 | |
| - Learning Rate: 0.0002 | |
| - Epochs: 3 | |
| - Batch Size: 4 | |
| ## Capabilities | |
| - Python, JavaScript, Java, C++, SQL | |
| - REST API generation | |
| - React components | |
| - Data science code | |
| - Algorithm implementation | |
| - Code refactoring | |
| - Bug fixing | |
| ## Limitations | |
| - May require specific prompt formatting | |
| - Best for code snippets up to 2048 tokens | |
| - Requires GPU for optimal inference | |
| ## Citation | |
| @misc{lokitrix2025, | |
| author = {Loki200519}, | |
| title = {Lokitrix: Fine-tuned Code Generation Agent}, | |
| year = {2025}, | |
| publisher = {Hugging Face}, | |
| } | |