LLM-OS-Models/LFM2.5-1.2B-Terminal-SFT-2Epoch-LiquidCLI-TemplateHoldout-8GPU
The LLM-OS-Models/LFM2.5-1.2B-Terminal-SFT-2Epoch-LiquidCLI-TemplateHoldout-8GPU is a 1.2 billion parameter model, based on LiquidAI/LFM2.5-1.2B-Instruct, specifically fine-tuned for terminal automation. It generates JSON-formatted commands based on user input and previous terminal states, excelling at predicting the next terminal action. This model is optimized for cost-effective and fast inference in specific terminal automation tasks, offering a conservative approach to command generation.
Loading preview...
Model Overview
This model, LLM-OS-Models/LFM2.5-1.2B-Terminal-SFT-2Epoch-LiquidCLI-TemplateHoldout-8GPU, is a 1.2 billion parameter language model fine-tuned for terminal automation. Built upon LiquidAI/LFM2.5-1.2B-Instruct, it is designed to predict and generate the next terminal command in JSON format, based on user input and the current terminal state. The training involved 2 epochs with Liquid-CLI style preprocessing and a chat-template aligned holdout split, making it efficient for repetitive evaluations and reinforcement learning experiments.
Key Capabilities & Characteristics
- Terminal Command Generation: Specializes in generating JSON-formatted terminal commands for automation.
- Cost-Effective Inference: Offers fast inference at a specific size and acceleration path, making it suitable for cost-sensitive applications.
- Conservative Command Output: Tends to issue fewer incorrect commands, prioritizing accuracy over completeness.
- Fast Execution: Achieves a speed of approximately 0.085 seconds per step.
- Liquid Chat Template Alignment: Optimized for the Liquid chat template and terminal SFT format, beneficial for lightweight and efficient experimentation.
Limitations and Considerations
- Lower Recall: May omit some necessary commands due to relatively lower recall.
- JSON Format Failures: Can occasionally produce invalid JSON, requiring parsing validation or retries before execution.
- Specialized Use: This model is an SFT model for automated terminal operations and does not guarantee general conversational or reasoning performance.
- Safety: Generated commands require safety measures like sandboxing, allowlisting, or human review before execution.