dirac-run/ec-0.6b

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:0.8BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Oct 5, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

dirac-run/ec-0.6b is a 0.8 billion parameter instruction-tuned causal language model developed by dirac-run, derived from Qwen/Qwen3-0.6B. This model is specifically fine-tuned to generate GNU/Linux Bash commands from English requests, outputting them in COMMAND JSON format. With a context length of 32768 tokens, it excels at translating natural language into executable shell commands, making it ideal for automation and command-line interface interactions.

Loading preview...

EasyCommand 0.6B Overview

EasyCommand 0.6B is a specialized 0.8 billion parameter language model, fine-tuned by dirac-run from the Qwen/Qwen3-0.6B base model. Its primary function is to translate natural language English requests into GNU/Linux Bash commands, formatted as COMMAND JSON. This model is provided as a merged BF16 Hugging Face model, along with its original FP32 LoRA adapter for further training.

Key Capabilities

  • Natural Language to Bash Commands: Generates executable GNU/Linux Bash commands from English prompts.
  • JSON Output: Commands are outputted in a structured COMMAND JSON format, facilitating programmatic use.
  • Trainable LoRA Adapter: Includes the original LoRA adapter, allowing for continued fine-tuning on specific use cases.
  • Small Footprint: At 0.8B parameters, it offers efficient command generation for resource-constrained environments.
  • High Context Length: Supports a context window of 32768 tokens, enabling complex command generation scenarios.

Good For

  • Command-Line Automation: Automating repetitive tasks by converting natural language instructions into shell scripts.
  • Developer Tools: Integrating natural language command generation into IDEs or custom developer workflows.
  • Educational Purposes: Helping users learn Bash commands by providing translations from English descriptions.
  • Specialized Linux Environments: Deploying in environments where precise, JSON-formatted Bash command generation is critical.

Limitations

  • The model targets English requests and GNU/Linux utilities; it does not inspect live filesystems or installed tools.
  • Generated commands may be incorrect, incomplete, or destructive; inspection before execution is crucial.
  • Evaluation results (ALFA-updated) are development measurements and not directly comparable with external original-ALFA benchmarks due to differing protocols.

Finetunes
2 models