Skills
skills/ascend-ai-coding/awesome-ascend-skills/mindspeed-llm-pipeline/Gen Agent Trust Hub

mindspeed-llm-pipeline

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Downloads model configuration files, tokenizers, and weight tensors (safetensors) from Hugging Face (huggingface.co).
  • [EXTERNAL_DOWNLOADS]: Fetches training and evaluation datasets in Parquet and JSONL formats from Hugging Face.
  • [COMMAND_EXECUTION]: Executes Python scripts for weight conversion (convert_ckpt.py) and dataset preprocessing (preprocess_data.py).
  • [COMMAND_EXECUTION]: Utilizes torchrun to initiate distributed training and fine-tuning (SFT/DPO) processes using posttrain_gpt.py and pretrain_gpt.py.
  • [COMMAND_EXECUTION]: Runs shell scripts for model inference and benchmarking evaluation (MMLU, GSM8K, etc.) located within the skill's example directory.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 11:53 AM
Security Audit — agent-trust-hub — mindspeed-llm-pipeline