Guide LLaMA Factory fine-tuning, preference training, and deployment via CLI and YAML configs.
npx skills add https://github.com/wenerme/ai --skill llamafactory-docs
Please help me install this Agent Skill. Skill: llamafactory-docs Source: https://github.com/wenerme/ai/tree/main/skills/llamafactory-docs Command: npx skills add https://github.com/wenerme/ai --skill llamafactory-docs
LLaMA Factory documentation helps you configure and run fine-tuning, reinforcement learning, and deployment workflows without getting stuck on framework details and dataset formatting requirements.
llamafactory-cli.Ask for a complete setup and training plan by running: llamafactory-cli train examples/train_lora/llama3_lora_sft.yaml.
High-intent search queries and answers about installing and using this skill.
LLM fine-tuning dataset formatting requires structuring data into alpaca or sharegpt formats. LLaMA Factory specifies exact field mappings, including dedicated multimodal fields for images, videos, and audios alongside standard text inputs.
Yes, LLaMA Factory supports distributed training using both DeepSpeed and FSDP. You can configure these frameworks within YAML configs to scale large language model training across multiple GPUs efficiently.
The best way to deploy a fine-tuned LLM is using LLaMA Factory's inference backends like vLLM and SGLang. You can serve models via an OpenAI-style API or use the LLaMA Board Web UI for chat interfaces.
Yes, LLaMA Factory supports multimodal training for images, videos, and audios. You need to configure the specific multimodal fields within your sharegpt or alpaca-style datasets to train models on these inputs.
You can run LoRA and QLoRA fine-tuning by executing `llamafactory-cli train` with a YAML config file. This command processes the training pipeline using documented CLI entry points and specified architecture settings.
LLaMA Factory supports multiple preference learning algorithms for LLM training, including DPO, KTO, PPO, ORPO, SimPO, CPO, and GRPO-style reinforcement learning workflows to align model behaviors.
ovachieverAutomate LLaMA-family model fine-tuning via the Llama-Factory WebUI.
vasilyu1983Optimize LLM inference performance and cost with vLLM and quantization.
blueskies1818Automate distributed training setup for PyTorch models with device placement and mixed precision.
huggingfaceTrain and fine-tune language models on Hugging Face Jobs using TRL methods.