Configure Axolotl YAML for LLM fine-tuning with LoRA/QLoRA and DPO/KTO/ORPO/GRPO preferences.
npx skills add https://github.com/MarbleSodas/Mavis --skill axolotl-marblesodas
Please help me install this Agent Skill. Skill: axolotl Source: https://github.com/MarbleSodas/Mavis/tree/main/optional-skills/mlops/training/axolotl Command: npx skills add https://github.com/MarbleSodas/Mavis --skill axolotl-marblesodas
This Skill helps you fine-tune LLMs correctly and efficiently by translating complex Axolotl training options into actionable guidance for YAML-based setups.
Ask the AI to generate an Axolotl YAML configuration for LoRA or QLoRA fine-tuning using DPO, including FSDP and context parallel settings for your GPU count.
High-intent search queries and answers about installing and using this skill.
To configure Axolotl YAML for LoRA and DPO fine-tuning, you define your LoRA parameters and preference-training datasets directly in the YAML file. This Skill provides actionable requirements to translate complex Axolotl training options into a valid configuration.
The best way to set up multi-GPU FSDP in Axolotl YAML is by specifying distributed training settings and context parallelism divisibility rules. This Skill guides you through optimizing these performance and scaling knobs for your specific GPU count.
Yes, Axolotl supports KTO and ORPO preference training workflows alongside DPO and GRPO. This Skill helps you choose and apply the correct RLHF-style training patterns aligned to Axolotl capabilities within your YAML setup.
To fix dataset preprocessing errors with packed sequences in Axolotl, you must handle dataset edge cases and validate common configuration pitfalls. This Skill provides compatibility checks and debug patterns for common dataset formats.
Yes, you can use QLoRA with context parallelism settings in Axolotl by configuring both in your YAML file. This Skill targets practitioners tuning LLMs and offers guidance on combining QLoRA with distributed training behavior effectively.
x-TheFoxFine-tune large language models with YAML-driven LoRA, DPO, and GRPO configurations.
SethyPagnaFine-tune large language models with parameter-efficient adapters on memory-constrained GPUs.
tadod12Configure DeepSpeed distributed training for large transformer models with ZeRO and mixed precision.
anxiety135790Configure and troubleshoot Axolotl fine-tuning workflows for large language models.
