[{"data":1,"prerenderedAt":182},["ShallowReactive",2],{"categories-init":3,"tool-details-llamafactory":4,"tool-pricing-llamafactory":5,"tool-rel-llamafactory":6,"tool-llamafactory":115,"tool-res-llamafactory":180,"tool-stacks-llamafactory":181},true,null,[],[7,43,66,92],{"relationship_type":8,"relationship_display_name":9,"relationship_description":10,"relationship_display_order":11,"tool":12,"strength":37,"notes":42},"works_with","Works well with","Complementary tools used side by side in the same stack, with nothing built specifically to connect them.",1,{"tool_id":13,"name":14,"slug":15,"tooltip_description":16,"logo_url":17,"logo_bg":18,"pricing_model":19,"learning_curve_score":23,"popularity_score":24,"hosting_assignment_type":25,"hosting_provider_restriction":26,"hosting_target_restriction":26,"hosting_compatible_tool_ids":4,"parent_tool_id":4,"category":27,"subcategory":31,"categories":35,"subcategories":36,"flexibility_score":24,"performance_score":37,"portability_score":37,"is_featured":38,"tags":39,"aliases":40,"score_reasonings":41,"published_date":4,"last_updated_date":4},213,"Hugging Face","hugging-face","The largest model hub and ecosystem in AI — 2M+ models, 500K+ datasets, and 1M+ Spaces, with three distinct ways to run inference (Serverless API, dedicated Inference Endpoints, or 200+ third-party Inference Providers).","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fhugging-face.svg","dark",{"slug":20,"display_name":21,"description":22},"freemium","Freemium","A free tier is available; additional features, usage limits, or managed hosting require a paid plan.",2,5,"managed_only","open",{"category_id":28,"name":29,"slug":30},21,"AI Infrastructure","ai-infrastructure",{"subcategory_id":32,"name":33,"slug":34},77,"AI Inference Providers","ai-inference-providers",[],[],4,false,[],[],{},"LlamaFactory is built on Hugging Face Transformers, PEFT, and TRL: it loads base models and datasets from the Hub by repository ID and exports merged weights in the standard Hub format.",{"relationship_type":44,"relationship_display_name":45,"relationship_description":46,"relationship_display_order":47,"tool":48,"strength":37,"notes":65},"framework_of","Framework of","A framework or layer built on top of another tool (e.g. Next.js → React).",3,{"tool_id":37,"name":49,"slug":50,"tooltip_description":51,"logo_url":52,"logo_bg":18,"pricing_model":53,"learning_curve_score":23,"popularity_score":24,"hosting_assignment_type":4,"hosting_provider_restriction":26,"hosting_target_restriction":26,"hosting_compatible_tool_ids":4,"parent_tool_id":4,"category":57,"subcategory":4,"categories":60,"subcategories":61,"flexibility_score":24,"performance_score":47,"portability_score":24,"is_featured":38,"tags":62,"aliases":63,"score_reasonings":64,"published_date":4,"last_updated_date":4},"Python","python","Python is a high-level, interpreted, dynamically typed programming language emphasising readability and simplicity. It dominates data science, machine learning, and general-purpose scripting.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fpython.svg",{"slug":54,"display_name":55,"description":56},"open_source","Open Source","Source code is publicly available and free to use, modify, and distribute. No paid plans from the project itself.",{"category_id":47,"name":58,"slug":59},"Programming Languages","programming-languages",[],[],[],[],{},"LlamaFactory is a Python toolkit installed with pip; its CLI and web UI run fine-tuning jobs on the Python machine learning stack.",{"relationship_type":67,"relationship_display_name":68,"relationship_description":69,"relationship_display_order":37,"tool":70,"strength":24,"notes":91},"requires","Requires","This tool depends on the listed tools to function.",{"tool_id":71,"name":72,"slug":73,"tooltip_description":74,"logo_url":75,"logo_bg":18,"pricing_model":76,"learning_curve_score":47,"popularity_score":47,"hosting_assignment_type":77,"hosting_provider_restriction":26,"hosting_target_restriction":26,"hosting_compatible_tool_ids":4,"parent_tool_id":4,"category":78,"subcategory":82,"categories":86,"subcategories":87,"flexibility_score":24,"performance_score":24,"portability_score":47,"is_featured":38,"tags":88,"aliases":89,"score_reasonings":90,"published_date":4,"last_updated_date":4},80,"PyTorch","pytorch","PyTorch is the leading open-source deep learning framework, providing dynamic computation graphs, GPU-accelerated tensor operations, and a rich ecosystem for building and deploying neural networks.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fpytorch.svg",{"slug":54,"display_name":55,"description":56},"library",{"category_id":79,"name":80,"slug":81},14,"Data & ML Libraries","data-ml-libraries",{"subcategory_id":83,"name":84,"slug":85},44,"ML Frameworks","ml-frameworks",[],[],[],[],{},"LlamaFactory trains models on PyTorch through Hugging Face Transformers and PEFT; there is no other training backend.",{"relationship_type":93,"relationship_display_name":94,"relationship_description":95,"relationship_display_order":96,"tool":97,"strength":24,"notes":114},"alternative_to","Alternative to","These tools serve a similar purpose — typically you would pick one, not both.",6,{"tool_id":98,"name":99,"slug":100,"tooltip_description":101,"logo_url":102,"logo_bg":18,"pricing_model":103,"learning_curve_score":47,"popularity_score":37,"hosting_assignment_type":77,"hosting_provider_restriction":26,"hosting_target_restriction":26,"hosting_compatible_tool_ids":4,"parent_tool_id":4,"category":104,"subcategory":105,"categories":109,"subcategories":110,"flexibility_score":37,"performance_score":24,"portability_score":24,"is_featured":38,"tags":111,"aliases":112,"score_reasonings":113,"published_date":4,"last_updated_date":4},236,"Unsloth","unsloth","Open-source desktop app, web UI, and Python library for running and fine-tuning open-weight models locally. It serves models through an OpenAI-compatible API and trains them about twice as fast with roughly 70% less memory than standard pipelines.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Funsloth.png",{"slug":20,"display_name":21,"description":22},{"category_id":79,"name":80,"slug":81},{"subcategory_id":106,"name":107,"slug":108},67,"Fine-Tuning","fine-tuning",[],[],[],[],{},"LlamaFactory covers more training methods and model families, driven from YAML configs or a no-code web UI, and scales to multi-GPU clusters; Unsloth is faster and lighter on memory on a single GPU and also runs models locally. LlamaFactory can use Unsloth's kernels as a backend, but a project picks one as its toolkit.",{"tool_id":116,"name":117,"slug":118,"tooltip_description":119,"logo_url":120,"logo_bg":121,"pricing_model":122,"learning_curve_score":47,"popularity_score":47,"hosting_assignment_type":77,"hosting_provider_restriction":26,"hosting_target_restriction":26,"hosting_compatible_tool_ids":4,"parent_tool_id":4,"category":123,"subcategory":124,"categories":125,"subcategories":128,"flexibility_score":24,"performance_score":37,"portability_score":24,"is_featured":38,"tags":130,"aliases":146,"score_reasonings":147,"published_date":153,"last_updated_date":4,"vendor":4,"website_url":154,"documentation_url":155,"github_url":154,"long_description":156,"tagline":157,"key_features":158,"pros":167,"cons":172,"social_links":177,"screenshots_urls":178,"pricing_tiers":179,"license_type":4,"community_size":4,"active_maintenance":3,"parent_tool":4},304,"LlamaFactory","llamafactory","Open-source toolkit for fine-tuning more than 100 open-weight language and vision models with LoRA, QLoRA, full training, and preference methods such as DPO. Runs from YAML configs on the command line or from the LLaMA Board web UI without writing code.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fllamafactory.png","white",{"slug":54,"display_name":55,"description":56},{"category_id":79,"name":80,"slug":81},{"subcategory_id":106,"name":107,"slug":108},[126],{"category_id":79,"name":80,"slug":81,"is_primary":3,"display_order":127},0,[129],{"subcategory_id":106,"name":107,"slug":108,"category_id":79,"is_primary":3,"display_order":127},[131,133,137,141],{"tag_id":11,"name":49,"slug":50,"tag_type":132},"technology",{"tag_id":134,"name":55,"slug":135,"tag_type":136},11,"open-source","feature",{"tag_id":138,"name":139,"slug":140,"tag_type":136},12,"Self-hostable","self-hostable",{"tag_id":142,"name":143,"slug":144,"tag_type":145},25,"Machine Learning","machine-learning","use_case",[],{"learning_curve":148,"flexibility":149,"performance":150,"popularity":151,"portability":152},"LLaMA Board makes a first run a matter of filling in forms, but choosing between training methods, quantization levels, and hyperparameters still requires a working grasp of fine-tuning concepts.","Over 100 model families, every common post-training method from SFT to PPO and SimPO, a dozen parameter-efficient techniques, and both config-file and web UI workflows.","FlashAttention-2, Liger Kernel, and the optional Unsloth backend bring single-GPU QLoRA close to Unsloth's own times, though its memory use is higher than Unsloth running alone.","Around 75K GitHub stars puts it level with Unsloth among fine-tuning projects; fine-tuning itself is work few application developers ever do, so the name rarely comes up outside ML engineering.","Apache-2.0, installs with pip or Docker on Linux, Windows, and macOS, runs on NVIDIA, AMD, and Ascend hardware, and exports standard Hugging Face weights that any runtime can load.","2026-10-02","https:\u002F\u002Fgithub.com\u002Fhiyouga\u002FLlamaFactory","https:\u002F\u002Fllamafactory.readthedocs.io","LlamaFactory, long written **LLaMA-Factory**, is an open-source framework for **fine-tuning open-weight language and vision-language models**. One codebase covers more than 100 model families, including Llama, Qwen, DeepSeek, Mistral, Gemma, GLM, and Phi, plus multimodal models such as Qwen-VL and LLaVA. The project began at Beihang University, was published at ACL 2024, and is one of the most-starred fine-tuning projects on GitHub.\n\nIt bundles the whole **post-training toolbox** in one place: continued pre-training, supervised fine-tuning, reward modeling, PPO, DPO, KTO, ORPO, and SimPO. Each can run as full fine-tuning, freeze-tuning, LoRA, or QLoRA with 2- to 8-bit quantization, alongside newer methods such as DoRA, PiSSA, GaLore, and OFT. Speed-ups come from FlashAttention-2, Liger Kernel, and Unsloth's kernels as an optional backend, and DeepSpeed and Megatron-core handle multi-GPU and multi-node runs.\n\nThere are two ways to drive it. The **`llamafactory-cli`** runs training, evaluation, merging, and export from YAML config files, which keeps runs reproducible. **LLaMA Board**, a Gradio web UI, exposes the same options as forms, so a first fine-tune needs no code at all. After training, the CLI can chat with the model or serve it through an OpenAI-compatible API backed by vLLM or SGLang, and export merged weights for other runtimes.\n\nLlamaFactory is **Apache-2.0** and free, with no hosted plan of its own. It runs on the user's own GPUs, a cloud GPU instance, or a Docker image, and the hardware needed ranges from a single consumer GPU for QLoRA on a 7B model to a cluster for full training of a 70B one.","Unified efficient fine-tuning of 100+ LLMs and VLMs.",[159,160,161,162,163,164,165,166],"Fine-tuning for 100+ language and vision-language model families","SFT, reward modeling, PPO, DPO, KTO, ORPO, and SimPO training","Full, freeze, LoRA, and 2- to 8-bit QLoRA tuning","LLaMA Board web UI for no-code training runs","YAML-driven CLI for reproducible training, merging, and export","FlashAttention-2, Liger Kernel, and Unsloth acceleration backends","DeepSpeed and Megatron-core for multi-GPU training","OpenAI-compatible inference through vLLM or SGLang after training",[168,169,170,171],"The web UI gets a first fine-tune running without writing code, which suits newcomers","Covers far more training methods and model families than most single-purpose tools","Can use Unsloth's kernels as a backend, so single-GPU runs get most of Unsloth's speed","Scales from one consumer GPU to multi-node clusters with DeepSpeed or Megatron",[173,174,175,176],"Uses noticeably more VRAM than Unsloth for the same QLoRA job in community benchmarks","The sheer number of options in the UI and configs is overwhelming once past the defaults","Documentation and community discussion lean toward Chinese-language sources","No hosted service, so users must provision and manage their own GPUs",{},[],[],[],[],1792679715867]