torchtune now officially supports Meta Llama3! Check out our recipes for Llama3-8B-Instruct with LoRA, QLoRA and Full fine-tune in the Llama3 section! We also support 70B fine-tuning with LoRA! 🚀 🦙
Introduction | Installation | Get Started | Documentation | Design Principles | Community Contributions | License
torchtune is a PyTorch-native library for easily authoring, fine-tuning and experimenting with LLMs. We're excited to announce our alpha release!
torchtune provides:
- Native-PyTorch implementations of popular LLMs using composable and modular building blocks
- Easy-to-use and hackable training recipes for popular fine-tuning techniques (LoRA, QLoRA) - no trainers, no frameworks, just PyTorch!
- YAML configs for easily configuring training, evaluation, quantization or inference recipes
- Built-in support for many popular dataset formats and prompt templates to help you quickly get started with training
torchtune focuses on integrating with popular tools and libraries from the ecosystem. These are just a few examples, with more under development:
- Hugging Face Hub for accessing model weights
- EleutherAI's LM Eval Harness for evaluating trained models
- Hugging Face Datasets for access to training and evaluation datasets
- PyTorch FSDP for distributed training
- torchao for lower precision dtypes and post-training quantization techniques
- Weights & Biases for logging metrics and checkpoints, and tracking training progress
- ExecuTorch for on-device inference using fine-tuned models
- bitsandbytes for low memory optimizers for our single-device recipes
torchtune currently supports the following models.
| Model | Sizes |
|---|---|
| Llama3 | 8B, 70B [models, configs] |
| Llama2 | 7B, 13B, 70B [models, configs] |
| Mistral | 7B [model, configs] |
| Gemma | 2B [model, configs] |
We'll be adding a number of new models in the coming weeks, including support for 70B versions and MoEs.
torchtune provides the following fine-tuning recipes.
| Training | Fine-tuning Method |
|---|---|
| Distributed Training [1 to 8 GPUs] | Full [code, example], LoRA [code, example] |
| Single Device / Low Memory [1 GPU] | Full [code, example], LoRA + QLoRA [code, example] |
| Single Device [1 GPU] | DPO [code, example] |
Memory efficiency is important to us. All of our recipes are tested on a variety of setups including commodity GPUs with 24GB of VRAM as well as beefier options found in data centers.
Single-GPU recipes expose a number of memory optimizations that aren't available in the distributed versions. These include support for low-precision optimizers from bitsandbytes and fusing optimizer step with backward to reduce memory footprint from the gradients (see example config). For memory-constrained setups, we recommend using the single-device configs as a starting point. For example, our default QLoRA config has a peak memory usage of ~9.3GB. Similarly LoRA on single device with batch_size=2 has a peak memory usage of ~17.1GB. Both of these are with dtype=bf16 and AdamW as the optimizer.
This table captures the minimum memory requirements for our different recipes using the associated configs.
| Example HW Resources | Finetuning Method | Config | Model | Peak Memory per GPU |
|---|---|---|---|---|
| 1 x RTX 4090 | QLoRA | qlora_finetune_single_device | Llama2-7B | 8.57 GB |
| 2 x RTX 4090 | LoRA | lora_finetune_distributed | Llama2-7B | 20.95 GB |
| 1 x RTX 4090 | LoRA | lora_finetune_single_device | Llama2-7B | 17.18 GB |
| 1 x RTX 4090 | Full finetune | full_finetune_single_device | Llama2-7B | 14.97 GB |
| 4 x RTX 4090 | Full finetune | full_finetune_distributed | Llama2-7B | 22.9 GB |
- these are averaged over multiple runs, but there might be some variance based on the setup. We'll update this table regularly.
torchtune supports fine-tuning for the Llama3 8B and 70B size models. We currently support LoRA, QLoRA and full fine-tune on a single GPU as well as LoRA and full fine-tune on multiple devices for the 8B model, and LoRA on multiple devices for the 70B model. For all the details, take a look at our tutorial.
Note: our Llama3 LoRA and QLoRA configs default to the instruct fine-tuned models. This is because not all special token embeddings are initialized in the base 8B and 70B models.
In our initial experiments for Llama3-8B, QLoRA has a peak allocated memory of ~9GB while LoRA on a single GPU has a peak allocated memory of ~19GB. To get started, you can use our default configs to kick off training.
LoRA 8B
tune run lora_finetune_single_device --config llama3/8B_lora_single_deviceQLoRA 8B
tune run lora_finetune_single_device --config llama3/8B_qlora_single_deviceFull 8B
tune run full_finetune_single_device --config llama3/8B_full_single_deviceFull 8B
tune run --nproc_per_node 4 full_finetune_distributed --config llama3/8B_fullLoRA 8B
tune run --nproc_per_node 2 lora_finetune_distributed --config llama3/8B_loraLoRA 70B
Note that the download command for the Meta-Llama3 70B model slightly differs from download commands for the 8B models. This is because we use the HuggingFace safetensor model format to load the model. To download the 70B model, run
tune download meta-llama/Meta-Llama-3-70b --hf-token <> --output-dir /tmp/Meta-Llama-3-70b --ignore-patterns "original/consolidated*"Then, a finetune can be kicked off:
tune run --nproc_per_node 8 lora_finetune_distributed --config recipes/configs/llama3/70B_lora.yamlYou can find a full list of all our Llama3 configs here.
Step 1: Install PyTorch. torchtune is tested with the latest stable PyTorch release (2.2.2) as well as the preview nightly version.
Step 2: The latest stable version of torchtune is hosted on PyPI and can be downloaded with the following command:
pip install torchtuneTo confirm that the package is installed correctly, you can run the following command:
tune --helpAnd should see the following output:
usage: tune [-h] {ls,cp,download,run,validate} ...
Welcome to the TorchTune CLI!
options:
-h, --help show this help message and exit
...
To get started with fine-tuning your first LLM with torchtune, see our tutorial on fine-tuning Llama2 7B. Our end-to-end workflow tutorial will show you how to evaluate, quantize and run inference with this model. The rest of this section will provide a quick overview of these steps with Llama2.
Follow the instructions on the official meta-llama repository to ensure you have access to the official Llama model weights. Once you have confirmed access, you can run the following command to download the weights to your local machine. This will also download the tokenizer model and a responsible use guide.
tune download meta-llama/Llama-2-7b-hf \
--output-dir /tmp/Llama-2-7b-hf \
--hf-token <HF_TOKEN> \tune download meta-llama/Meta-Llama-3-8B \
--output-dir /tmp/Meta-Llama-3-8B \
--hf-token <HF_TOKEN> \Tip: Set your environment variable
HF_TOKENor pass in--hf-tokento the command in order to validate your access. You can find your token at https://huggingface.co/settings/tokens
Llama2 7B + LoRA on single GPU:
tune run lora_finetune_single_device --config llama2/7B_lora_single_deviceFor distributed training, tune CLI integrates with torchrun. Llama2 7B + LoRA on two GPUs:
tune run --nproc_per_node 2 full_finetune_distributed --config llama2/7B_fullTip: Make sure to place any torchrun commands before the recipe specification. Any CLI args after this will override the config and not impact distributed training.
There are two ways in which you can modify configs:
Config Overrides
You can easily overwrite config properties from the command-line:
tune run lora_finetune_single_device \
--config llama2/7B_lora_single_device \
batch_size=8 \
enable_activation_checkpointing=True \
max_steps_per_epoch=128Update a Local Copy
You can also copy the config to your local directory and modify the contents directly:
tune cp llama2/7B_full ./my_custom_config.yaml
Copied to ./7B_full.yamlThen, you can run your custom recipe by directing the tune run command to your local files:
tune run full_finetune_distributed --config ./my_custom_config.yaml
Check out tune --help for all possible CLI commands and options. For more information on using and updating configs, take a look at our config deep-dive.
torchtune embodies PyTorch’s design philosophy [details], especially "usability over everything else".
torchtune is a native-PyTorch library. While we provide integrations with the surrounding ecosystem (eg: Hugging Face Datasets, EleutherAI Eval Harness), all of the core functionality is written in PyTorch.
torchtune is designed to be easy to understand, use and extend.
- Composition over implementation inheritance - layers of inheritance for code re-use makes the code hard to read and extend
- No training frameworks - explicitly outlining the training logic makes it easy to extend for custom use cases
- Code duplication is preferred over unnecessary abstractions
- Modular building blocks over monolithic components
torchtune provides well-tested components with a high-bar on correctness. The library will never be the first to provide a feature, but available features will be thoroughly tested. We provide
- Extensive unit-tests to ensure component-level numerical parity with reference implementations
- Checkpoint-tests to ensure model-level numerical parity with reference implementations
- Integration tests to ensure recipe-level performance parity with reference implementations on standard benchmarks
We really value our community and the contributions made by our wonderful users. We'll use this section to call out some of these contributions! If you'd like to help out as well, please see the CONTRIBUTING guide.
- @solitude-alive for adding the Gemma 2B model to torchtune, including recipe changes, numeric validations of the models and recipe correctness
- @yechenzhi for adding DPO to torchtune, including the recipe and config along with correctness checks
The Llama2 code in this repository is inspired by the original Llama2 code.
We want to give a huge shout-out to EleutherAI, Hugging Face and Weights & Biases for being wonderful collaborators and for working with us on some of these integrations within torchtune.
We also want to acknowledge some awesome libraries and tools from the ecosystem:
- gpt-fast for performant LLM inference techniques which we've adopted OOTB
- llama recipes for spring-boarding the llama2 community
- bitsandbytes for bringing several memory and performance based techniques to the PyTorch ecosystem
- @winglian and axolotl for early feedback and brainstorming on torchtune's design and feature set.
- lit-gpt for pushing the LLM fine-tuning community forward.
- HF TRL for making reward modeling more accessible to the PyTorch community.
torchtune is released under the BSD 3 license. However you may have other legal obligations that govern your use of other content, such as the terms of service for third-party models.