- Notifications
You must be signed in to change notification settings - Fork216
Foundation Architecture for (M)LLMs
License
microsoft/torchscale
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
TorchScale is a PyTorch library that allows researchers and developers to scale up Transformers efficiently and effectively.
Fundamental research to develop new architectures for foundation models and A(G)I, focusing on modeling generality and capability, as well as training stability and efficiency.
- Stability -DeepNet: scaling Transformers to 1,000 Layers and beyond
- Generality -Foundation Transformers (Magneto): towards true general-purpose modeling across tasks and modalities (including language, vision, speech, and multimodal)
- Capability - ALength-Extrapolatable Transformer
- Efficiency -X-MoE: scalable & finetunable sparse Mixture-of-Experts (MoE)
- BitNet: 1-bit Transformers for Large Language Models
- RetNet: Retentive Network: A Successor to Transformer for Large Language Models
- LongNet: Scaling Transformers to 1,000,000,000 Tokens
- December, 2023:LongNet andLongViT released
- October, 2023: Update RMSNorm and SwiGLU as the default module in RetNet
- November, 2022: TorchScale 0.1.1 released [Paper] [PyPI]
To install:
pip install torchscaleAlternatively, you can develop it locally:
git clone https://github.com/microsoft/torchscale.gitcd torchscalepip install -e .For faster training installFlash Attention for Turing, Ampere, Ada, or Hopper GPUs:
pip install flash-attnorxFormers for Volta, Turing, Ampere, Ada, or Hopper GPUs:
# cuda 11.8 versionpip3 install -U xformers --index-url https://download.pytorch.org/whl/cu118# cuda 12.1 versionpip3 install -U xformers --index-url https://download.pytorch.org/whl/cu121It takes only several lines of code to create a model with the above fundamental research features enabled. Here is how to quickly obtain a BERT-like encoder:
>>>fromtorchscale.architecture.configimportEncoderConfig>>>fromtorchscale.architecture.encoderimportEncoder>>>config=EncoderConfig(vocab_size=64000)>>>model=Encoder(config)>>>print(model)
We also support theDecoder architecture and theEncoderDecoder architecture:
# Creating a decoder model>>>fromtorchscale.architecture.configimportDecoderConfig>>>fromtorchscale.architecture.decoderimportDecoder>>>config=DecoderConfig(vocab_size=64000)>>>decoder=Decoder(config)>>>print(decoder)# Creating a encoder-decoder model>>>fromtorchscale.architecture.configimportEncoderDecoderConfig>>>fromtorchscale.architecture.encoder_decoderimportEncoderDecoder>>>config=EncoderDecoderConfig(vocab_size=64000)>>>encdec=EncoderDecoder(config)>>>print(encdec)
It takes only several lines of code to create a RetNet model:
# Creating a RetNet model>>>importtorch>>>fromtorchscale.architecture.configimportRetNetConfig>>>fromtorchscale.architecture.retnetimportRetNetDecoder>>>config=RetNetConfig(vocab_size=64000)>>>retnet=RetNetDecoder(config)>>>print(retnet)
For LongNet models (Flash Attention required):
>>>importtorch>>>fromtorchscale.architecture.configimportEncoderConfig,DecoderConfig>>>fromtorchscale.model.longnetimportLongNetEncoder,LongNetDecoder# Creating a LongNet encoder with the dilated pattern of segment_length=[2048,4096] and dilated_ratio=[1,2]>>>config=EncoderConfig(vocab_size=64000,segment_length='[2048,4096]',dilated_ratio='[1,2]',flash_attention=True)>>>longnet=LongNetEncoder(config)# Creating a LongNet decoder with the dilated pattern of segment_length=[2048,4096] and dilated_ratio=[1,2]>>>config=DecoderConfig(vocab_size=64000,segment_length='[2048,4096]',dilated_ratio='[1,2]',flash_attention=True)>>>longnet=LongNetDecoder(config)
DeepNorm to improve the training stability of Post-LayerNorm Transformers
- enabled by settingdeepnorm=True in the
Configclass. - It adjusts both the residual connection and the initialization method according to the model architecture (i.e., encoder, decoder, or encoder-decoder).
- enabled by settingdeepnorm=True in the
SubLN for the model generality and the training stability
- enabled bysubln=True. This is enabled by default.
- It introduces another LayerNorm to each sublayer and adjusts the initialization according to the model architecture.
- Note that SubLN and DeepNorm cannot be used in one single model.
X-MoE: efficient and finetunable sparse MoE modeling
- enabled byuse_xmoe=True.
- It replaces every'moe_freq'
FeedForwardNetworklayers with the X-MoE layers.
Multiway architecture for multimodality
- enabled bymultiway=True.
- It provides a pool of Transformer's parameters used for different modalities.
Extrapolatable position embedding (Xpos)
- enabled byxpos_rel_pos=True.
- enabled by adjustingrel_pos_buckets andmax_rel_pos.
SparseClip: improving the gradient clipping for sparse MoE models
- we provide asample code that can be easily adapted to the FairSeq (or other) repo.
Retentive Network: A Successor to Transformer for Large Language Models
- created by
config = RetNetConfig(vocab_size=64000)andretnet = RetNetDecoder(config).
- created by
Most of the features above can be used by simply passing the corresponding parameters to the config. For example:
>>>fromtorchscale.architecture.configimportEncoderConfig>>>fromtorchscale.architecture.encoderimportEncoder>>>config=EncoderConfig(vocab_size=64000,deepnorm=True,multiway=True)>>>model=Encoder(config)>>>print(model)
We have examples of how to use TorchScale in the following scenarios/tasks:
Language
Vision
ViT/BEiT [In progress]
Speech
Multimodal
We plan to provide more examples regarding different tasks (e.g. vision pretraining and speech recognition) and various deep learning toolkits (e.g.DeepSpeed andMegatron-LM). Any comments or PRs are welcome!
Some implementations in TorchScale are either adapted from or inspired by theFairSeq repository and theUniLM repository.
If you find this repository useful, please consider citing our work:
@article{torchscale, author = {Shuming Ma and Hongyu Wang and Shaohan Huang and Wenhui Wang and Zewen Chi and Li Dong and Alon Benhaim and Barun Patra and Vishrav Chaudhary and Xia Song and Furu Wei}, title = {{TorchScale}: {Transformers} at Scale}, journal = {CoRR}, volume = {abs/2211.13184}, year = {2022}}@article{deepnet, author = {Hongyu Wang and Shuming Ma and Li Dong and Shaohan Huang and Dongdong Zhang and Furu Wei}, title = {{DeepNet}: Scaling {Transformers} to 1,000 Layers}, journal = {CoRR}, volume = {abs/2203.00555}, year = {2022},}@article{magneto, author = {Hongyu Wang and Shuming Ma and Shaohan Huang and Li Dong and Wenhui Wang and Zhiliang Peng and Yu Wu and Payal Bajaj and Saksham Singhal and Alon Benhaim and Barun Patra and Zhun Liu and Vishrav Chaudhary and Xia Song and Furu Wei}, title = {Foundation {Transformers}}, journal = {CoRR}, volume = {abs/2210.06423}, year = {2022}}@inproceedings{xmoe, title={On the Representation Collapse of Sparse Mixture of Experts}, author={Zewen Chi and Li Dong and Shaohan Huang and Damai Dai and Shuming Ma and Barun Patra and Saksham Singhal and Payal Bajaj and Xia Song and Xian-Ling Mao and Heyan Huang and Furu Wei}, booktitle={Advances in Neural Information Processing Systems}, year={2022}, url={https://openreview.net/forum?id=mWaYC6CZf5}}@article{retnet, author={Yutao Sun and Li Dong and Shaohan Huang and Shuming Ma and Yuqing Xia and Jilong Xue and Jianyong Wang and Furu Wei}, title = {Retentive Network: A Successor to {Transformer} for Large Language Models}, journal = {ArXiv}, volume = {abs/2307.08621}, year = {2023}}@article{longnet, author={Jiayu Ding and Shuming Ma and Li Dong and Xingxing Zhang and Shaohan Huang and Wenhui Wang and Nanning Zheng and Furu Wei}, title = {{LongNet}: Scaling Transformers to 1,000,000,000 Tokens}, journal = {ArXiv}, volume = {abs/2307.02486}, year = {2023}}@article{longvit, title = {When an Image is Worth 1,024 x 1,024 Words: A Case Study in Computational Pathology}, author = {Wenhui Wang and Shuming Ma and Hanwen Xu and Naoto Usuyama and Jiayu Ding and Hoifung Poon and Furu Wei}, journal = {ArXiv}, volume = {abs/2312.03558}, year = {2023}}This project welcomes contributions and suggestions. Most contributions require you to agree to aContributor License Agreement (CLA) declaring that you have the right to, and actually do, grant usthe rights to use your contribution. For details, visithttps://cla.opensource.microsoft.com.
When you submit a pull request, a CLA bot will automatically determine whether you need to providea CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructionsprovided by the bot. You will only need to do this once across all repos using our CLA.
This project has adopted theMicrosoft Open Source Code of Conduct.For more information, see theCode of Conduct FAQ orcontactFuru Wei andShuming Ma with any additional questions or comments.
This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must followMicrosoft's Trademark & Brand Guidelines.Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship.Any use of third-party trademarks or logos is subject to those third-party's policies.