Skip to content
LAMB Optimizer for Large Batch Training (TensorFlow version)
Branch: master
Clone or download
Latest commit 8788e45 Jan 17, 2020
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitattributes add gitattr Jul 17, 2019 add full results Jun 6, 2019
LICENSE Initial commit Jun 5, 2019 add official LAMB optimizer Jan 17, 2020
algorithm.png initial commit Jun 5, 2019
mnist_tensorflow.ipynb add ipynb Jun 6, 2019 add official LAMB optimizer Jan 17, 2020 add opt Jun 6, 2019

LAMB Optimizer (TensorFlow)

This is a simple implementation of LAMB Optimizer, which appeared in the paper "Large Batch Optimization for Deep Learning: Training BERT in 76 minutes".

The older name of the paper was "Reducing BERT Pre-Training Time from 3 Days to 76 Minutes"

Update: official implementation of LAMB optimizer is now available:


  • This is NOT an official implementation.
  • LAMB optimizer changes slightly from arXiv v1 ~ v3.
  • We implement v3 version (which is the latest version on June, 2019.).
  • Some uncertain parts are clarified by consulting original authors (such as scaling function).


LAMB optimizer is originally designed for large batch learning in neural networks, but could also used in small batch size as indicated by authors.



The implementation is based on BERT repository, which uses AdamWeightDecayOptimizer (appears in for pre-training and fine-tuning.

  • Just use LAMBOptimizer as a regular optimizer in TensorFlow, similar to Adam or AdamWeightDecayOptimizer.
  • Find LAMB optimizer in
  • There is nothing special to tune other than initial learning_rate.

Results on MNIST

  • I don't have TPU Pod to test its scalability on BERT with large batch 😂, but tested on MNIST for verify its effectiveness.
  • All optimizers use an initial learning rate of 0.001 (default settings), and did NOT scale to the batch size (may bring another gain, but leave it for you to test).
  • All the experiments are done on NVIDIA TESLA T4.

Here are the numbers on several three classical neural networks (MLP, CNN, Bi-RNN, Bi-GRU, Bi-LSTM) with different optimizers (Adam, AdamW, LAMB).

I only list results of batch={64, 128, 1024, 16384}. For full results, please see


Optimizer MLP CNN Bi-RNN Bi-GRU Bi-LSTM Note
Adam 97.03 98.93 96.24 98.92 99.04 Just ordinary Adam
AdamW 97.11 99.01 96.50 99.11 99.04 Used in BERT
LAMB 98.27 99.33 97.73 98.83 98.94 New optimizer for large batch


Optimizer MLP CNN Bi-RNN Bi-GRU Bi-LSTM Note
Adam 96.38 98.76 97.73 99.08 99.09 Just ordinary Adam
AdamW 96.57 98.72 98.05 98.96 99.00 Used in BERT
LAMB 97.90 99.20 98.04 98.87 98.76 New optimizer for large batch


Optimizer MLP CNN Bi-RNN Bi-GRU Bi-LSTM Note
Adam 93.05 97.92 98.10 98.94 98.67 Just ordinary Adam
AdamW 93.67 98.00 98.19 98.86 98.82 Used in BERT
LAMB 97.68 98.82 98.27 98.61 98.47 New optimizer for large batch


Optimizer MLP CNN Bi-RNN Bi-GRU Bi-LSTM Note
Adam 88.46 95.06 95.98 97.81 97.74 Just ordinary Adam
AdamW 91.46 96.57 96.34 98.45 98.39 Used in BERT
LAMB 93.23 97.89 93.76 87.60 80.36 New optimizer for large batch

Several Conclusions

Note: The conclusions are only made by the results above.

  • LAMB consistently outperforms Adam and AdamW in most of the times, and shows consistent results among different batch sizes.
  • LAMB shows big advantage than Adam and AdamW on large batch, showing its excellent scalability.
  • LAMB failed to outperform than Adam and AdamW on complex RNN-based models, despite batch size.


Check mnist_tensorflow.ipynb for details.

Note: You know the GPU/TPU won't get exactly the same results even we use fixed random seed.



For help or issues, please submit a GitHub issue.

You can’t perform that action at this time.