# Request: Mask-LM Training Google Colab
**URL:**
**Author:**  [@zanderbush](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fzanderbush)%5C
**Post date:** [November 19, 2020, 2:45am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Frequest-mask-lm-training-google-colab%2F2119%2F1 "2020-11-19T02:45:38Z")
Does anyone have a Google Colab notebook they can share? (If possible, one that has been adapted to the new model upload)
---
**Author:**  [@Jung](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FJung)%5C
**Post date:** [November 19, 2020, 3:57am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Frequest-mask-lm-training-google-colab%2F2119%2F2 "2020-11-19T03:57:00Z")
Hi, not exactly what you requested (ie. not the latest Huggingface version),
but it should give the main idea with a bonus on working with TPU:
Kaggle notebook on **implementation of MLM finetuning on TPUs with a custom training loop in TensorFlow 2**.
[https://www.kaggle.com/riblidezso/finetune-xlm-roberta-on-jigsaw-test-data-with-mlm](https://www.kaggle.com/riblidezso/finetune-xlm-roberta-on-jigsaw-test-data-with-mlm)
---
**Author:**  [@zanderbush](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fzanderbush)%5C
**Post date:** [November 20, 2020, 2:10am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Frequest-mask-lm-training-google-colab%2F2119%2F3 "2020-11-20T02:10:47Z")
Thank you!