# Creating custom model

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fcreating-custom-model%2F6531%3E%5C
**Category:** Beginners\
**Created:** [June 2, 2021, 4:28pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fcreating-custom-model%2F6531 "2021-06-02T16:28:20Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Roman](https://avatars.discourse-cdn.com/v4/letter/r/ed8c4c/32.png) [@Roman](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FRoman)%5C
**Post date:** [June 2, 2021, 4:28pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fcreating-custom-model%2F6531%2F1 "2021-06-02T16:28:20Z")

</div>

Hello.  
I’d like to train model on my custom dataset(I don’t want to use pretrained tokenizer and model). According to my understanding I should do following:

1. Selecting desired network architecture - for example let it be ‘ibert’
2. Get coresponding config file: config = AutoConfig.for\_model(‘ibert’)
3. Create model from config: model = AutoModel.from\_config(config)
4. Create tokenizer  
Here is first questin - can I create and train any tokenizer from ‘tokenizers’ pakage when I train own model from scratch?  
The second question - if I’d like to get tokenizer config from existing model, for example ‘allenai/scibert\_scivocab\_uncased’ - how can I do it? I don’t need pretrained tokenizer - I’d like to train it on my own dataset from scratch?