> ## Documentation Index
> Fetch the complete documentation index at: https://training-docs.cerebras.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Convert to Hugging Face

> Learn how to use the Model Zoo CLI to convert to Hugging Face.

Once you have trained a language model on the cluster, you can port it to Hugging Face to generate outputs. For more information, refer to [Convert Checkpoints and Model Configs](../../model-zoo/migration/convert-checkpoints-and-model-configs/convert-checkpoints-and-model-configs).

<Note>
  This guide uses an example from Spring 2023 with CerebrasGPT which uses release 1.8 (cs-1.8) as the source format.
</Note>

## Procedure

1. To use the conversion tool, activate the Cerebras virtual environment and specify the following flags:

| Flag                                             | Description                                                                   |
| ------------------------------------------------ | ----------------------------------------------------------------------------- |
| `--model gpt2`                                   | Model architecture the checkpoint corresponds to                              |
| `--src-fmt cs-1.8`                               | the source format of the checkpoint corresponding to Cerebras Model Zoo(R1.8) |
| `--tgt-fmt hf`                                   | Target format of the checkpoint corresponding to Hugging Face                 |
| `--config custom_config_GPT111M.yaml`            | `yaml` file configuration used for training the model                         |
| `--output-dir hf_dir_train_from_scratch_GPT111M` | Directory containing the output configuration and checkpoint                  |

```bash theme={null}
cd $PARENT_CS
source venv_cerebras_pt/bin/activate
```

```bash theme={null}
cszoo checkpoint convert \
  --model gpt2 \
  --src-fmt cs-1.8 \
  --tgt-fmt hf \
  --config custom_config_GPT111M.yaml \
  --output-dir hf_dir_train_from_scratch_GPT111M \
  train_from_scratch_GPT111M/checkpoint_10.mdl

```

2. Convert the checkpoint obtained from fine-tuning “Cerebras-GPT 111M” included in the model directory finetune\_GPT111M.

```bash theme={null}
cd $PARENT_CS
source venv_cerebras_pt/bin/activate
```

```bash theme={null}
cszoo checkpoint convert \
  --model gpt2 \
  --src-fmt cs-1.8 \
  --tgt-fmt hf \
  --config custom_config_GPT111M.yaml \
  --output-dir hf_dir_finetune_GPT111M \
  finetune_GPT111M/checkpoint_10.mdl

```

3. To facilitate importing the model in Hugging Face, modify the configuration file’s name to include gpt2 in it.

```bash theme={null}
mv hf_dir_train_from_scratch_GPT111M/custom_config_GPT111M_to_hf.json hf_dir_train_from_scratch_GPT111M/config_to_hf_gpt2.json

mv hf_dir_finetune_GPT111M/custom_config_GPT111M_to_hf.json hf_dir_finetune_GPT111M/config_to_gpt2.json
```

Now the trained models can be used with Hugging Face.

4. Create a Python virtual environment to use Hugging Face.

Before you proceed, deactivate the CS virtual environment.

```bash theme={null}
python -m venv hf_env
```

```bash theme={null}
source hf_env/bin/activate
```

```bash theme={null}
pip install 'transformers[torch]'
```

```bash theme={null}
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('we love y

```

5. Once you have set up the virtual environment, you can now generate outputs using Hugging Face. The tokenizer can be found in Cerebras-GPT-111M available in Hugging Face. Here is an example using the model trained from scratch:

```bash theme={null}
python
```

```python theme={null}
Python 3.8.13 (default, Aug 16 2022, 12:16:29)
[GCC 9.3.1 20200408 (Red Hat 9.3.1-2)] on linux
Type "help", "copyright", "credits" or "license" for more information.

>>> from transformers import AutoTokenizer, AutoModelForCausalLM, AutoConfig

>>>  from transformers import pipeline

>>>  tokenizer = AutoTokenizer.from_pretrained("cerebras/Cerebras-GPT-111M")

>>>  config = AutoConfig.from_pretrained("./hf_dir_train_from_scratch_GPT111M/config_to_hf_gpt2.json")

>>>  model = AutoModelForCausalLM.from_pretrained(pretrained_model_name_or_path="./hf_dir_train_from_scratch_GPT111M/checkpoint_10_to_hf.bin", config = config)

>>>  text = "Generative AI is "

>>>  pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

>>>  generated_text = pipe(text, max_length=50, do_sample=False, no_repeat_ngram_size=2)[0]

Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.

>>>  print(generated_text['generated_text'])

Generative AI is  tossed separatist separatist,, separatist Fantasy, heading, Po, green, confession confession Po Po?!", 113, Pitch, counselormot newfound,ioch confession, Christopher, newfoundmotmot confessionrealDonaldTrump,Vict,icity confessionmot resear

>>>  exit()
```

Here is an example using the model **fine-tuned using Cerebras-GPT checkpoint:**

```python theme={null}
Python 3.8.13 (default, Aug 16 2022, 12:16:29)
[GCC 9.3.1 20200408 (Red Hat 9.3.1-2)] on linux
Type "help", "copyright", "credits" or "license" for more information.

>>> from transformers import AutoTokenizer, AutoModelForCausalLM, AutoConfig

>>> from transformers import pipeline

>>> tokenizer = AutoTokenizer.from_pretrained("cerebras/Cerebras-GPT-111M")

>>> config = AutoConfig.from_pretrained("./hf_dir_finetune_GPT111M/config_to_hf_gpt2.json")

>>> model = AutoModelForCausalLM.from_pretrained("./hf_dir_finetune_GPT111M/checkpoint_10_to_hf.bin", config = config)

>>> text = "Generative AI is "

>>> pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

>>> generated_text = pipe(text, max_length=50, do_sample=False, no_repeat_ngram_size=2)[0]

Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.

>>> print(generated_text['generated_text'])

Generative AI is

The following is a list of the most common types of AI in the world.
AI is the number of human AI classes in which the AI class is defined. The AI type is
the number that the human

>>> exit()
```