README.md

---
pipeline_tag: text-generation
inference: true
widget:
- text: 'Hello!'
  example_title: Hello world
  group: Python
library_name: transformers
---

This model is randomly initialized, using the config from [EleutherAI/gpt-j-6b](https://huggingface.co/EleutherAI/gpt-j-6b) but with smaller size. 
Note the model is in float16.

Codes:
```python
from transformers import pipeline
from huggingface_hub import create_repo, upload_folder
import torch
import transformers
import os

model_id = 'EleutherAI/gpt-j-6b'
save_path = '/tmp/yujiepan/gptj-tiny-random'
repo_id = 'yujiepan/gptj-tiny-random'

config = transformers.AutoConfig.from_pretrained(model_id)
config.hidden_size = 16
config.n_embd = 16
config.num_attention_heads = 2
config.n_head = 2
config.rotary_dim = 4
config.num_hidden_layers = 2
config.n_layer = 2
config.torch_dtype = torch.float16
print(config)

model = transformers.AutoModelForCausalLM.from_config(config, torch_dtype=torch.float16)
model = model.half()
model.save_pretrained(save_path)

tokenizer = transformers.AutoTokenizer.from_pretrained(model_id)
tokenizer.save_pretrained(save_path)

# from optimum.intel.openvino import OVModelForCausalLM
# ovmodel = OVModelForCausalLM.from_pretrained(save_path, export=True)
# ovmodel.save_pretrained(save_path)

os.system(f'ls -alh {save_path}')
create_repo(repo_id, exist_ok=True)
upload_folder(repo_id=repo_id, folder_path=save_path)
```