---
title: gpt2-from-scratch-pytorch
canonical_url: "https://www.modelscope.cn/models/AI-ModelScope/gpt2-from-scratch-pytorch"
md_url: "https://www.modelscope.cn/models/AI-ModelScope/gpt2-from-scratch-pytorch.md"
repository: AI-ModelScope/gpt2-from-scratch-pytorch
last_updated: 2025-04-02
license: mit
parameters: 3.2B
tensor_type:
  - F32
library_name:
  - pytorch
  - safetensors
frameworks:
  - pytorch
language:
  - en
downloads: 498
stars: 1
tags:
  - pytorch
  - gpt
  - gpt-2
---

# gpt2-from-scratch-pytorch

> gpt2-from-scratch-pytorch - AI-ModelScope 在 ModelScope 开源的模型。The original GPT-2 model weights from https://openaipublic.blob.core.windows.net/gpt-2/models converted from TensorFlow to PyTorch state dicts and PyTorch safetensors files.

AI-ModelScope/gpt2-from-scratch-pytorch 是 ModelScope 魔搭社区上的 3.2B 参数机器学习模型，采用 mit 许可。

- **Repository**: AI-ModelScope/gpt2-from-scratch-pytorch
- **License**: mit
- **Parameters**: 3.2B
- **Tags**: pytorch, gpt, gpt-2
- **Downloads**: 498
- **Stars**: 1
- **Last updated**: 2025-04-02

Source: https://www.modelscope.cn/models/AI-ModelScope/gpt2-from-scratch-pytorch

---

# GPT-2 PyTorch



The original GPT-2 model weights from [https://openaipublic.blob.core.windows.net/gpt-2/models](https://openaipublic.blob.core.windows.net/gpt-2/models) converted from TensorFlow to PyTorch state dicts and PyTorch safetensors files.


 
## Usage

The section below explain how the model weights can be used.

 
### Setup

Install the [llms-from-scratch](https://pypi.org/project/llms-from-scratch/) PyPI package:

```python
pip install llms_from_scratch
```

Or copy the and paste the `GPTModel` class and dependencies from [GitHub](https://github.com/rasbt/LLMs-from-scratch/blob/main/ch05/01_main-chapter-code/previous_chapters.py).


 
### Loading the model weights

The following shows how to load the weights into the 355M parameter model:

```python
import torch
from llms_from_scratch.ch04 import GPTModel

GPT_CONFIG_BASE = {
    "vocab_size": 50257,    # Vocabulary size
    "context_length": 1024, # Original context length
    "emb_dim": 768,         # Embedding dimension
    "n_heads": 12,          # Number of attention heads
    "n_layers": 12,         # Number of layers
    "drop_rate": 0.0,       # Dropout rate
    "qkv_bias": True        # Query-key-value bias
}

model_configs = {
    "gpt2-small (124M)": {"emb_dim": 768, "n_layers": 12, "n_heads": 12},
    "gpt2-medium (355M)": {"emb_dim": 1024, "n_layers": 24, "n_heads": 16},
    "gpt2-large (774M)": {"emb_dim": 1280, "n_layers": 36, "n_heads": 20},
    "gpt2-xl (1558M)": {"emb_dim": 1600, "n_layers": 48, "n_heads": 25},
}

model_name = "gpt2-medium (355M)"  # Example model name
NEW_CONFIG = GPT_CONFIG_BASE.copy()
NEW_CONFIG.update(model_configs[model_name])

model = GPTModel(NEW_CONFIG)

# Option A: state dict
model.load_state_dict(torch.load("gpt2-medium-355M.pth", weights_only=True));
model.eval();

# Option B: safetensors
# from safetensors.torch import load_file
# model.load_state_dict(load_file("gpt2-medium-355M.safetensors"))

model.eval();
```

To use the other models, simply replace the model names:

```
model_name = "gpt2-medium (355M)"
...
model.load_state_dict(torch.load("gpt2-medium-355M.pth"))
# or
model.load_state_dict(load_file("gpt2-medium-355M.safetensors"))
```

with the desired model names. For example:

```
model_name = "gpt2-small (124M)"
...
model.load_state_dict(torch.load("gpt2-small-124M.pth"))
# or
model.load_state_dict(load_file("gpt2-small-124M.safetensors"))
```


 
### Generating text

The following showcases how the model can then be used to generate text.

```python
import tiktoken
from llms_from_scratch.ch04 import generate_text_simple

tokenizer = tiktoken.get_encoding("gpt2")

prompt = "Ever effort moves"
enc_prompt = tokenizer.encode(prompt)
enc_prompt = torch.tensor([enc_prompt])

token_ids = generate_text_simple(
    model=model,
    idx=enc_prompt, 
    max_new_tokens=25, 
    context_size=NEW_CONFIG["context_length"]
)

output = tokenizer.decode(token_ids.squeeze().tolist())
print(output)
```

```
Ever effort moves the needle.

The first step is to understand the difference between a "good" and a "bad" goal.
```