> ## 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 CS Checkpoints for GPUs

> Learn how to work with Cerebras-format checkpoints, including how to load, convert, and reuse them in your training workflows.

## PyTorch Checkpoint Format

Our large model-optimized checkpoint format is based off the standard [HDF5 file format](https://docs.h5py.org/en/stable/index.html). At a high-level, when saving a checkpoint, the Cerebras stack will take a PyTorch state dictionary, flatten it, and store it in an HDF5 file.

For example, the following state dictionary is flattened and stored into the H5 file as follows:

```python theme={null}
{
    "a": {
        "b": 0.1,
        "c": 0.001,
    },
    "d": [0.1, 0.2, 0.3]
}

```

Flattened H5:

```python theme={null}
{
    "a.b": 0.1,
    "a.c": 0.001,
    "d.0": 0.1,
    "d.1": 0.2,
    "d.2": 0.3,
}

```

A model/optimizer state dictionary can be saved in the new checkpoint format using the `cstorch.save` method:

```python theme={null}
import cerebras.pytorch as cstorch

...

state_dict = {
    "model": model.state_dict(),
    "optimizer": optimizer.state_dict(),
}
cstorch.save(state_dict, "path/to/checkpoint")

...

```

A checkpoint saved using the above can be loaded using the `cstorch.load` method:

```python theme={null}
import cerebras.pytorch as cstorch

...

state_dict = cstorch.load("path/to/checkpoint")

model.load_state_dict(state_dict["model"])
optimizer.load_state_dict(state_dict["optimizer"])

...
```

### Convert to Pickle Format

If using `cstorch.load` is not a sufficient solution for loading the checkpoint into memory, a simple conversion can be done to the pickle format:

```python theme={null}
import torch
import cerebras.pytorch as cstorch

state_dict = cstorch.load("path/to/checkpoint")
torch.save(state_dict, "path/to/new/checkpoint")

```