---
language:
  - en
license: apache-2.0
library_name: transformers
tags:
  - open-reason
  - causal-lm
  - cpu
datasets:
  - theworker02/open-reason
base_model: gpt2-scratch
---

# Open Reason small (CPU)

This is a **small** GPT-2-style causal LM trained from scratch on the
Open Reason SFT split. It is **not** a 1B model and is **not**
`theworker02/open-reason-1b`.

- Parameters (approx): 1334016
- Steps: 200
- Backend: cpu-host
- CUDA used: False
- Rows: 2348
- Final loss: 6.095415115356445

Hardware: CPU (Docker when available). AMD GPU training is not used.

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
tok = AutoTokenizer.from_pretrained("theworker02/open-reason-small")
model = AutoModelForCausalLM.from_pretrained("theworker02/open-reason-small")
```