|
--- |
|
license: apache-2.0 |
|
tags: |
|
- trl |
|
- transformers |
|
- reinforcement-learning |
|
--- |
|
|
|
# TRL Model |
|
|
|
This is a [TRL language model](https://github.com/huggingface/trl) that has been fine-tuned with reinforcement learning to |
|
guide the model outputs according to a value, function, or human feedback. The model can be used for text generation. |
|
This was used as a test model in the reward interpretability study at https://arxiv.org/abs/2310.08164. |
|
|
|
## Usage |
|
|
|
To use this model for inference, first install the TRL library: |
|
|
|
```bash |
|
python -m pip install trl |
|
``` |
|
|
|
You can then generate text as follows: |
|
|
|
```python |
|
from transformers import pipeline |
|
|
|
generator = pipeline("text-generation", model="amirabdullah19852020//tmp/tmpz1w2usia/amirabdullah19852020/pythia-70m_sentiment_reward") |
|
outputs = generator("Hello, my llama is cute") |
|
``` |
|
|
|
If you want to use the model for training or to obtain the outputs from the value head, load the model as follows: |
|
|
|
```python |
|
from transformers import AutoTokenizer |
|
from trl import AutoModelForCausalLMWithValueHead |
|
|
|
tokenizer = AutoTokenizer.from_pretrained("amirabdullah19852020//tmp/tmpz1w2usia/amirabdullah19852020/pythia-70m_sentiment_reward") |
|
model = AutoModelForCausalLMWithValueHead.from_pretrained("amirabdullah19852020//tmp/tmpz1w2usia/amirabdullah19852020/pythia-70m_sentiment_reward") |
|
|
|
inputs = tokenizer("Hello, my llama is cute", return_tensors="pt") |
|
outputs = model(**inputs, labels=inputs["input_ids"]) |
|
``` |
|
|
