File size: 1,451 Bytes
f07371a f87bf24 f07371a 91e805d f07371a 91e805d f07371a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 |
---
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"])
```
|