dshin's picture
Push model using huggingface_hub.
license: apache-2.0
- trl
- transformers
- reinforcement-learning
# TRL Model
This is a [TRL language model](https://github.com/lvwerra/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.
## Usage
To use this model for inference, first install the TRL library:
python -m pip install trl
You can then generate text as follows:
from transformers import pipeline
generator = pipeline("text-generation", model="dshin//tmp/tmpxz2bcbbl/dshin/flan-t5-ppo-user-h-batch-size-8-epoch-1")
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:
from transformers import AutoTokenizer
from trl import AutoModelForCausalLMWithValueHead
tokenizer = AutoTokenizer.from_pretrained("dshin//tmp/tmpxz2bcbbl/dshin/flan-t5-ppo-user-h-batch-size-8-epoch-1")
model = AutoModelForCausalLMWithValueHead.from_pretrained("dshin//tmp/tmpxz2bcbbl/dshin/flan-t5-ppo-user-h-batch-size-8-epoch-1")
inputs = tokenizer("Hello, my llama is cute", return_tensors="pt")
outputs = model(**inputs, labels=inputs["input_ids"])