[Python] TypeError: PPOTrainer.__init__() got an unexpected keyword argument 'config'

I am trying to initialize a PPO_trainer but have issues.

from trl import PPOTrainer, PPOConfig

ppo_config = PPOConfig(
batch_size=4,
learning_rate=1e-5,
mini_batch_size=2,
use_cpu=True
)

ppo_trainer = PPOTrainer(
config=ppo_config,
model=model,
tokenizer=tokenizer,
train_dataset=dataset,
)


And keep getting all kinds of errors regarding the arguments, but the most common one is TypeError: PPOTrainer.__init__() got an unexpected keyword argument 'config'

I am using trl==0.4.7.

Thanks.

Continue reading...