Exploring pre-trained LLMs

Reinforcement Learning from Human Feedback (RLHF)

Mina Parham

AI Engineer

The importance of fine-tuning

A diagram illustrating the RLHF process.

Reinforcement Learning from Human Feedback (RLHF)

The importance of fine-tuning

A diagram representing an LLM with input and output.

Reinforcement Learning from Human Feedback (RLHF)

A step-by-step guide to fine-tuning an LLM

An icon representing an LLM trained to recognize the sentiment of tweets.

Reinforcement Learning from Human Feedback (RLHF)

A step-by-step guide to fine-tuning an LLM

An icon representing an LLM trained to recognize the sentiment of tweets giving the wrong output.

Reinforcement Learning from Human Feedback (RLHF)

A step-by-step guide to fine-tuning an LLM

An icon representing an LLM for recognizing the sentiment of tweets, ore-trained with a large dataset.

Reinforcement Learning from Human Feedback (RLHF)

Step 1: load the data to use

from datasets import load_dataset
import pandas as pd

# `load_dataset` simplifies loading and preprocessing datasets from various sources
# It provides easy access to a wide range of datasets with minimal setup
dataset = load_dataset("mteb/tweet_sentiment_extraction")
df = pd.DataFrame(dataset['train'])

    id               text                                        label   label_text
0   cb774db0d1       I'd have responded, if I were going         1       neutral
1   549e992a42       Sooo SAD I will miss you in San Diego!!!    0       negative
2   08ac60f138       my boss is bullying me...                   0       negative
Reinforcement Learning from Human Feedback (RLHF)

Step 2: choose a pre-trained model

from transformers import AutoModelForCausalLM

# AutoModelForCausalLM simplifies loading and switching models
model = AutoModelForCausalLM.from_pretrained("openai-gpt")

 

  • Causal models: previous tokens "cause" subsequent ones
Reinforcement Learning from Human Feedback (RLHF)

Step 3: tokenizer

from transformers import AutoTokenizer

# `AutoTokenizer` loads the correct tokenizer for the specified model
tokenizer = AutoTokenizer.from_pretrained("openai-gpt")
tokenizer.add_special_tokens({'pad_token': '[PAD]'})
model.resize_token_embeddings(len(tokenizer))

 

  • Padding: to have equal-sized batches of text
Reinforcement Learning from Human Feedback (RLHF)

Step 3: tokenizer

def tokenize_function(examples):
    tokenized = tokenizer(examples["content"], padding="max_length", truncation=True) 
    return tokenized

tokenized_datasets = dataset.map(tokenize_function, batched=True)

 

  • Batched parameter: for faster processing
Reinforcement Learning from Human Feedback (RLHF)

Step 4: fine-tune using the Trainer method

training_args = TrainingArguments(
   output_dir="test_trainer",
   per_device_train_batch_size=1,
   per_device_eval_batch_size=1,  
   gradient_accumulation_steps=4)

trainer = Trainer(
   model=model,
   args=training_args,
   train_dataset=tokenized_dataset["train"],
   eval_dataset=tokenized_dataset["test"])
trainer.train()
Reinforcement Learning from Human Feedback (RLHF)

Let's practice!

Reinforcement Learning from Human Feedback (RLHF)

Preparing Video For Download...