Parâmetros do LLM
Como qualquer modelo de aprendizado de máquina, os grandes modelos de linguagem têm vários parâmetros que controlam a variância da saída de texto gerada. Nós começamos uma série de várias partes para explicar detalhadamente o impacto desses parâmetros. Concluiremos com o equilíbrio perfeito na geração de conteúdo usando todos esses parâmetros discutidos na nossa série de várias partes.
Bem-vindo à segunda parte, onde discutimos outro parâmetro bem conhecido, “Top-P”.
Top-P (Amostragem de Núcleo)
Se o objetivo é controlar a diversidade da saída do modelo, então o Top-P é a opção ideal para você. Um valor de Top-P baixo força o modelo a usar as palavras mais prováveis, enquanto um valor de Top-P alto força o modelo a usar palavras mais diversas, aumentando a criatividade.
Vamos ver o Top-P em ação com o código e a saída a seguir.
Python
import torch
from transformers import GPT2LMHeadModel, GPT2Tokenizer
# Load GPT-2 model and tokenizer
tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
model = GPT2LMHeadModel.from_pretrained("gpt2")
# Add pad token to tokenizer (GPT-2 doesn't have it by default)
tokenizer.pad_token = tokenizer.eos_token
# Function to generate response with varying top_p
def generate_with_top_p(prompt, top_p):
inputs = tokenizer(prompt, return_tensors='pt', padding=True)
# Set the attention_mask and pad_token_id
outputs = model.generate(
inputs.input_ids,
attention_mask=inputs['attention_mask'],
do_sample=True,
max_length=200,
top_p=top_p,
pad_token_id=tokenizer.eos_token_id
)
return tokenizer.decode(outputs[0], skip_special_tokens=True)
prompt = "What are some effective ways to manage stress in daily life?"
# List of top-p values and their descriptions
top_p_values = {
0.1: "Very conservative: Generates highly probable and safe responses.",
0.3: "Conservative: Generates probable responses with less risk.",
0.5: "Balanced: A mix of safe and creative responses.",
0.7: "Creative: Generates more diverse and creative responses.",
0.9: "Very creative: Allows for highly diverse and less probable responses."
}
# Test top_p variations
for top_p, description in top_p_values.items():
print(f"\nTop-p {top_p} ({description}):\n")
print(generate_with_top_p(prompt, top_p=top_p))
Saída:
PowerShell
python test_top_p.py
Top-p 0.1 (Very conservative: Generates highly probable and safe responses.):
What are some effective ways to manage stress in daily life?
The following are some of the most common ways to manage stress in daily life.
1. Avoiding stress
The most common way to avoid stress is to avoid it.
The most common way to avoid stress is to avoid it.
2. Avoiding stress
The most common way to avoid stress is to avoid it.
The most common way to avoid stress is to avoid it.
3. Avoiding stress
The most common way to avoid stress is to avoid it.
The most common way to avoid stress is to avoid it.
4. Avoiding stress
The most common way to avoid stress is to avoid it.
The most common way to avoid stress is to avoid it.
5. Avoiding stress
The most common way to avoid stress is to avoid it.
The most common way to avoid stress
Top-p 0.3 (Conservative: Generates probable responses with less risk.):
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
What are some effective ways to manage stress in daily life?
63
What are some effective ways to manage stress in daily life?
64
65
What are some effective ways to manage stress in daily life?
66
67
What are some effective ways to manage stress in daily life?
68
69
What are some effective
70
71
Top-p 0.5 (Balanced: A mix of safe and creative responses.):
72
73
What are some effective ways to manage stress in daily life?
74
75
1. Stay on top of your body's natural stress levels
76
77
When you're stressed, your body's natural stress levels are low.
78
79
If you're stressed, your body's natural stress levels are high.
80
81
If you're stressed, your body's natural stress levels are low.
82
83
2. Avoid excessive exercise
84
85
Exercise can make you feel better.
86
87
Exercise can make you feel better.
88
89
3. Get up early to avoid fatigue
90
91
Exercise can make you feel better.
92
93
Exercise can make you feel better.
94
95
4. Avoid the temptation to take the wrong thing
96
97
Exercise can make you feel better.
98
99
Exercise can make you feel better.
100
101
5. Avoid eating the wrong foods
102
103
Exercise can make you feel better.
104
105
Exercise can make you feel better.
106
107
6. Avoid the temptation to
108
109
Top-p 0.7 (Creative: Generates more diverse and creative responses.):
110
111
What are some effective ways to manage stress in daily life?
112
113
I am talking about a very specific situation. The person I am talking about has been stressed, but has not been doing much work for a long time. I want to tell you, because this person has had a lot of stress in his life, that it is not something you can just go back to. But what I'm trying to say is, that if you don't have a job, you have to go back to work every day, so you can spend more time with your family. So I've been doing that for a long time now. And so, that is a very common occurrence.
114
115
But what do you think is the best way to deal with the stress?
116
117
You know, it's not easy to deal with it. It is very difficult to deal with the stress that we experience. So, that is a very good way to deal with it. So, I think it's the
118
119
Top-p 0.9 (Very creative: Allows for highly diverse and less probable responses.):
120
121
What are some effective ways to manage stress in daily life?
122
123
There are many things that can be done by daily meditation and practice. As a general rule of thumb, meditation can help you stay mindful of your own needs, goals, feelings, desires, emotions, and the life and emotions around you.
124
125
The purpose of meditation is to feel a deep desire to practice more, to be more mindful, and to be more productive. It also serves to enhance your overall well-being.
126
127
1. Be active, be creative, be mindful, and be optimistic.
128
129
This is where the first step towards meditation comes from. If we're looking for inspiration, there's a whole section on being "active" and "creative."
130
131
While I'm not sure I know much about meditation, I know some of its practitioners and some that I never met. My mom used to tell me that she'd always find a way to make her feel more connected and involved.
Agora vamos entender a saída.
- Top-P 0.1 – Muito Conservador:Como o modelo seleciona das palavras seguintes mais prováveis no top 10%, há muita repeticão no conteúdo gerado. Portanto, essa resposta carece de diversidade e é também muitas vezes não informativa.
- Top-P 0.3 – Conservador: O modelo seleciona das primeiras 30% das opções de palavra prováveis seguintes, portanto, é um pouco menos conservador que o anterior conjunto de Top-P. Como você pode ver na saída, isso não melhorou a geração de conteúdo e o prompt foi repetido ao longo da completação. Neste caso, a repetição do prompt significa que a seqüência mais provável após o prompt para o modelo parece ser o próprio prompt.
- Top-P 0.5 – Equilibrado: Aqui você começa a ver o modelo listando algumas estratégias numeradas pela primeira vez. Ainda há alguma repetição nesta configuração, também. Mas o ponto é que neste nível de Top-P, o modelo começa a incorporar uma gama mais ampla de palavras. A saída é uma mistura de conselhos padrão com algumas inconsistências. Este valor de Top-P permite uma melhoria na criatividade, mas ainda luta com a profundidade de informação.
- Top-P 0.7 – Criativo: Neste caso, o modelo pode selecionar de uma gama mais ampla de palavras, e como você pode ver, a resposta está mudando para um estilo narrativo. O conteúdo é mais criativo, já que envolve agora um cenário onde uma pessoa está lidando com estresse. O lado negativo é a perda de foco, já que o foco não era na gestão do estresse, mas nas dificuldades em lidar com o estresse.
- Top-P 0.9 – Muito Criativo: Nesta configuração, o modelo tem acesso a uma ampla gama de vocabulário e ideias, incluindo palavras e conceitos menos prováveis. Essa configuração permite que o modelo use linguagem mais expressiva. Novamente, o lado negativo de ser muito criativo é que o modelo se desvia do prompt na busca por conteúdo rico e variado.
O ponto crítico a serem notados a partir do exercício acima é como o conteúdo muda com a mudança no valor de Top-P. Também dá-nos a ideia de que este parâmetro não é o único que precisa ser tratado para variar o conteúdo e sua relevância.
Agora, vamos olhar para o impacto de Top-P em alguns casos de uso, justamente como na parte anterior desta série sobre “Geração de Histórias Criativas” e “Explicações Técnicas”.
Python
1
import torch
2
from transformers import GPT2LMHeadModel, GPT2Tokenizer
3
4
# Load GPT-2 model and tokenizer
5
tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
6
model = GPT2LMHeadModel.from_pretrained("gpt2")
7
8
# Add pad token to tokenizer (GPT-2 doesn't have it by default)
9
tokenizer.pad_token = tokenizer.eos_token
10
11
# Function to generate response based on top_p
12
def generate_with_top_p(prompt, top_p, max_length=250):
13
inputs = tokenizer(prompt, return_tensors='pt')
14
outputs = model.generate(
15
inputs.input_ids,
16
attention_mask=inputs.attention_mask,
17
do_sample=True,
18
max_length=max_length,
19
top_p=top_p,
20
pad_token_id=tokenizer.eos_token_id,
21
eos_token_id=tokenizer.eos_token_id,
22
no_repeat_ngram_size=2 # Prevents repetition of phrases
23
)
24
return tokenizer.decode(outputs[0], skip_special_tokens=True)
25
26
### USE CASE 1: CREATIVE STORY GENERATION ###
27
28
def creative_story_generation():
29
prompt = ("In the mystical land of Eldoria, a young warrior found an ancient map "
30
"that led to a hidden treasure guarded by dragons. He knew that courage and "
31
"wisdom would be his allies on this perilous journey.")
32
33
# Negative Impact: Low top_p for creative writing (less creative)
34
print("\n=== Creative Story with Low top_p (0.2) - Negative Impact: ===")
35
low_top_p_story = generate_with_top_p(prompt, top_p=0.2)
36
print(low_top_p_story)
37
38
# Perfect Impact: High top_p for creative writing (more creative)
39
print("\n=== Creative Story with High top_p (0.95) - Perfect Impact: ===")
40
high_top_p_story = generate_with_top_p(prompt, top_p=0.95)
41
print(high_top_p_story)
42
43
### USE CASE 2: TECHNICAL EXPLANATION ###
44
45
def technical_explanation():
46
prompt = ("Explain step by step how the internet works, focusing on how computers "
47
"use IP addresses and data packets to communicate with each other.")
48
49
# Negative Impact: High top_p for technical writing (less precise)
50
print("\n=== Technical Explanation with High top_p (0.95) - Negative Impact: ===")
51
high_top_p_explanation = generate_with_top_p(prompt, top_p=0.95)
52
print(high_top_p_explanation)
53
54
# Perfect Impact: Optimal top_p for technical writing (accurate)
55
print("\n=== Technical Explanation with Optimal top_p (0.5) - Perfect Impact: ===")
56
optimal_top_p_explanation = generate_with_top_p(prompt, top_p=0.5)
57
print(optimal_top_p_explanation)
58
59
# Run both use cases
60
creative_story_generation()
61
technical_explanation()
62
Saída:
PowerShell
1
python top_p_multiple.py
2
3
=== Creative Story with Low top_p (0.2) - Negative Impact: ===
4
In the mystical land of Eldoria, a young warrior found an ancient map that led to a hidden treasure guarded by dragons. He knew that courage and wisdom would be his allies on this perilous journey.
5
6
The Dragon King
7
...
8
(The Book of the Dragon)
9
,
10
-
11
: The Dragon Lord is a legendary warrior who has been the focus of many legends. The dragon king is the most powerful of all the dragons in the world. In the magical land, he is known as the "Dragon King". He is also known to be the leader of a group of dragons called the Black Dragons. His name is derived from the dragon's name, "the dragon".
12
"The Black Dragon" is an important symbol of power and powerlessness. It is said that the black dragon is able to create a dragon that can defeat the strongest of his enemies. However, the true power of this dragon lies in his ability to manipulate the minds of others. This ability is called "The Dark Dragon". The Dark dragon has a powerful sense of self-preservation and is capable of manipulating others to his will. When he has control over others, his power is so great that he can destroy entire cities. As a result
13
14
=== Creative Story with High top_p (0.95) - Perfect Impact: ===
15
In the mystical land of Eldoria, a young warrior found an ancient map that led to a hidden treasure guarded by dragons. He knew that courage and wisdom would be his allies on this perilous journey.
16
17
Spirits are like gods. In this world, there are no gods without secrets. There are also no secrets about being a fighter or a thief. But every dragon has a special hidden skill, and he or she can use that skill to destroy and gain strength or hide something hidden in the secret. Many dragons are skilled at their martial arts, while most are unaware of the secrets of their true power. These dragons cannot only use these skills, but that will only allow them to escape the dragons' clutches. Because their training will be tested before they're even born, dragon fighting has never been so hard, even without training, so they should be able to break a dragon's body.
18
19
=== Technical Explanation with High top_p (0.95) - Negative Impact: ===
20
Explain step by step how the internet works, focusing on how computers use IP addresses and data packets to communicate with each other. If a person with the same identity as a user on the US government's private network uses the online address bar, then this data is sent to a server on a computer on your local network. Your IP address is a small byte in the string. The IP and network address are identical. Do you remember, you just want to do that instead of using IPs or numbers. In addition, remember that IP can be used to verify a particular IP for you and your computer. For instance, your name does not always match an address on our government network and you should have your public IP in this country. This does seem quite unusual and perhaps a bit bizarre.
21
22
There was a time in Silicon Valley when you could set your identity out. But in most of today's world, how do you set up your own address and how does one look for it? What about the public? The internet itself was different. It was just a set of rules around data flow that you were supposed to follow. Now, even in today the "internet in general" seems a little more complicated to define. Let's say
23
24
=== Technical Explanation with Optimal top_p (0.5) - Perfect Impact: ===
25
Explain step by step how the internet works, focusing on how computers use IP addresses and data packets to communicate with each other.
26
27
"We've been trying to understand how it works and what it means for the future," says James. "It's not just about the IP address, it's about how people communicate. It's also about what's going on with the data. We want to see how this works. What is the Internet going to look like in the next 10 years?"
28
, the director of the Computer Science and Artificial Intelligence Laboratory at the University of Michigan, says that while there's still a lot of work to be done, "we've got to start to think about it."
Agora vamos desconstruir e analisar a saída para a geração de histórias criativas e explicações técnicas com base nas configurações de Top-P e como a saída foi afetada.
Para demonstrar eficazmente o impacto de Top-P, nós integramos melhores prompts para orientar a saída de modo que o impacto seja observado facilmente.
Geração de Histórias Criativas
- Alto Top-P (Impacto Bom):Como você pode ver com o menor Top-P, o modelo é restrito ao uso de palavras ou expressões e causa, portanto, repetição e redundância. A criatividade é também limitada neste caso, já que o modelo tenta não apresentar ideias novas. Mas se você notar, o fluxo lógico ainda é mantido e o modelo fica no tópico, o que é típico de valores de Top-P baixos.
- Baixo Top-P (Impacto Ruim):Neste caso, o modelo apresenta novos conceitos e adiciona uma angulação criativa à narrativa. A vocabulário é mais amplo, adicionando profundidade e ricaça ao texto. No entanto, devido à criatividade aumentada, o fluxo lógico foi comprometido.
A contraste entre as duas narrativas mostra claramente o impacto de Top-P, fazendo com que seja fácil entender como ele afeta a escrita criativa.
Explicação Técnica
- Alta Top-P (Impacto Negativo): Como podem ver, a alta Top-P afeta negativamente as explicações técnicas, impedindo um fluxo lógico e desviante do tópico. O modelo também introduce informações irrelevantes que não são pertinentes à explicação.
- Top-P Ótima (Impacto Perfeito): Com a Top-P ótima, a explicação é mais coerente e próxima do tópico. O conteúdo se ajusta melhor ao prompt e equilibra bem a precisão e a expressão. A confiabilidade da informação é melhorada, pois o modelo é limitado a palavras mais prováveis.
Conclusão
Com este experimento, mostrámos com sucesso a importância do parâmetro Top-P para controlar a randomness e a criatividade do texto gerado. Primeiro, olhámos para um único prompt e como o resultado varia com diferentes Top-P e depois adoptámos uma abordagem baseada em casos de uso para ver como o Top-P controla o resultado com base no caso de uso.
No entanto, das partes anteriores e desta série, notamos que individualmente, cada parâmetro não é o suficiente para a qualidade de geração de conteúdo. É por isso que é essencial olhar para o impacto de todos esses parâmetros e isso será feito na última parte desta série.
Source:
https://dzone.com/articles/decoding-llm-parameters-top-p