Parametri dell’LM
Come qualsiasi modello di apprendimento automatico, i grandi modelli di linguaggio hanno diversi parametri che controllano la varianza dell’output del testo generato. Abbiamo iniziato una serie multipart per spiegare in dettaglio l’impatto di questi parametri. Concluderai con il bilanciamento perfetto nella generazione del contenuto usando tutti questi parametri discussi nella nostra serie multipart.
Benvenuti nel secondo capitolo, dove discutiamo di un altro parametro ben noto, “Top-P”.
Top-P (Sampling Nucleus)
Se il tuo obiettivo è controllare la diversità dell’output del modello, allora Top-P è la scelta giusta per te. Un basso Top-P costringe il modello ad usare le parole più probabili, mentre un alto Top-P costringe il modello ad usare parole più diverse, aumentando la creatività.
Osserviamo Top-P in azione con il seguente codice e output.
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))
Output:
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.
Adesso spiegheremo l’output.
- Top-P 0.1 – Molto conservatore: Poiché il modello sceglie dalla top 10% delle scelte di parola probabili seguenti, ci sono molte ripetizioni nel contenuto generato. Quindi questa risposta manca di diversità e anche di informazione la maggior parte del tempo.
- Top-P 0.3 – Conservatore: Il modello sceglie dalle prime 30% delle scelte di parola probabili successive, quindi è leggermente meno conservatore della precedente impostazione Top-P. Come vedete dall’output, questo non ha migliorato la generazione del contenuto, e il prompt è stato ripetuto più volte nella completazione. In questo caso, la ripetizione del prompt significa che la prosecuzione più probabile dopo il prompt per il modello sembra essere il prompt stesso.
- Top-P 0.5 – Equilibrato: Ecco dove vedete il modello elencare alcune strategie numerate per la prima volta. Anche in questa impostazione si nota ancora qualche ripetizione. Ma il punto fondamentale è che a questa impostazione Top-P, il modello comincia a incorporare una gamma più ampia di parole. L’output è una miscela di consigli standard con alcune incongruenze. Questo valore Top-P consente una migliore creatività ma ancora soffre di una profondità di informazione limitata.
- Top-P 0.7 – Creativo: In questo caso, il modello può scegliere da una gamma più ampia di parole, e come vedete, la risposta si sta spostando verso uno stile narrativo. Il contenuto è più creativo poiché ora coinvolge una situazione in cui una persona sta affrontando stress. Il lato negativo è la perdita di focus, poiché l’emphasis non era sulla gestione dello stress ma sulle difficoltà nell’affrontarlo.
- Top-P 0.9 – Molto Creativo: In questa impostazione, il modello ha accesso a una vasta gamma di vocabolari e idee, inclusi parole e concetti meno probabili. Questa impostazione ha permesso al modello di usare un linguaggio più espressivo. Ancora una volta, il lato negativo dell’essere molto creativo è che il modello si allontana dal prompt nella ricerca di produrre contenuti ricchi e variegati.
La cosa fondamentale da notare dall’esercizio precedente è come il contenuto cambia con il cambiamento della impostazione Top-P. Dà anche un’idea che questo parametro non è l’unico che deve essere gestito per la variazione nel contenuto e nella sua rilevanza.
Adesso, consideriamo l’impatto di Top-P su alcuni casi d’uso, proprio come nella parte precedente di questa serie su “Generazione di Storie Creative” e “Esplicazioni Tecniche”.
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
Output:
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."
Adesso, analizziamo in dettaglio l’output per la generazione di storie creative e le spiegazioni tecniche in base alle impostazioni di Top-P e come queste hanno influenzato l’output.
Per dimostrare efficacemente l’impatto di Top-P, abbiamo integrato prompt migliori per guidare l’output in modo da renderlo facile da osservare.
Generazione di Storie Creative
- Impostazione Top-P bassa (Impatto negativo): Come vedete con l’alta Top-P, il modello è limitato nell’uso di parole o frasi e quindi causa ripetizione e redundanza. La creatività è anche limitata in questo caso, poiché il modello cerca di non introdurre nuove idee. Ma se notate, la logica del flusso è ancora mantenuta e il modello rimane sul topic, caratteristica tipica di valori bassi di Top-P.
- Impostazione Top-P alta (Impatto perfetto): In questo caso, il modello introduce nuovi concetti e aggiunge un angolo creativo alla narrazione. Viene utilizzata una più ampia vocabolario, aggiungendo profondità e ricchezza al testo. Tuttavia, a causa dell’aumentata creatività, il flusso logico è stato limitato.
La contrapposizione tra le due narrazioni chiaramente mostra l’impatto di Top-P, rendendolo facile da capire come influenza la scrittura creativa.
Spiegazione Tecnica
- Alto Top-P (Impatto Negativo):Come potete vedere, un alto Top-P ha un impatto negativo sulla spiegazione tecnica, impedendo un flusso logico e deviando dalla discussione principale. Il modello introduce anche informazioni non pertinenti alla spiegazione.
- Top-P Ottimale (Impatto perfetto):Con un Top-P ottimale, l’esposizione è più coerente e vicina al topic. Il contenuto si allinea meglio con il prompt e equilibra bene accuratezza e espressione. La affidabilità delle informazioni viene incrementata perché il modello è limitato a parole più probabili.
Conclusione
Con questo esperimento, abbiamo dimostrato con successo l’importanza del parametro Top-P nel controllare la casualità e la creatività del testo generato. Iniziammo guardando un singolo prompt e come l’output varia con il variare del Top-P, poi abbiamo adottato un approcio basato su casi d’uso per capire come Top-P controlla l’output in base all’uso.
Tuttavia, dalla parte precedente e da questa parte della serie, abbiamo notato che individualmente, ciascun parametro non riesce a garantire una generazione di contenuti di qualità adeguata. Per questo motivo, è importante considerare l’impatto di tutti questi parametri, e lo faremo come parte finale di questa serie.
Source:
https://dzone.com/articles/decoding-llm-parameters-top-p