/ Модель

Характеристики

Контекст
32 768 токенов
Параметры
142 млрд
Размер весов
286 ГБ
Лицензия
Своя лицензия
Архитектура
qwen2 · dense
Базовая модель
Qwen2.5-72B-Instruct
Библиотека
transformers
Опубликована
23 сен 2024

Qwen2.5-143B-Doubled72B-Instruct-Mergekit-Merge

Генерация текстаТекст → текст

Экспериментальная 141,7-миллиардная модель Solshine, полученная дублированием слоёв Qwen2.5-72B-Instruct для генерации текста и диалогов.

119скачиваний за 30 дней2лайкавеса обновлены17 окт 2024

Бенчмарки

Как модель справилась с задачами пользователей ailist и что о ней публикуют открытые лидерборды

Тесты пользователейпока нет

Никто ещё не опубликовал тест этой модели. Соберите бенчмарк из своих задач — результат появится здесь.

  1. Добавьте модель в бенчмарк

    Вместе с другими моделями категории «Текст» — сравнение идёт внутри одной категории.

  2. Опишите свои задачи

    Напишите запросы, при необходимости приложите файлы и эталонные ответы.

  3. Запустите и опубликуйте

    Модели ответят на каждую задачу, судья оценит ответы. Опубликованный результат появится на этой странице.

Публичные оценки

Из открытых лидербордов. Сравнивайте результаты одного теста при одинаковых условиях — отсутствие оценки не значит, что модель хуже.

В подключённых источниках результатов для этой модели пока нет.

Что умеет модель

Описание составлено по карточке модели и метаданным Hugging Face

Solshine создал экспериментальную языковую модель слиянием через mergekit на основе Qwen/Qwen2.5-72B-Instruct. Метод passthrough дублирует слои исходной модели по порядку; итоговый чекпойнт содержит 141,7 млрд параметров. После слияния модель не дообучали, а тестирование было минимальным.

Модель предназначена для диалогов и генерации текста, включая структурированные ответы, программирование и математику — задачи, поддерживаемые базовой Qwen2.5-72B-Instruct. Контекст составляет до 32 768 токенов; доступны интерфейсы чата и вызова инструментов. Качество и поведение модели после дублирования слоёв не подтверждены широкой оценкой. В метаданных указана лицензия «other», поэтому конкретные условия использования следует проверять в карточке репозитория.

Сценарии применения

  • ведение диалогов с инструкционными ответами
  • генерация и редактирование текстов
  • подготовка структурированных текстовых ответов
  • помощь с программированием
  • решение математических задач

Ограничения

Слияние не проходило дополнительного дообучения, а модель прошла лишь минимальное тестирование, поэтому её качество и стабильность не подтверждены. Указана лицензия «other»; условия использования следует уточнить в карточке репозитория.

Карточка модели

README от автора на Hugging Face, как он его опубликовал

Открыть на Hugging Face

Qwen2.5-143B-Doubled72B-Instruct-Mergekit-Merge by Solshine (Caleb DeLeeuw)

image/png

merge

This is a merge of pre-trained language models created using mergekit.

This model recieved no post merge retraining (yet) and minimal testing. Please contribute any feedback or evaluations of any kind via the community tab.

License

Hippocratic License 3.0 + Ecocide module, + Extractive Industries module, + Copyleft Hippocratic License HL3-CL-ECO-EXTR https://firstdonoharm.dev/version/3/0/cl-eco-extr.txt

Merge Details

Merge Method

This model was merged using the passthrough merge method. Every layer is doubled in order, from Qwen/Qwen2.5-72B-Instruct, creating 143B parameters. No additional fine-tune has been done in this merged model.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [0, 1]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [0, 1]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [1, 2]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [1, 2]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [2, 3]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [2, 3]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [3, 4]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [3, 4]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [4, 5]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [4, 5]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [5, 6]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [5, 6]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [6, 7]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [6, 7]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [7, 8]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [7, 8]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [8, 9]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [8, 9]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [9, 10]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [9, 10]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [10, 11]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [10, 11]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [11, 12]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [11, 12]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [12, 13]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [12, 13]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [13, 14]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [13, 14]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [14, 15]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [14, 15]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [15, 16]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [15, 16]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [16, 17]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [16, 17]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [17, 18]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [17, 18]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [18, 19]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [18, 19]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [19, 20]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [19, 20]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [20, 21]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [20, 21]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [21, 22]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [21, 22]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [22, 23]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [22, 23]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [23, 24]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [23, 24]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [24, 25]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [24, 25]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [25, 26]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [25, 26]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [26, 27]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [26, 27]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [27, 28]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [27, 28]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [28, 29]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [28, 29]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [29, 30]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [29, 30]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [30, 31]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [30, 31]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [31, 32]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [31, 32]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [32, 33]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [32, 33]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [33, 34]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [33, 34]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [34, 35]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [34, 35]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [35, 36]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [35, 36]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [36, 37]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [36, 37]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [37, 38]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [37, 38]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [38, 39]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [38, 39]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [39, 40]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [39, 40]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [40, 41]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [40, 41]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [41, 42]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [41, 42]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [42, 43]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [42, 43]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [43, 44]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [43, 44]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [44, 45]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [44, 45]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [45, 46]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [45, 46]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [46, 47]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [46, 47]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [47, 48]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [47, 48]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [48, 49]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [48, 49]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [49, 50]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [49, 50]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [50, 51]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [50, 51]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [51, 52]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [51, 52]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [52, 53]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [52, 53]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [53, 54]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [53, 54]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [54, 55]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [54, 55]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [55, 56]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [55, 56]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [56, 57]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [56, 57]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [57, 58]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [57, 58]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [58, 59]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [58, 59]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [59, 60]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [59, 60]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [60, 61]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [60, 61]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [61, 62]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [61, 62]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [62, 63]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [62, 63]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [63, 64]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [63, 64]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [64, 65]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [64, 65]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [65, 66]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [65, 66]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [66, 67]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [66, 67]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [67, 68]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [67, 68]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [68, 69]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [68, 69]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [69, 70]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [69, 70]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [70, 71]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [70, 71]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [71, 72]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [71, 72]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [72, 73]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [72, 73]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [73, 74]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [73, 74]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [74, 75]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [74, 75]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [75, 76]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [75, 76]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [76, 77]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [76, 77]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [77, 78]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [77, 78]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [78, 79]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [78, 79]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [79, 80]
  - sources:
      - model: Qwen/Qwen2.5-72B-Instruct
        layer_range: [79, 80]
merge_method: passthrough
dtype: float16

Как запустить

На своей GPU в ailist. Цена — по текущим предложениям провайдеров

GPU в ailist

vLLM

Можно запустить

Подбираем GPU…

Класс архитектуры
Qwen2ForCausalLM
Ревизия весов
bc6709e

0 моделей