Delete README.md
Browse filesРасчет каскада с ОЭ BD235
README.md
DELETED
|
@@ -1,130 +0,0 @@
|
|
| 1 |
-
---
|
| 2 |
-
library_name: transformers
|
| 3 |
-
model_name: Vikhr-Gemma-2B-instruct
|
| 4 |
-
base_model:
|
| 5 |
-
- google/gemma-2-2b-it
|
| 6 |
-
language:
|
| 7 |
-
- ru
|
| 8 |
-
license: apache-2.0
|
| 9 |
-
datasets:
|
| 10 |
-
- Vikhrmodels/GrandMaster-PRO-MAX
|
| 11 |
-
---
|
| 12 |
-
|
| 13 |
-
# 💨 Vikhr-Gemma-2B-instruct
|
| 14 |
-
|
| 15 |
-
#### RU
|
| 16 |
-
|
| 17 |
-
Мощная инструктивная модель на основе Gemma 2 2B, обученная на русскоязычном датасете GrandMaster-PRO-MAX.
|
| 18 |
-
|
| 19 |
-
#### EN
|
| 20 |
-
|
| 21 |
-
A powerful instructive model based on Gemma 2 2B, trained on the Russian-language dataset GrandMaster-PRO-MAX.
|
| 22 |
-
|
| 23 |
-
## GGUF
|
| 24 |
-
|
| 25 |
-
- [Vikhrmodels/Vikhr-Gemma-2B-instruct-GGUF](https://huggingface.co/Vikhrmodels/Vikhr-Gemma-2B-instruct-GGUF)
|
| 26 |
-
|
| 27 |
-
## Особенности:
|
| 28 |
-
|
| 29 |
-
- 📚 Основа / Base: [gemma-2-2b-it](https://huggingface.co/google/gemma-2-2b-it)
|
| 30 |
-
- 🇷🇺 Специализация / Specialization: **RU**
|
| 31 |
-
- 💾 Датасет / Dataset: [GrandMaster-PRO-MAX](https://huggingface.co/datasets/Vikhrmodels/GrandMaster-PRO-MAX)
|
| 32 |
-
|
| 33 |
-
## Попробовать / Try now:
|
| 34 |
-
|
| 35 |
-
[](https://colab.research.google.com/drive/1htw3x1OS73vIJrMYvdQfflGg4ASdGg9P)
|
| 36 |
-
|
| 37 |
-
## Описание:
|
| 38 |
-
|
| 39 |
-
#### RU
|
| 40 |
-
|
| 41 |
-
Vikhr-Gemma-2B-instruct — это мощная и компактная языковая модель, обученная на датасете GrandMaster-PRO-MAX, специально доученная для обработки русского языка.
|
| 42 |
-
|
| 43 |
-
#### EN
|
| 44 |
-
|
| 45 |
-
Vikhr-Gemma-2B-instruct is a powerful and compact language model trained on the GrandMaster-PRO-MAX dataset, specifically designed for processing the Russian language.
|
| 46 |
-
|
| 47 |
-
## Пример кода для запуска / Sample code to run:
|
| 48 |
-
|
| 49 |
-
```python
|
| 50 |
-
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 51 |
-
|
| 52 |
-
# Загрузка модели и токенизатора
|
| 53 |
-
model_name = "Vikhrmodels/Vikhr-Gemma-2B-instruct"
|
| 54 |
-
model = AutoModelForCausalLM.from_pretrained(model_name)
|
| 55 |
-
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 56 |
-
|
| 57 |
-
# Подготовка входного текста
|
| 58 |
-
input_text = "Напиши стихотворение о весне в России."
|
| 59 |
-
|
| 60 |
-
# Токенизация и генерация текста
|
| 61 |
-
input_ids = tokenizer.encode(input_text, return_tensors="pt")
|
| 62 |
-
output = model.generate(input_ids, max_length=200, num_return_sequences=1, no_repeat_ngram_size=2)
|
| 63 |
-
|
| 64 |
-
# Декодирование и вывод результата
|
| 65 |
-
generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
|
| 66 |
-
print(generated_text)
|
| 67 |
-
```
|
| 68 |
-
|
| 69 |
-
#### Ответ модели / Model response:
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
> Весна в России – это время обновления природы, когда природа пробуждается >от зимнего сна. Вот стихотворение, отражающее эту красоту:
|
| 73 |
-
>
|
| 74 |
-
> ---
|
| 75 |
-
>
|
| 76 |
-
> **Весна в России**
|
| 77 |
-
>
|
| 78 |
-
> Зимняя тишина утихла,
|
| 79 |
-
> Весна в России пришла.
|
| 80 |
-
> Солнце светит, словно в сказке,
|
| 81 |
-
> В небесах – птицы в полете.
|
| 82 |
-
>
|
| 83 |
-
> Снег пошел, ушел вдаль,
|
| 84 |
-
> И в каждом уголке – весна.
|
| 85 |
-
> Лед промерз, вода в реке –
|
| 86 |
-
> Ветры вьют, и листья поют.
|
| 87 |
-
>
|
| 88 |
-
> Цветы распустились, как будто
|
| 89 |
-
> В честь весны, в честь жизни.
|
| 90 |
-
> Зеленая трава, как полотно,
|
| 91 |
-
> Под ногами – мягкость.
|
| 92 |
-
>
|
| 93 |
-
> Весна в России – это чудо,
|
| 94 |
-
> Счастье, что в сердце живет.
|
| 95 |
-
> И каждый день – праздник,
|
| 96 |
-
> Когда природа в цвету.
|
| 97 |
-
>
|
| 98 |
-
> ---
|
| 99 |
-
>
|
| 100 |
-
> Надеюсь, это стихотворение передало дух и красоту весны в России.
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
## Метрики на ru_arena_general / Metrics on ru_arena_general
|
| 104 |
-
|
| 105 |
-
| Model | Score | 95% CI | Avg Tokens | Std Tokens | LC Score |
|
| 106 |
-
| ---------------------------------------------- | --------- | --------------- | ---------- | ---------- | --------- |
|
| 107 |
-
| suzume-llama-3-8B-multilingual-orpo-borda-half | 90.89 | +1.1 / -1.1 | 2495.38 | 1211.62 | 55.86 |
|
| 108 |
-
| mistral-nemo-instruct-2407 | 50.53 | +2.5 / -2.2 | 403.17 | 321.53 | 50.08 |
|
| 109 |
-
| sfr-iterative-dpo-llama-3-8b-r | 50.06 | +2.1 / -2.1 | 516.74 | 316.84 | 50.01 |
|
| 110 |
-
| gpt-3.5-turbo-0125 | 50.00 | +0.0 / -0.0 | 220.83 | 170.30 | 50.00 |
|
| 111 |
-
| glm-4-9b-chat | 49.75 | +1.9 / -2.3 | 568.81 | 448.76 | 49.96 |
|
| 112 |
-
| c4ai-command-r-v01 | 48.95 | +2.6 / -1.7 | 529.34 | 368.98 | 49.85 |
|
| 113 |
-
| llama-3-instruct-8b-sppo-iter3 | 47.45 | +2.0 / -2.2 | 502.27 | 304.27 | 49.63 |
|
| 114 |
-
| **Vikhrmodels-vikhr-gemma-2b-it** | **45.82** | **+2.4 / -2.0** | **722.83** | **710.71** | **49.40** |
|
| 115 |
-
| suzume-llama-3-8b-multilingual | 45.71 | +2.4 / -1.7 | 641.18 | 858.96 | 49.38 |
|
| 116 |
-
| yandex_gpt_pro | 45.11 | +2.2 / -2.5 | 345.30 | 277.64 | 49.30 |
|
| 117 |
-
| hermes-2-theta-llama-3-8b | 44.07 | +2.0 / -2.2 | 485.99 | 390.85 | 49.15 |
|
| 118 |
-
| gpt-3.5-turbo-1106 | 41.48 | +1.9 / -2.0 | 191.19 | 177.31 | 48.77 |
|
| 119 |
-
| llama-3-smaug-8b | 40.80 | +2.1 / -1.6 | 524.02 | 480.56 | 48.68 |
|
| 120 |
-
| llama-3-8b-saiga-suzume-ties | 39.94 | +2.0 / -1.7 | 763.27 | 699.39 | 48.55 |
|
| 121 |
-
|
| 122 |
-
```
|
| 123 |
-
@article{nikolich2024vikhr,
|
| 124 |
-
title={Vikhr: The Family of Open-Source Instruction-Tuned Large Language Models for Russian},
|
| 125 |
-
author={Aleksandr Nikolich and Konstantin Korolev and Sergey Bratchikov and Nikolay Kompanets and Artem Shelmanov},
|
| 126 |
-
journal={arXiv preprint arXiv:2405.13929},
|
| 127 |
-
year={2024},
|
| 128 |
-
url={https://arxiv.org/pdf/2405.13929}
|
| 129 |
-
}
|
| 130 |
-
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|