Qwen3-14B

MERA Создан 15.03.2026 21:06
0.409
Общий результат

Оценки по задачам лидерборда

Таблица скроллится влево

Задача Результат Метрика
YABLoCo 0.048 / 0.005
EM pass@k
stRuCom 0.291
chrF
RealCode 0.283 / 0.959
pass@k execution_success
UnitTests 0.3
CodeBLEU
ruCodeEval 0.674 / 0.717 / 0.732
pass@k
JavaTestGen 0.22 / 0.463
pass@k compile@1
ruHumanEval 0.775 / 0.8 / 0.805
pass@k
RealCodeJava 0.295 / 0.977
pass@k execution_success
CodeLinterEval 0.389 / 0.538 / 0.591
pass@k
ruCodeReviewer 0.041 / 0.2 / 0.07 / 0.167 / 0.216
chrF BLEU judge@1 judge@5 judge@10
CodeCorrectness 0.812
EM

Информация о сабмите

Версия MERA
v1.0.0
Версия Torch
2.9.1
Версия кодовой базы
7c56310
Версия CUDA
12.8
Precision весов модели
bfloat16
Сид
1234
Батч
1
Версия transformers
4.56.1
Количество GPU и их тип
4 x NVIDIA A100-SXM4-80GB
Архитектура
local-chat-completions

Команда:

MERA

Название ML-модели:

Qwen3-14B

Ссылка на ML-модель:

https://huggingface.co/Qwen/Qwen3-14B

Размер модели

14.0B

Тип модели:

Открытая

SFT

Дополнительные ссылки:

https://huggingface.co/Qwen/Qwen3-14B

Описание архитектуры:

Qwen3-14B is a decoder-only transformer language model from the Qwen3 family with approximately 14 billion parameters. The model is designed for general language understanding and generation tasks, including reasoning, coding, and multilingual applications, and supports long-context inputs.

Описание обучения:

The model follows the standard Qwen3 multi-stage training pipeline consisting of large-scale pretraining followed by post-training. Post-training includes instruction tuning and alignment techniques to improve instruction following, reasoning ability, and response quality.

Данные претрейна:

The model was pretrained on a large multilingual corpus of approximately 36 trillion tokens covering 119 languages. It was further post-trained on instruction-following and reasoning-oriented datasets to improve instruction following and reasoning performance.

Лицензия:

Apache License 2.0

Параметры инференса

Параметры генерации:
codecorrectness - until=["<|im_end|>"];do_sample=false;temperature=0;max_gen_toks=10000; \ncodelintereval - do_sample=true;temperature=0.6;max_gen_toks=10000;until=["<|im_end|>"]; \njavatestgen - do_sample=true;max_gen_toks=10000;temperature=0.2;top_p=0.9;until=["<|im_end|>"]; \nrealcode - do_sample=true;max_gen_toks=10000;temperature=0.7;repetition_penalty=1.05;top_p=0.8;until=["<|im_end|>"]; \nrealcodejava - do_sample=true;max_gen_toks=10000;temperature=0.7;repetition_penalty=1.05;top_p=0.8;until=["<|im_end|>"]; \nrucodeeval_code - do_sample=true;temperature=0.6;max_gen_toks=10000;until=["<|im_end|>"]; \nrucodereviewer - temperature=0;do_sample=false;max_gen_toks=10000;until=["<|im_end|>"]; \nruhumaneval_code - do_sample=true;temperature=0.6;max_gen_toks=10000;until=["<|im_end|>"]; \nstrucom - do_sample=false;max_gen_toks=10000;until=["<|im_end|>"]; \nunittests - do_sample=false;max_gen_toks=10000;until=["<|im_end|>"]; \nyabloco_oracle - max_gen_toks=10000;do_sample=false;until=["<|im_end|>"];

Описание темплейта:
{%- if tools %} {{- '<|im_start|>system \n' }} {%- if messages[0].role == 'system' %} {{- messages[0].content + ' \n \n' }} {%- endif %} {{- "# Tools \n \nYou may call one or more functions to assist with the user query. \n \nYou are provided with function signatures within <tools></tools> XML tags: \n<tools>" }} {%- for tool in tools %} {{- " \n" }} {{- tool | tojson }} {%- endfor %} {{- " \n</tools> \n \nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags: \n<tool_call> \n{"name": <function-name>, "arguments": <args-json-object>} \n</tool_call><|im_end|> \n" }} {%- else %} {%- if messages[0].role == 'system' %} {{- '<|im_start|>system \n' + messages[0].content + '<|im_end|> \n' }} {%- endif %} {%- endif %} {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} {%- for message in messages[::-1] %} {%- set index = (messages|length - 1) - loop.index0 %} {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %} {%- set ns.multi_step_tool = false %} {%- set ns.last_query_index = index %} {%- endif %} {%- endfor %} {%- for message in messages %} {%- if message.content is string %} {%- set content = message.content %} {%- else %} {%- set content = '' %} {%- endif %} {%- if (message.role == "user") or (message.role == "system" and not loop.first) %} {{- '<|im_start|>' + message.role + ' \n' + content + '<|im_end|>' + ' \n' }} {%- elif message.role == "assistant" %} {%- set reasoning_content = '' %} {%- if message.reasoning_content is string %} {%- set reasoning_content = message.reasoning_content %} {%- else %} {%- if '</think>' in content %} {%- set reasoning_content = content.split('</think>')[0].rstrip(' \n').split('<think>')[-1].lstrip(' \n') %} {%- set content = content.split('</think>')[-1].lstrip(' \n') %} {%- endif %} {%- endif %} {%- if loop.index0 > ns.last_query_index %} {%- if loop.last or (not loop.last and reasoning_content) %} {{- '<|im_start|>' + message.role + ' \n<think> \n' + reasoning_content.strip(' \n') + ' \n</think> \n \n' + content.lstrip(' \n') }} {%- else %} {{- '<|im_start|>' + message.role + ' \n' + content }} {%- endif %} {%- else %} {{- '<|im_start|>' + message.role + ' \n' + content }} {%- endif %} {%- if message.tool_calls %} {%- for tool_call in message.tool_calls %} {%- if (loop.first and content) or (not loop.first) %} {{- ' \n' }} {%- endif %} {%- if tool_call.function %} {%- set tool_call = tool_call.function %} {%- endif %} {{- '<tool_call> \n{"name": "' }} {{- tool_call.name }} {{- '", "arguments": ' }} {%- if tool_call.arguments is string %} {{- tool_call.arguments }} {%- else %} {{- tool_call.arguments | tojson }} {%- endif %} {{- '} \n</tool_call>' }} {%- endfor %} {%- endif %} {{- '<|im_end|> \n' }} {%- elif message.role == "tool" %} {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %} {{- '<|im_start|>user' }} {%- endif %} {{- ' \n<tool_response> \n' }} {{- content }} {{- ' \n</tool_response>' }} {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} {{- '<|im_end|> \n' }} {%- endif %} {%- endif %} {%- endfor %} {%- if add_generation_prompt %} {{- '<|im_start|>assistant \n' }} {%- if enable_thinking is defined and enable_thinking is false %} {{- '<think> \n \n</think> \n \n' }} {%- endif %} {%- endif %}