diff --git a/data/docker/mysql-init/init.sql b/data/docker/mysql-init/init.sql deleted file mode 100644 index be9e12884440dc4eafea4520a156c6dc35a5b86e..0000000000000000000000000000000000000000 --- a/data/docker/mysql-init/init.sql +++ /dev/null @@ -1,627 +0,0 @@ --- phpMyAdmin SQL Dump --- version 5.2.1 --- https://www.phpmyadmin.net/ --- --- 主机： 10.128.253.115 --- 生成日期： 2025-03-20 12:31:46 --- 服务器版本： 8.3.0 --- PHP 版本： 8.2.18 - -SET SQL_MODE = "NO_AUTO_VALUE_ON_ZERO"; -START TRANSACTION; -SET time_zone = "+00:00"; - - -/*!40101 SET @OLD_CHARACTER_SET_CLIENT=@@CHARACTER_SET_CLIENT */; -/*!40101 SET @OLD_CHARACTER_SET_RESULTS=@@CHARACTER_SET_RESULTS */; -/*!40101 SET @OLD_COLLATION_CONNECTION=@@COLLATION_CONNECTION */; -/*!40101 SET NAMES utf8mb4 */; - --- --- 数据库： `lmsp_api_server` --- -CREATE DATABASE IF NOT EXISTS `lmsp_api_server` DEFAULT CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci; -USE `lmsp_api_server`; - --- -------------------------------------------------------- - --- --- 表的结构 `apikey_tbl` --- - -CREATE TABLE `apikey_tbl` ( - `id_pk` bigint UNSIGNED NOT NULL, - `user_name` char(255) NOT NULL, - `kclkuserid` char(255) NOT NULL, - `apikey_name` varchar(32) DEFAULT NULL, - `api_key` varchar(64) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `status` char(8) NOT NULL, - `permissions` json DEFAULT NULL, - `create_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP, - `update_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP, - `extra_info` varchar(1024) NOT NULL, - `is_deleted` tinyint(1) NOT NULL DEFAULT '0' -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `api_certs` --- - -CREATE TABLE `api_certs` ( - `id` varchar(64) NOT NULL, - `name` varchar(32) NOT NULL, - `access_domain` text, - `theme` text, - `issuer` text, - `certificate_chain` text NOT NULL, - `private_key` text NOT NULL, - `created_at` datetime NOT NULL, - `updated_at` datetime NOT NULL, - `extra_info` text, - `status` int NOT NULL, - `user_info` text NOT NULL, - `user_name` varchar(32) NOT NULL, - `tenant_name` varchar(32) NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `api_gateways` --- - -CREATE TABLE `api_gateways` ( - `id` varchar(36) NOT NULL, - `name` varchar(32) NOT NULL, - `description` text NOT NULL, - `status` int NOT NULL, - `config` text NOT NULL COMMENT 'JSON format of ApiGatewayConfig', - `created_at` datetime NOT NULL, - `updated_at` datetime NOT NULL, - `user_info` text NOT NULL, - `user_name` varchar(32) NOT NULL, - `tenant_name` varchar(32) NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `dataset_info` --- - -CREATE TABLE `dataset_info` ( - `id_pk` bigint NOT NULL, - `dataset_id` varchar(64) NOT NULL, - `version_id` varchar(64) NOT NULL, - `name` varchar(64) NOT NULL, - `version` varchar(32) NOT NULL, - `history_version` varchar(32) NOT NULL, - `type` int NOT NULL, - `format` int NOT NULL, - `status` int NOT NULL, - `amount` int NOT NULL, - `path` varchar(128) NOT NULL, - `created_at` datetime NOT NULL, - `modified_at` datetime NOT NULL, - `description` text NOT NULL, - `source` int NOT NULL, - `shared` tinyint(1) NOT NULL, - `user_info` text NOT NULL, - `username` varchar(32) NOT NULL, - `tenant_name` varchar(32) NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `infer_request_stats` --- - -CREATE TABLE `infer_request_stats` ( - `id_pk` bigint NOT NULL, - `infer_task_id` varchar(64) NOT NULL, - `req_id` bigint NOT NULL, - `req_status` tinyint(1) NOT NULL COMMENT '0: success, 1: failed', - `input_tokens` int NOT NULL, - `output_tokens` int NOT NULL, - `req_begin_time` datetime NOT NULL COMMENT 'in UTC', - `time_to_first_token` int NOT NULL COMMENT 'in miliseconds', - `time_to_last_token` int NOT NULL COMMENT 'in miliseconds', - `user_name` varchar(32) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL COMMENT 'unique user name', - `tenant_name` varchar(32) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `apikey_name` varchar(64) NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `infer_task_plans` --- - -CREATE TABLE `infer_task_plans` ( - `id_pk` bigint NOT NULL, - `client_id` varchar(64) NOT NULL, - `name` varchar(64) NOT NULL, - `description` text NOT NULL, - `model_id` varchar(64) NOT NULL, - `model_source` varchar(32) NOT NULL, - `model_placement` text NOT NULL COMMENT 'json format of ModelPlacement', - `model_runtime_info` text CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL COMMENT 'json format of ModelRuntimeInfo', - `infer_task_id` varchar(64) NOT NULL, - `generated_at` varchar(32) NOT NULL, - `infer_engine_info` text NOT NULL COMMENT 'json format of HostInfo', - `rank` int NOT NULL, - `infer_engine_list` text NOT NULL COMMENT 'json format of HostInfo list', - `infer_server_endpoint` text NOT NULL, - `infer_engine_type` int NOT NULL, - `task_status` int NOT NULL, - `user_info` text NOT NULL COMMENT 'json format of UserInfo', - `user_name` varchar(32) NOT NULL COMMENT 'unique user name', - `tenant_name` varchar(32) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `hostname` varchar(256) NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `infer_templates` --- - -CREATE TABLE `infer_templates` ( - `id` varchar(36) NOT NULL, - `name` varchar(32) NOT NULL, - `description` text NOT NULL, - `gpu_name` varchar(128) NOT NULL, - `gpu_count` int NOT NULL, - `cpu_count` int NOT NULL, - `memory_size` int NOT NULL COMMENT 'in MiB', - `model_id` varchar(64) NOT NULL, - `infer_precision` varchar(8) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `kilo_input_token_price` float NOT NULL, - `kilo_output_token_price` float NOT NULL, - `status` int NOT NULL, - `created_at` datetime NOT NULL, - `updated_at` datetime NOT NULL, - `extra_info` text NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `instance_details` --- - -CREATE TABLE `instance_details` ( - `id_pk` bigint NOT NULL, - `client_id` varchar(64) NOT NULL, - `instance_id` varchar(64) NOT NULL, - `name` varchar(64) NOT NULL, - `description` text NOT NULL, - `ip_address` varchar(64) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `instance_status` int NOT NULL, - `compute_resource` text CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL COMMENT 'json format of ComputeResource', - `charge_type` int NOT NULL, - `image_name` varchar(256) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `shared_storage_list` text CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL COMMENT 'json format of SharedStorage list', - `ssh_pub_key` text NOT NULL, - `enable_jupyterlab` tinyint(1) NOT NULL, - `expired_at` varchar(32) NOT NULL COMMENT 'RFC3339 format', - `created_at` varchar(32) NOT NULL COMMENT 'RFC3339 format', - `instance_engine_info` text NOT NULL COMMENT 'json format of HostInfo', - `ssh_port` int NOT NULL, - `user_info` text NOT NULL COMMENT 'json format of UserInfo', - `user_name` varchar(32) NOT NULL COMMENT 'unique user name', - `tenant_name` varchar(32) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `hostname` varchar(256) NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `model_display_info` --- - -CREATE TABLE `model_display_info` ( - `id_pk` bigint NOT NULL, - `name` varchar(64) NOT NULL, - `create_at` datetime NOT NULL, - `source` varchar(32) NOT NULL, - `tags` varchar(64) NOT NULL, - `gpus` varchar(64) NOT NULL, - `summary` text NOT NULL, - `description` longtext NOT NULL, - `username` varchar(32) NOT NULL, - `tenant_name` varchar(32) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `model_id` varchar(64) NOT NULL, - `base_model_id` varchar(64) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `task_name` varchar(64) NOT NULL, - `registered` tinyint(1) NOT NULL, - `shared` tinyint(1) NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- --- 转存表中的数据 `model_display_info` --- - -INSERT INTO `model_display_info` (`id_pk`, `name`, `create_at`, `source`, `tags`, `gpus`, `summary`, `description`, `username`, `tenant_name`, `model_id`, `base_model_id`, `task_name`, `registered`, `shared`) VALUES -(1, 'Qwen2.5-7B-Instruct', '2024-09-16 00:00:00', 'public_models', '文本生成', 'Nvidia,Enflame,Denglin', 'Qwen是Qwen团队研发的大语言模型和大型多模态模型系列，Qwen2.5-7B-Instruct是对外开源的7B规模的模型，模型支持 131,072 tokens上下文。', 'Qwen2.5是Qwen大型语言模型系列的最新成果。对于Qwen2.5，发布了从0.5到720亿参数不等的基础语言模型及指令调优语言模型。Qwen2.5相比Qwen2带来了以下改进：\n\n- 显著增加的知识量，在编程与数学领域的能力有了极大提升，这得益于我们在这些领域的专业专家模型。\n- 在遵循指令、生成长文本（超过8K个token）、理解结构化数据（如表格）及生成特别是JSON格式的结构化输出方面有显著提升。对系统提示的多样性更具韧性，增强了聊天机器人的角色扮演实现和条件设定功能。\n- 支持长上下文，最多可达128K个token，并能生成最多8K个token的文本。\n- 支持超过29种语言的多语言能力，包括中文、英语、法语、西班牙语、葡萄牙语、德语、意大利语、俄语、日语、韩语、越南语、泰语、阿拉伯语等。\n\nQwen2.5-7B-Instruct的特点如下：\n\n- 类型：因果语言模型\n- 训练阶段：预训练与后训练调优\n- 架构：采用RoPE、SwiGLU、RMSNorm、注意力QKV偏置及词嵌入绑定的transformers架构\n- 参数数量：7.61B\n- 非嵌入参数数量：6.53B\n- 层数：28\n- 注意力头数（GQA）：28个Q，4个KV\n- 上下文长度：完整131,072个tokens，生成上限8192个tokens', '', '', '62676e8f-9b34-4f38-aad5-33412d9ae006', '', '', 0, 0), -(2, 'Qwen2.5-14B-Instruct', '2024-09-16 00:00:00', 'public_models', '文本生成', 'Nvidia,Enflame,Denglin', 'Qwen是Qwen团队研发的大语言模型和大型多模态模型系列，Qwen2.5-14B-Instruct是对外开源的14B规模的模型，模型支持 131,072 tokens上下文。', 'Qwen2.5是Qwen大型语言模型系列的最新成果。对于Qwen2.5，发布了从0.5到720亿参数不等的基础语言模型及指令调优语言模型。Qwen2.5相比Qwen2带来了以下改进：\n\n- 显著增加的知识量，在编程与数学领域的能力有了极大提升，这得益于我们在这些领域的专业专家模型。\n- 在遵循指令、生成长文本（超过8K个token）、理解结构化数据（如表格）及生成特别是JSON格式的结构化输出方面有显著提升。对系统提示的多样性更具韧性，增强了聊天机器人的角色扮演实现和条件设定功能。\n- 支持长上下文，最多可达128K个token，并能生成最多8K个token的文本。\n- 支持超过29种语言的多语言能力，包括中文、英语、法语、西班牙语、葡萄牙语、德语、意大利语、俄语、日语、韩语、越南语、泰语、阿拉伯语等。\n\nQwen2.5-14B-Instruct的特点如下：\n\n- 类型：因果语言模型\n- 训练阶段：预训练与后训练调优\n- 架构：采用RoPE、SwiGLU、RMSNorm、注意力QKV偏置及词嵌入绑定的transformers架构\n- 参数数量：14.7B\n- 非嵌入参数数量：13.1B\n- 层数：48\n- 注意力头数（GQA）：40个Q，8个KV\n- 上下文长度：完整131,072个tokens，生成上限8192个tokens', '', '', '4396460c-4459-4fe3-84ea-8de3b8be4d1e', '', '', 0, 0), -(3, 'Qwen2.5-32B-Instruct', '2024-09-17 00:00:00', 'public_models', '文本生成', 'Nvidia,Enflame,Denglin', 'Qwen是Qwen团队研发的大语言模型和大型多模态模型系列，Qwen2.5-32B-Instruct是对外开源的32B规模的模型，模型支持 131,072 tokens上下文。', 'Qwen2.5是Qwen大型语言模型系列的最新成果。对于Qwen2.5，发布了从0.5到720亿参数不等的基础语言模型及指令调优语言模型。Qwen2.5相比Qwen2带来了以下改进：\n\n- 显著增加的知识量，在编程与数学领域的能力有了极大提升，这得益于我们在这些领域的专业专家模型。\n- 在遵循指令、生成长文本（超过8K个token）、理解结构化数据（如表格）及生成特别是JSON格式的结构化输出方面有显著提升。对系统提示的多样性更具韧性，增强了聊天机器人的角色扮演实现和条件设定功能。\n- 支持长上下文，最多可达128K个token，并能生成最多8K个token的文本。\n- 支持超过29种语言的多语言能力，包括中文、英语、法语、西班牙语、葡萄牙语、德语、意大利语、俄语、日语、韩语、越南语、泰语、阿拉伯语等。\n\nQwen2.5-32B-Instruct的特点如下：\n\n- 类型：因果语言模型\n- 训练阶段：预训练与后训练调优\n- 架构：采用RoPE、SwiGLU、RMSNorm、注意力QKV偏置及词嵌入绑定的transformers架构\n- 参数数量：32.5B\n- 非嵌入参数数量：31.0B\n- 层数：64\n- 注意力头数（GQA）：40个Q，8个KV\n- 上下文长度：完整131,072个tokens，生成上限8192个tokens', '', '', 'c6625af2-e82a-4602-8b0b-e22ed1c0e406', '', '', 0, 0), -(4, 'Qwen2.5-Coder-7B-Instruct', '2024-09-18 00:00:00', 'public_models', '文本生成', 'Nvidia,Enflame,Denglin', 'Qwen2.5-Coder 是最新的代码专用 Qwen 大型语言模型系列，Qwen2.5-Coder-7B-Instruct是对外开源的7B规模的模型，模型支持 131,072 tokens上下文。', 'Qwen2.5-Coder 是最新的代码专用 Qwen 大型语言模型系列（以前称为 CodeQwen）。目前，Qwen2.5-Coder 已经覆盖了六种主流的模型大小，分别是 0.5、1.5、3、7、14 和 320 亿参数，以满足不同开发者的需求。Qwen2.5-Coder 相对于 CodeQwen1.5 带来了以下改进：\n\n- 在代码生成、代码推理和代码修复方面有显著提升。基于强大的 Qwen2.5，我们将训练tokens扩展到了 5.5 万亿，包括源代码、文本-代码对接、合成数据等。\n- 为诸如代码代理等实际应用提供了更全面的基础。不仅增强了编码能力，还保持了在数学和通用能力方面的优势。\n- 长上下文支持高达 128K tokens。\n\n此仓库包含指令微调后的 7B Qwen2.5-Coder 模型，具有以下特点：\n\n- 类型：因果语言模型\n- 训练阶段：预训练 & 后训练\n- 架构：带有 RoPE、SwiGLU、RMSNorm 和注意力 QKV 偏置的 transformers\n- 参数数量：7.61B\n- 非嵌入参数数量：6.53B\n- 层数：28\n- 注意力头数（GQA）：Q 为 28 个，KV 为 4 个\n- 上下文长度：完整的 131,072 令牌', '', '', '50f4990f-66f6-4201-84de-322a7bba35a7', '', '', 0, 0), -(5, 'Qwen2.5-Coder-32B-Instruct', '2024-11-06 00:00:00', 'public_models', '文本生成', 'Nvidia,Enflame,Denglin', 'Qwen2.5-Coder 是最新的代码专用 Qwen 大型语言模型系列，Qwen2.5-Coder-32B-Instruct是对外开源的32B规模的模型，模型支持 131,072 tokens上下文。', 'Qwen2.5-Coder 是最新的代码专用 Qwen 大型语言模型系列（以前称为 CodeQwen）。目前，Qwen2.5-Coder 已经覆盖了六种主流的模型大小，分别是 0.5、1.5、3、7、14 和 320 亿参数，以满足不同开发者的需求。Qwen2.5-Coder 相对于 CodeQwen1.5 带来了以下改进：\n\n- 在代码生成、代码推理和代码修复方面有显著提升。基于强大的 Qwen2.5，我们将训练tokens扩展到了 5.5 万亿，包括源代码、文本-代码对接、合成数据等。\n- 为诸如代码代理等实际应用提供了更全面的基础。不仅增强了编码能力，还保持了在数学和通用能力方面的优势。\n- 长上下文支持高达 128K tokens。\n\n此仓库包含指令微调后的 32B Qwen2.5-Coder 模型，具有以下特点：\n\n- 类型：因果语言模型\n- 训练阶段：预训练 & 后训练\n- 架构：带有 RoPE、SwiGLU、RMSNorm 和注意力 QKV 偏置的 transformers\n- 参数数量：32.5B\n- 非嵌入参数数量：31.0B\n- 层数：64\n- 注意力头数（GQA）：Q 为 40 个，KV 为 8 个\n- 上下文长度：完整的 131,072 令牌', '', '', 'ab2e8e37-50ca-44b4-8a83-e8bad8b32ca4', '', '', 0, 0), -(6, 'DeepSeek-R1-Distill-Qwen-7B', '2025-01-20 00:00:00', 'public_models', '文本生成', 'Nvidia,Enflame,Denglin', 'DeepSeek-R1-Distill 系列模型是基于开源模型并利用 DeepSeek-R1 生成的样本进行微调的，本模型源自 Qwen-2.5 系列，模型支持 131,072 tokens上下文。', '使用由 DeepSeek-R1 生成的推理数据，对研究社区广泛使用的一些密集模型进行了微调。评估结果表明，提炼后的小型密集模型在基准测试中的表现非常出色。基于 Qwen2.5 和 Llama3 系列提炼出 1.5B、7B、8B、14B、32B 和 70B 的模型。', '', '', '5b406c08-771a-4669-ab58-949c0c188c0c', '', '', 0, 0), -(7, 'DeepSeek-R1-Distill-Llama-8B', '2025-01-20 00:00:00', 'public_models', '文本生成', 'Nvidia,Enflame,Denglin', 'DeepSeek-R1-Distill 系列模型是基于开源模型并利用 DeepSeek-R1 生成的样本进行微调的，本模型源自Llama3.1-8B-Base，模型支持 131,072 tokens上下文。', '使用由 DeepSeek-R1 生成的推理数据，对研究社区广泛使用的一些密集模型进行了微调。评估结果表明，提炼后的小型密集模型在基准测试中的表现非常出色。基于 Qwen2.5 和 Llama3 系列提炼出 1.5B、7B、8B、14B、32B 和 70B 的模型。', '', '', '153ae9a0-70b7-4495-9e61-b197d58eaa5d', '', '', 0, 0), -(8, 'DeepSeek-R1-Distill-Qwen-14B', '2025-01-20 00:00:00', 'public_models', '文本生成', 'Nvidia,Enflame,Denglin', 'DeepSeek-R1-Distill 系列模型是基于开源模型并利用 DeepSeek-R1 生成的样本进行微调的，本模型源自Qwen-2.5 系列，模型支持 131,072 tokens上下文。', '使用由 DeepSeek-R1 生成的推理数据，对研究社区广泛使用的一些密集模型进行了微调。评估结果表明，提炼后的小型密集模型在基准测试中的表现非常出色。基于 Qwen2.5 和 Llama3 系列提炼出 1.5B、7B、8B、14B、32B 和 70B 的模型。', '', '', 'a7932136-d636-47dc-8239-38ec40c6c6f1', '', '', 0, 0), -(9, 'DeepSeek-R1-Distill-Qwen-32B', '2025-01-20 00:00:00', 'public_models', '文本生成', 'Nvidia,Enflame,Denglin', 'DeepSeek-R1-Distill 系列模型是基于开源模型并利用 DeepSeek-R1 生成的样本进行微调的，本模型源自Qwen-2.5 系列，模型支持 131,072 tokens上下文。', '使用由 DeepSeek-R1 生成的推理数据，对研究社区广泛使用的一些密集模型进行了微调。评估结果表明，提炼后的小型密集模型在基准测试中的表现非常出色。基于 Qwen2.5 和 Llama3 系列提炼出 1.5B、7B、8B、14B、32B 和 70B 的模型。', '', '', '59d16cf2-304e-4fb2-89aa-546dbfd1884b', '', '', 0, 0); - --- -------------------------------------------------------- - --- --- 表的结构 `model_serve_info` --- - -CREATE TABLE `model_serve_info` ( - `id_pk` int NOT NULL, - `architecture` varchar(16) NOT NULL, - `name` varchar(32) NOT NULL, - `type` varchar(16) NOT NULL, - `file_path` varchar(128) NOT NULL, - `block_count` int NOT NULL, - `context_len` int NOT NULL, - `embedding_len` int NOT NULL, - `ff_len` int NOT NULL, - `attn_head_count` int NOT NULL, - `attn_head_count_kv` int NOT NULL, - `rope_freq_base` int NOT NULL, - `attn_layer_norm_rms_epsilon` float NOT NULL, - `vocabulary_size` int NOT NULL, - `eos_token_id` int NOT NULL, - `padding_token_id` int NOT NULL, - `bos_token_id` int NOT NULL, - `quantization_version` int NOT NULL, - `model_parameters` int NOT NULL, - `model_id` varchar(64) NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- --- 转存表中的数据 `model_serve_info` --- - -INSERT INTO `model_serve_info` (`id_pk`, `architecture`, `name`, `type`, `file_path`, `block_count`, `context_len`, `embedding_len`, `ff_len`, `attn_head_count`, `attn_head_count_kv`, `rope_freq_base`, `attn_layer_norm_rms_epsilon`, `vocabulary_size`, `eos_token_id`, `padding_token_id`, `bos_token_id`, `quantization_version`, `model_parameters`, `model_id`) VALUES -(1, 'qwen', 'Qwen2.5-7B-Instruct', 'safetensors', '/models/Qwen/Qwen2.5-7B-Instruct', 28, 32768, 3584, 18944, 28, 4, 1000000.0, 0.000001, 152064, 151645, 151643, 151643, 2, 7000, '62676e8f-9b34-4f38-aad5-33412d9ae006'), -(2, 'qwen', 'Qwen2.5-14B-Instruct', 'safetensors', '/models/Qwen/Qwen2.5-14B-Instruct', 70, 32768, 5120, 13824, 40, 8, 1000000.0, 0.000001, 152064, 151645, 151643, 151643, 2, 14000, '4396460c-4459-4fe3-84ea-8de3b8be4d1e'), -(3, 'qwen', 'Qwen2.5-32B-Instruct', 'safetensors', '/models/Qwen/Qwen2.5-32B-Instruct', 70, 32768, 5120, 27648, 40, 8, 1000000.0, 0.000001, 152064, 151645, 151643, 151643, 2, 32000, 'c6625af2-e82a-4602-8b0b-e22ed1c0e406'), -(4, 'qwen', 'Qwen2.5-Coder-7B-Instruct', 'safetensors', '/models/Qwen/Qwen2.5-Coder-7B-Instruct', 28, 32768, 3584, 18944, 28, 4, 1000000.0, 0.000001, 152064, 151645, 151643, 151643, 2, 7000, '50f4990f-66f6-4201-84de-322a7bba35a7'), -(5, 'qwen', 'Qwen2.5-Coder-32B-Instruct', 'safetensors', '/models/Qwen/Qwen2.5-Coder-32B-Instruct', 64, 32768, 5120, 27648, 40, 8, 1000000.0, 0.000001, 152064, 151645, 151643, 151643, 2, 32000, 'ab2e8e37-50ca-44b4-8a83-e8bad8b32ca4'), -(6, 'qwen', 'DeepSeek-R1-Distill-Qwen-7B', 'safetensors', '/models/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B', 28, 131072, 3584, 18944, 28, 4, 10000, 0.000001, 152064, 151643, 151643, 151643, 2, 7000, '5b406c08-771a-4669-ab58-949c0c188c0c'), -(7, 'llama', 'DeepSeek-R1-Distill-Llama-8B', 'safetensors', '/models/deepseek-ai/DeepSeek-R1-Distill-Llama-8B', 32, 131072, 4096, 14336, 32, 8, 500000, 0.00001, 128256, 128001, 128001, 128000, 2, 8000, '153ae9a0-70b7-4495-9e61-b197d58eaa5d'), -(8, 'qwen', 'DeepSeek-R1-Distill-Qwen-14B', 'safetensors', '/models/deepseek-ai/DeepSeek-R1-Distill-Qwen-14B', 48, 131072, 5120, 13824, 40, 8, 1000000, 0.00001, 152064, 151643, 151643, 151643, 2, 14000, 'a7932136-d636-47dc-8239-38ec40c6c6f1'), -(9, 'qwen', 'DeepSeek-R1-Distill-Qwen-32B', 'safetensors', '/models/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B', 64, 131072, 5120, 27648, 40, 8, 1000000.0, 0.00001, 152064, 151643, 151643, 151643, 2, 32000, '59d16cf2-304e-4fb2-89aa-546dbfd1884b'); - --- -------------------------------------------------------- - --- --- 表的结构 `server_stats` --- - -CREATE TABLE `server_stats` ( - `id_pk` bigint NOT NULL, - `hostname` varchar(256) NOT NULL, - `cpu_usage` decimal(5,2) NOT NULL, - `ram_usage` decimal(5,2) NOT NULL, - `gpu_usage` decimal(5,2) NOT NULL, - `vram_usage` decimal(5,2) NOT NULL, - `details` text NOT NULL, - `update_at` datetime NOT NULL COMMENT 'in UTC' -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `train_task_details` --- - -CREATE TABLE `train_task_details` ( - `id_pk` bigint NOT NULL, - `train_task_id` varchar(64) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `train_task_detail_list` text CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL COMMENT 'json format of TrainTaskDetail list' -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- -------------------------------------------------------- - --- --- 表的结构 `train_task_plans` --- - -CREATE TABLE `train_task_plans` ( - `id_pk` bigint NOT NULL, - `client_id` varchar(64) NOT NULL, - `name` varchar(64) NOT NULL, - `description` text NOT NULL, - `model_id` varchar(64) NOT NULL, - `model_source` varchar(32) NOT NULL, - `training_config` text CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL COMMENT 'json format of TrainingConfig', - `train_task_id` varchar(64) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `generated_at` varchar(32) NOT NULL, - `train_engine_info` text CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL COMMENT 'json format of HostInfo', - `task_status` int NOT NULL, - `percentage` float NOT NULL, - `compute_resource` text CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `user_info` text NOT NULL COMMENT 'json format of UserInfo', - `user_name` varchar(32) NOT NULL COMMENT 'unique user name', - `tenant_name` varchar(32) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `hostname` varchar(256) NOT NULL -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- --- 转储表的索引 --- - --- --- 表的索引 `apikey_tbl` --- -ALTER TABLE `apikey_tbl` - ADD PRIMARY KEY (`id_pk`), - ADD UNIQUE KEY `unique_api_key` (`api_key`), - ADD KEY `user_name` (`user_name`); - --- --- 表的索引 `api_certs` --- -ALTER TABLE `api_certs` - ADD PRIMARY KEY (`id`), - ADD KEY `idx_user_name` (`user_name`), - ADD KEY `idx_tenant_name` (`tenant_name`); - --- --- 表的索引 `api_gateways` --- -ALTER TABLE `api_gateways` - ADD PRIMARY KEY (`id`), - ADD KEY `idx_user_name` (`user_name`), - ADD KEY `idx_tenant_name` (`tenant_name`); - --- --- 表的索引 `dataset_info` --- -ALTER TABLE `dataset_info` - ADD PRIMARY KEY (`id_pk`); - --- --- 表的索引 `infer_request_stats` --- -ALTER TABLE `infer_request_stats` - ADD PRIMARY KEY (`id_pk`), - ADD KEY `idx_user_name` (`user_name`), - ADD KEY `idx_apikey_name` (`apikey_name`), - ADD KEY `idx_tenant_name` (`tenant_name`) USING BTREE, - ADD KEY `idx_req_begin_time` (`req_begin_time`); - --- --- 表的索引 `infer_task_plans` --- -ALTER TABLE `infer_task_plans` - ADD PRIMARY KEY (`id_pk`), - ADD UNIQUE KEY `idx_infer_task_id` (`infer_task_id`) USING BTREE, - ADD KEY `idx_user_name` (`user_name`), - ADD KEY `idx_hostname` (`hostname`), - ADD KEY `idx_tenant_name` (`tenant_name`) USING BTREE; - --- --- 表的索引 `infer_templates` --- -ALTER TABLE `infer_templates` - ADD PRIMARY KEY (`id`), - ADD KEY `idx_model_id` (`model_id`); - --- --- 表的索引 `instance_details` --- -ALTER TABLE `instance_details` - ADD PRIMARY KEY (`id_pk`), - ADD UNIQUE KEY `idx_instance_id` (`instance_id`) USING BTREE, - ADD KEY `idx_user_name` (`user_name`), - ADD KEY `idx_hostname` (`hostname`) USING BTREE, - ADD KEY `idx_tenant_name` (`tenant_name`) USING BTREE; - --- --- 表的索引 `model_display_info` --- -ALTER TABLE `model_display_info` - ADD PRIMARY KEY (`id_pk`); - --- --- 表的索引 `model_serve_info` --- -ALTER TABLE `model_serve_info` - ADD PRIMARY KEY (`id_pk`); - --- --- 表的索引 `server_stats` --- -ALTER TABLE `server_stats` - ADD PRIMARY KEY (`id_pk`), - ADD KEY `idx_update_at` (`update_at`); - --- --- 表的索引 `train_task_details` --- -ALTER TABLE `train_task_details` - ADD PRIMARY KEY (`id_pk`), - ADD KEY `idx_train_task_id` (`train_task_id`) USING BTREE; - --- --- 表的索引 `train_task_plans` --- -ALTER TABLE `train_task_plans` - ADD PRIMARY KEY (`id_pk`), - ADD UNIQUE KEY `idx_train_task_id` (`train_task_id`) USING BTREE, - ADD KEY `idx_user_name` (`user_name`), - ADD KEY `idx_hostname` (`hostname`) USING BTREE, - ADD KEY `idx_tenant_name` (`tenant_name`) USING BTREE; - --- --- 在导出的表使用AUTO_INCREMENT --- - --- --- 使用表AUTO_INCREMENT `apikey_tbl` --- -ALTER TABLE `apikey_tbl` - MODIFY `id_pk` bigint UNSIGNED NOT NULL AUTO_INCREMENT; - --- --- 使用表AUTO_INCREMENT `dataset_info` --- -ALTER TABLE `dataset_info` - MODIFY `id_pk` bigint NOT NULL AUTO_INCREMENT; - --- --- 使用表AUTO_INCREMENT `infer_request_stats` --- -ALTER TABLE `infer_request_stats` - MODIFY `id_pk` bigint NOT NULL AUTO_INCREMENT; - --- --- 使用表AUTO_INCREMENT `infer_task_plans` --- -ALTER TABLE `infer_task_plans` - MODIFY `id_pk` bigint NOT NULL AUTO_INCREMENT; - --- --- 使用表AUTO_INCREMENT `instance_details` --- -ALTER TABLE `instance_details` - MODIFY `id_pk` bigint NOT NULL AUTO_INCREMENT; - --- --- 使用表AUTO_INCREMENT `model_display_info` --- -ALTER TABLE `model_display_info` - MODIFY `id_pk` bigint NOT NULL AUTO_INCREMENT, AUTO_INCREMENT=30; - --- --- 使用表AUTO_INCREMENT `model_serve_info` --- -ALTER TABLE `model_serve_info` - MODIFY `id_pk` int NOT NULL AUTO_INCREMENT, AUTO_INCREMENT=29; - --- --- 使用表AUTO_INCREMENT `server_stats` --- -ALTER TABLE `server_stats` - MODIFY `id_pk` bigint NOT NULL AUTO_INCREMENT; - --- --- 使用表AUTO_INCREMENT `train_task_details` --- -ALTER TABLE `train_task_details` - MODIFY `id_pk` bigint NOT NULL AUTO_INCREMENT; - --- --- 使用表AUTO_INCREMENT `train_task_plans` --- -ALTER TABLE `train_task_plans` - MODIFY `id_pk` bigint NOT NULL AUTO_INCREMENT; -COMMIT; - -/*!40101 SET CHARACTER_SET_CLIENT=@OLD_CHARACTER_SET_CLIENT */; -/*!40101 SET CHARACTER_SET_RESULTS=@OLD_CHARACTER_SET_RESULTS */; -/*!40101 SET COLLATION_CONNECTION=@OLD_COLLATION_CONNECTION */; - - - - - --- phpMyAdmin SQL Dump --- version 5.2.1 --- https://www.phpmyadmin.net/ --- --- 主机： 10.128.253.115 --- 生成日期： 2025-03-20 12:32:04 --- 服务器版本： 8.3.0 --- PHP 版本： 8.2.18 - -SET SQL_MODE = "NO_AUTO_VALUE_ON_ZERO"; -START TRANSACTION; -SET time_zone = "+00:00"; - - -/*!40101 SET @OLD_CHARACTER_SET_CLIENT=@@CHARACTER_SET_CLIENT */; -/*!40101 SET @OLD_CHARACTER_SET_RESULTS=@@CHARACTER_SET_RESULTS */; -/*!40101 SET @OLD_COLLATION_CONNECTION=@@COLLATION_CONNECTION */; -/*!40101 SET NAMES utf8mb4 */; - --- --- 数据库： `lmsp_inference_server` --- -CREATE DATABASE IF NOT EXISTS `lmsp_inference_server` DEFAULT CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci; -USE `lmsp_inference_server`; - --- -------------------------------------------------------- - --- --- 表的结构 `model_tbl` --- - -CREATE TABLE `model_tbl` ( - `id_pk` int NOT NULL, - `model_id` varchar(128) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL, - `owner` char(32) NOT NULL COMMENT 'The owner''s user name.', - `model_backend` varchar(64) NOT NULL, - `model_backend_url` varchar(64) NOT NULL, - `template_content` text CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NOT NULL COMMENT 'JSON format', - `create_time` datetime NOT NULL, - `update_time` datetime NOT NULL, - `extra_info` varchar(1024) NOT NULL, - `is_deleted` tinyint(1) NOT NULL DEFAULT '0' -) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci; - --- --- 转存表中的数据 `model_tbl` --- - -INSERT INTO `model_tbl` (`id_pk`, `model_id`, `owner`, `model_backend`, `model_backend_url`, `template_content`, `create_time`, `update_time`, `extra_info`, `is_deleted`) VALUES -(1, 'lmsp:Qwen2.5-7B-Instruct', 'default-user', 'lmsp', '', '{\n \"architecture\": \"qwen2.5\",\n \"template_system\": \"<|im_start|>system\\n{{ .System }}<|im_end|>\\n\",\n \"template_user\": \"<|im_start|>user\\n{{ .Prompt }}<|im_end|>\\n\",\n \"template_assistant\": \"<|im_start|>assistant\\n{{ .Answer }}<|im_end|>\\n\",\n \"template_system_key\": \"{{ .System }}\",\n \"template_user_key\": \"{{ .Prompt }}\",\n \"template_assistant_key\": \"{{ .Answer }}\",\n \"template_assistant_head\": \"<|im_start|>assistant\\n\",\n \"system_prompt\": \"You are Qwen,created by Alibaba Cloud. You are a helpful assistant.\",\n \"knowledge_base_prompt\": \"使用以下信息：\\n\\n{{ .Knowledge }}\\n\\n简洁地用中文回答：{{ .Prompt }}\",\n \"template_knowledge_key\": \"{{ .Knowledge }}\",\n \"knowledge_not_found_response\": \"在知识库中没有找到相关信息。请更换问题后重试。\",\n \"stop_words\": [\n \"<|im_start|>\",\n \"<|im_end|>\",\n \"\"\n ],\n \"bad_words\": null\n}', '2025-03-14 14:32:00', '2025-03-14 14:32:00', '', 0), -(2, 'lmsp:Qwen2.5-14B-Instruct', 'default-user', 'lmsp', '', '{\n \"architecture\": \"qwen2.5\",\n \"template_system\": \"<|im_start|>system\\n{{ .System }}<|im_end|>\\n\",\n \"template_user\": \"<|im_start|>user\\n{{ .Prompt }}<|im_end|>\\n\",\n \"template_assistant\": \"<|im_start|>assistant\\n{{ .Answer }}<|im_end|>\\n\",\n \"template_system_key\": \"{{ .System }}\",\n \"template_user_key\": \"{{ .Prompt }}\",\n \"template_assistant_key\": \"{{ .Answer }}\",\n \"template_assistant_head\": \"<|im_start|>assistant\\n\",\n \"system_prompt\": \"You are Qwen,created by Alibaba Cloud. You are a helpful assistant.\",\n \"knowledge_base_prompt\": \"使用以下信息：\\n\\n{{ .Knowledge }}\\n\\n简洁地用中文回答：{{ .Prompt }}\",\n \"template_knowledge_key\": \"{{ .Knowledge }}\",\n \"knowledge_not_found_response\": \"在知识库中没有找到相关信息。请更换问题后重试。\",\n \"stop_words\": [\n \"<|im_start|>\",\n \"<|im_end|>\",\n \"\"\n ],\n \"bad_words\": null\n}', '2025-03-14 14:32:00', '2025-03-14 14:32:00', '', 0), -(3, 'lmsp:Qwen2.5-32B-Instruct', 'default-user', 'lmsp', '', '{\n \"architecture\": \"qwen2.5\",\n \"template_system\": \"<|im_start|>system\\n{{ .System }}<|im_end|>\\n\",\n \"template_user\": \"<|im_start|>user\\n{{ .Prompt }}<|im_end|>\\n\",\n \"template_assistant\": \"<|im_start|>assistant\\n{{ .Answer }}<|im_end|>\\n\",\n \"template_system_key\": \"{{ .System }}\",\n \"template_user_key\": \"{{ .Prompt }}\",\n \"template_assistant_key\": \"{{ .Answer }}\",\n \"template_assistant_head\": \"<|im_start|>assistant\\n\",\n \"system_prompt\": \"You are Qwen,created by Alibaba Cloud. You are a helpful assistant.\",\n \"knowledge_base_prompt\": \"使用以下信息：\\n\\n{{ .Knowledge }}\\n\\n简洁地用中文回答：{{ .Prompt }}\",\n \"template_knowledge_key\": \"{{ .Knowledge }}\",\n \"knowledge_not_found_response\": \"在知识库中没有找到相关信息。请更换问题后重试。\",\n \"stop_words\": [\n \"<|im_start|>\",\n \"<|im_end|>\",\n \"\"\n ],\n \"bad_words\": null\n}', '2025-03-14 14:32:00', '2025-03-14 14:32:00', '', 0), -(4, 'lmsp:Qwen2.5-Coder-7B-Instruct', 'default-user', 'lmsp', '', '{\n \"architecture\": \"qwen2.5\",\n \"template_system\": \"<|im_start|>system\\n{{ .System }}<|im_end|>\\n\",\n \"template_user\": \"<|im_start|>user\\n{{ .Prompt }}<|im_end|>\\n\",\n \"template_assistant\": \"<|im_start|>assistant\\n{{ .Answer }}<|im_end|>\\n\",\n \"template_system_key\": \"{{ .System }}\",\n \"template_user_key\": \"{{ .Prompt }}\",\n \"template_assistant_key\": \"{{ .Answer }}\",\n \"template_assistant_head\": \"<|im_start|>assistant\\n\",\n \"system_prompt\": \"You are Qwen,created by Alibaba Cloud. You are a helpful assistant.\",\n \"knowledge_base_prompt\": \"使用以下信息：\\n\\n{{ .Knowledge }}\\n\\n简洁地用中文回答：{{ .Prompt }}\",\n \"template_knowledge_key\": \"{{ .Knowledge }}\",\n \"knowledge_not_found_response\": \"在知识库中没有找到相关信息。请更换问题后重试。\",\n \"stop_words\": [\n \"<|im_start|>\",\n \"<|im_end|>\",\n \"\"\n ],\n \"bad_words\": null\n}', '2025-03-14 14:32:00', '2025-03-14 14:32:00', '', 0), -(5, 'lmsp:Qwen2.5-Coder-32B-Instruct', 'default-user', 'lmsp', '', '{\n \"architecture\": \"qwen2.5\",\n \"template_system\": \"<|im_start|>system\\n{{ .System }}<|im_end|>\\n\",\n \"template_user\": \"<|im_start|>user\\n{{ .Prompt }}<|im_end|>\\n\",\n \"template_assistant\": \"<|im_start|>assistant\\n{{ .Answer }}<|im_end|>\\n\",\n \"template_system_key\": \"{{ .System }}\",\n \"template_user_key\": \"{{ .Prompt }}\",\n \"template_assistant_key\": \"{{ .Answer }}\",\n \"template_assistant_head\": \"<|im_start|>assistant\\n\",\n \"system_prompt\": \"You are Qwen,created by Alibaba Cloud. You are a helpful assistant.\",\n \"knowledge_base_prompt\": \"使用以下信息：\\n\\n{{ .Knowledge }}\\n\\n简洁地用中文回答：{{ .Prompt }}\",\n \"template_knowledge_key\": \"{{ .Knowledge }}\",\n \"knowledge_not_found_response\": \"在知识库中没有找到相关信息。请更换问题后重试。\",\n \"stop_words\": [\n \"<|im_start|>\",\n \"<|im_end|>\",\n \"\"\n ],\n \"bad_words\": null\n}', '2025-03-14 14:32:00', '2025-03-14 14:32:00', '', 0), -(6, 'lmsp:DeepSeek-R1-Distill-Qwen-7B', 'default-user', 'lmsp', '', '{\r\n \"architecture\": \"deepseek-r1\",\r\n \"template_system\": \"<｜begin▁of▁sentence｜>{{ .System }}\",\r\n \"template_user\": \"<｜User｜>{{ .Prompt }}\",\r\n \"template_assistant\": \"<｜Assistant｜>{{ .Answer }}<｜end▁of▁sentence｜>\",\r\n \"template_system_key\": \"{{ .System }}\",\r\n \"template_user_key\": \"{{ .Prompt }}\",\r\n \"template_assistant_key\": \"{{ .Answer }}\",\r\n \"template_assistant_head\": \"<｜Assistant｜>\\n\",\r\n \"system_prompt\": \"\",\r\n \"knowledge_base_prompt\": \"使用以下信息：\\n\\n{{ .Knowledge }}\\n\\n简洁地用中文回答：{{ .Prompt }}\",\r\n \"template_knowledge_key\": \"{{ .Knowledge }}\",\r\n \"knowledge_not_found_response\": \"在知识库中没有找到相关信息。请更换问题后重试。\",\r\n \"stop_words\": [\r\n \"<｜begin▁of▁sentence｜>\",\r\n \"<｜end▁of▁sentence｜>\",\r\n \"\"\r\n ],\r\n \"bad_words\": null\r\n}', '2025-03-17 11:09:00', '2025-03-17 11:09:00', '', 0), -(7, 'lmsp:DeepSeek-R1-Distill-Llama-8B', 'default-user', 'lmsp', '', '{\r\n \"architecture\": \"deepseek-r1\",\r\n \"template_system\": \"<｜begin▁of▁sentence｜>{{ .System }}\",\r\n \"template_user\": \"<｜User｜>{{ .Prompt }}\",\r\n \"template_assistant\": \"<｜Assistant｜>{{ .Answer }}<｜end▁of▁sentence｜>\",\r\n \"template_system_key\": \"{{ .System }}\",\r\n \"template_user_key\": \"{{ .Prompt }}\",\r\n \"template_assistant_key\": \"{{ .Answer }}\",\r\n \"template_assistant_head\": \"<｜Assistant｜>\\n\",\r\n \"system_prompt\": \"\",\r\n \"knowledge_base_prompt\": \"使用以下信息：\\n\\n{{ .Knowledge }}\\n\\n简洁地用中文回答：{{ .Prompt }}\",\r\n \"template_knowledge_key\": \"{{ .Knowledge }}\",\r\n \"knowledge_not_found_response\": \"在知识库中没有找到相关信息。请更换问题后重试。\",\r\n \"stop_words\": [\r\n \"<｜begin▁of▁sentence｜>\",\r\n \"<｜end▁of▁sentence｜>\",\r\n \"\"\r\n ],\r\n \"bad_words\": null\r\n}', '2025-03-17 11:09:00', '2025-03-17 11:09:00', '', 0), -(8, 'lmsp:DeepSeek-R1-Distill-Qwen-14B', 'default-user', 'lmsp', '', '{\n \"architecture\": \"deepseek-r1\",\n \"template_system\": \"<｜begin▁of▁sentence｜>{{ .System }}\",\n \"template_user\": \"<｜User｜>{{ .Prompt }}\",\n \"template_assistant\": \"<｜Assistant｜>{{ .Answer }}<｜end▁of▁sentence｜>\",\n \"template_system_key\": \"{{ .System }}\",\n \"template_user_key\": \"{{ .Prompt }}\",\n \"template_assistant_key\": \"{{ .Answer }}\",\n \"template_assistant_head\": \"<｜Assistant｜>\\n\",\n \"system_prompt\": \"\",\n \"knowledge_base_prompt\": \"使用以下信息：\\n\\n{{ .Knowledge }}\\n\\n简洁地用中文回答：{{ .Prompt }}\",\n \"template_knowledge_key\": \"{{ .Knowledge }}\",\n \"knowledge_not_found_response\": \"在知识库中没有找到相关信息。请更换问题后重试。\",\n \"stop_words\": [\n \"<｜begin▁of▁sentence｜>\",\n \"<｜end▁of▁sentence｜>\",\n \"\"\n ],\n \"bad_words\": null\n}', '2025-03-17 11:09:00', '2025-03-17 11:09:00', '', 0), -(9, 'lmsp:DeepSeek-R1-Distill-Qwen-32B', 'default-user', 'lmsp', '', '{\n \"architecture\": \"deepseek-r1\",\n \"template_system\": \"<｜begin▁of▁sentence｜>{{ .System }}\",\n \"template_user\": \"<｜User｜>{{ .Prompt }}\",\n \"template_assistant\": \"<｜Assistant｜>{{ .Answer }}<｜end▁of▁sentence｜>\",\n \"template_system_key\": \"{{ .System }}\",\n \"template_user_key\": \"{{ .Prompt }}\",\n \"template_assistant_key\": \"{{ .Answer }}\",\n \"template_assistant_head\": \"<｜Assistant｜>\\n\",\n \"system_prompt\": \"\",\n \"knowledge_base_prompt\": \"使用以下信息：\\n\\n{{ .Knowledge }}\\n\\n简洁地用中文回答：{{ .Prompt }}\",\n \"template_knowledge_key\": \"{{ .Knowledge }}\",\n \"knowledge_not_found_response\": \"在知识库中没有找到相关信息。请更换问题后重试。\",\n \"stop_words\": [\n \"<｜begin▁of▁sentence｜>\",\n \"<｜end▁of▁sentence｜>\",\n \"\"\n ],\n \"bad_words\": null\n}', '2025-03-17 11:09:00', '2025-03-17 11:09:00', '', 0); - --- --- 转储表的索引 --- - --- --- 表的索引 `model_tbl` --- -ALTER TABLE `model_tbl` - ADD PRIMARY KEY (`id_pk`), - ADD UNIQUE KEY `idx_model_id` (`model_id`); - --- --- 在导出的表使用AUTO_INCREMENT --- - --- --- 使用表AUTO_INCREMENT `model_tbl` --- -ALTER TABLE `model_tbl` - MODIFY `id_pk` int NOT NULL AUTO_INCREMENT, AUTO_INCREMENT=29; -COMMIT; - -/*!40101 SET CHARACTER_SET_CLIENT=@OLD_CHARACTER_SET_CLIENT */; -/*!40101 SET CHARACTER_SET_RESULTS=@OLD_CHARACTER_SET_RESULTS */; -/*!40101 SET COLLATION_CONNECTION=@OLD_COLLATION_CONNECTION */; - --- --- 允许所有IP访问 --- -USE `mysql`; -UPDATE `user` SET `host`= '%' WHERE `user`= 'root' LIMIT 1; diff --git a/seabed/docker-compose.yaml b/seabed/docker-compose.yaml deleted file mode 100644 index 3411a7b2c397dc6fc6e689a8b690351247929368..0000000000000000000000000000000000000000 --- a/seabed/docker-compose.yaml +++ /dev/null @@ -1,108 +0,0 @@ -name: metastone-all-in-one -services: - mysql_8_3: - ports: - - 3306:3306 - volumes: - - ../data/docker/mysql_data:/var/lib/mysql - - ../data/docker/mysql-init:/docker-entrypoint-initdb.d - environment: - - MYSQL_ROOT_PASSWORD=123456 - image: swr.cn-east-3.myhuaweicloud.com/cloud-mdgx/mysql_8_3:v0.0.1 - - llmapi-server-prod: - init: true - container_name: llmapi-server - hostname: llmapi-server - restart: always - ports: - - 8900:8900 - - 8901:8901 - volumes: - - /var/log/lmsp-api-server:/app/logs - - ../data/docker/datasets:/datasets - healthcheck: - test: curl -fs http://${HOST_IP}:8901/v1/ping || exit 1 - interval: 30s - timeout: 5s - retries: "2" - extra_hosts: - - "sy-s60-vllm-ubuntu2204:${HOST_IP}" - environment: - - LMSP_API_SERVER_ENABLE_CONSOLE_LOG=1 - - LMSP_API_SERVER_LOG_FILE_PATH=/app/logs/lmsp-api-server.log - - LMSP_API_SERVER_LOG_PANIC_FILE_PATH=/app/logs/lmsp-api-server-panic.log - - LMSP_API_SERVER_LOG_LEVEL=0 - - LMSP_API_SERVER_GRPC_LISTEN_ADDRESS=:8900 - - LMSP_API_SERVER_HTTP_LISTEN_ADDRESS=:8901 - - LMSP_API_SERVER_INFER_SRV_ENDPOINT=http://${HOST_IP}:8889/v1/chat/completions - - LMSP_API_SERVER_DATABASE_ADDRESS=${HOST_IP}:3306 - - LMSP_API_SERVER_DATABASE_USERNAME=root - - LMSP_API_SERVER_DATABASE_PASSWORD=123456 - - LMSP_API_SERVER_DATASET_ROOT=/datasets - - LMSP_API_SERVER_DISABLE_AUTH=true - - LMSP_API_SERVER_ALL_IN_ONE=true - image: swr.cn-east-3.myhuaweicloud.com/cloud-mdgx/llmapi-svr-prod:v0.0.1 - inference-svr-prod: - init: true - container_name: inference-server - hostname: inference-server - restart: always - ports: - - 8889:8889 - volumes: - - /var/log/inference-server:/app/logs - healthcheck: - test: curl -fs http://localhost:8889/v1/ping || exit 1 - interval: 30s - timeout: 5s - retries: "2" - extra_hosts: - - "sy-s60-vllm-ubuntu2204:${HOST_IP}" - environment: - - INFERENCE_SERVER_ENABLE_CONSOLE_LOG=1 - - INFERENCE_SERVER_LOG_FILE_PATH=/app/logs/inference-server.log - - INFERENCE_SERVER_LOG_PANIC_FILE_PATH=/app/logs/inference-server-panic.log - - INFERENCE_SERVER_LOG_LEVEL=1 - - INFERENCE_SERVER_LISTEN_ADDRESS=:8889 - - INFERENCE_SERVER_DATABASE_ADDRESS=${HOST_IP}:3306 - - INFERENCE_SERVER_DATABASE_USERNAME=root - - INFERENCE_SERVER_DATABASE_PASSWORD=123456 - - INFERENCE_SERVER_TENSORRT_LLM_SRV_ADDR=http://${HOST_IP}:8000 - - INFERENCE_SERVER_OLLAMA_SRV_ADDR=http://${HOST_IP}:11434 - - INFERENCE_SERVER_API_GRPC_ADDR=${HOST_IP}:8900 - - INFERENCE_SERVER_API_HTTP_ADDR=http://${HOST_IP}:8901 - - INFERENCE_SERVER_DISABLE_AUTH=true - image: swr.cn-east-3.myhuaweicloud.com/cloud-mdgx/inference-svr-prod:v0.0.1 - sy-s60-vllm: - stdin_open: true - tty: true - container_name: sy-s60-vllm-ubuntu2204 - hostname: sy-s60-vllm-ubuntu2204 - restart: always - volumes: - - ../data/docker/data/enflame-vllm/:/data/ - - ../models/:/models/ - - ../data/docker/lmsp_finetuned_models:/lmsp_finetuned_models - - /var/run/docker.sock:/var/run/docker.sock - working_dir: /data/ - privileged: true - network_mode: host - environment: - - LMSP_API_SERVER_ADDRESS=${HOST_IP}:8900 - - LMSP_AGENT_HOSTNAME=sy-s60-vllm-ubuntu2204 - - LMSP_AGENT_SERVER_CAPABILITIES=0,2 - - LMSP_AGENT_LOG_LEVEL=0 - image: swr.cn-east-3.myhuaweicloud.com/cloud-mdgx/sy-s60-vllm-ubuntu2204:v0.0.1 - web-ui-prod: - environment: - - HOST=${HOST_IP} - - NEXTAUTH_URL=http://${HOST_IP}:18088 - - PORT=3000 - - AUTH_SECRET=lQjkRoFlrjfBkDeAaBkv+Hodt40wO1pBW29gwKKC0U8= - - API_SERVER_URL=http://${HOST_IP}:8901/v1 - - INFERENCE_SERVER_URL=http://${HOST_IP}:8889/v1 - - NEXT_PUBLIC_DEPLOY_TO_MSCLOUD=false - ports: - - 18088:3000 - image: swr.cn-east-3.myhuaweicloud.com/cloud-mdgx/web-ui-prod:v0.0.1 diff --git a/seabed/install_seabed.sh b/seabed/install_seabed.sh deleted file mode 100755 index e8ce5b68a9625ee06af9aabbe94016556b54f859..0000000000000000000000000000000000000000 --- a/seabed/install_seabed.sh +++ /dev/null @@ -1,90 +0,0 @@ -#/bin/bash - -# install preqeruiement env; - -#!/bin/bash -#username="mdgx" # 自定义用户名 -#password="mdgx" # 自定义密码 -# -## 检查用户是否存在 -#if id "$username" &>/dev/null; then -# echo "mdgx用户已存在" -#else -# # 创建用户并设置密码（非交互式） -# sudo useradd -m -s /bin/bash "$username" # 创建用户及主目录[2,5](@ref) -# echo "$username:$password" | sudo chpasswd # 非交互式设置密码[2,5](@ref) -#fi - -#这个命令把mdgx 授予sudo执行权限 -sudo usermod -aG sudo $(whoami) - -#这个命令把mdgx 授予docker命令执行权限 -sudo usermod -aG docker $(whoami) - - -sudo chown -R $(whoami):$(whoami) $HOME/mdgx/ -cd $HOME/mdgx - -#set DNS server -#sudo echo 'nameserver 114.114.114.114' >> /etc/resolv.conf -#sudo echo 'nameserver 8.8.8.8' >> /etc/resolv.conf - -#apt source change -sudo cp /etc/apt/sources.list /etc/apt/sources.list.bak -sudo ./tools/change-apt-mirror.sh - -#set timezone -sudo timedatectl set-timezone Asia/Shanghai - -#apt-get update -sudo apt-get update - -sudo apt-get install python3-pip unzip -y - -#common tools -sudo apt-get install curl -y - -#SSH Server Configuration -sudo apt-get install openssh-server -y -#sed -i 's/PermitRootLogin prohibit-password/PermitRootLogin yes/' /etc/ssh/sshd_config && \ -#echo 'PasswordAuthentication yes' >> /etc/ssh/ssh_config - -#restart ssh server -sudo systemctl restart sshd.service - - -#install docker -sudo apt-get install jq -y -sudo apt-get install docker.io -y - -#add docker mirrors -sudo ./tools/change-docker-mirror.sh - -sudo systemctl daemon-reload -sudo systemctl restart docker - -#set pip mirror -#pip config set global.index-url https://mirrors.aliyun.com/pypi/simple -sudo pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple -sudo pip install modelscope -#sudo pip install docker-compose -./tools/install-dockercompose.sh - -#install GPU driver -sudo bash ./hwv/enflame/drivers/TopsRider_i3x_3.3.20250205_deb_amd64.run -y - -#for docker compose env -export HOST_IP=$(hostname -I | awk '{print $1}') -sudo echo "HOST_IP=$HOST_IP" > ./seabed/.env - -#log into HW Cloud and download image. -sudo docker login -u cn-east-3@TCPETCYF3CQUNEKNKNGM -p b818bd25b93aeeb355cbedd95e40cf20f270ff599120fa1aa24eb473224943d4 swr.cn-east-3.myhuaweicloud.com - -cd $HOME/mdgx/seabed - -sudo HOST_IP=$HOST_IP docker-compose pull - -#sudo docker pull swr.cn-east-3.myhuaweicloud.com/cloud-mdgx/mysql_8_3:v0.0.1 -#sudo docker pull swr.cn-east-3.myhuaweicloud.com/cloud-mdgx/llmapi-svr-prod:v0.0.1 -#sudo docker pull swr.cn-east-3.myhuaweicloud.com/cloud-mdgx/inference-svr-prod:v0.0.1 -#sudo docker pull swr.cn-east-3.myhuaweicloud.com/cloud-mdgx/sy-s60-vllm-0.7.2-prod-v0.0.1 diff --git a/seabed/start_seabed.sh b/seabed/start_seabed.sh deleted file mode 100755 index 8104ca4e09395991e33af5f6f8772a13a7a74f98..0000000000000000000000000000000000000000 --- a/seabed/start_seabed.sh +++ /dev/null @@ -1,3 +0,0 @@ - -cd $HOME/mdgx/seabed -sudo HOST_IP=$HOST_IP docker-compose up -d diff --git a/seabed/stop_seabed.sh b/seabed/stop_seabed.sh deleted file mode 100755 index f779a65249a6c12759c3414047241cdec4599e64..0000000000000000000000000000000000000000 --- a/seabed/stop_seabed.sh +++ /dev/null @@ -1,3 +0,0 @@ - -cd $HOME/mdgx/seabed -sudo HOST_IP=$HOST_IP docker-compose down diff --git a/tests/call-api/chat-completions.sh b/tests/call-api/chat-completions.sh deleted file mode 100755 index cc7c0556f28e7c68580187addc726bdac6199173..0000000000000000000000000000000000000000 --- a/tests/call-api/chat-completions.sh +++ /dev/null @@ -1,19 +0,0 @@ -curl --location 'http://10.10.6.246:8889/v1/chat/completions?infer-task-id=b20cd5ef-312f-4240-b2cf-79432a2c38c3' \ ---header 'Content-Type: application/json' \ ---data '{ - "model": "lmsp:DeepSeek-R1-Distill-Qwen-7B", - "messages": [ - { - "role": "user", - "content": "say you are a test" - } - ], - "stream": false, - "max_tokens": 10, - "temperature": 0.7, - "top_p": 0.9, - "top_k": 20 -}' - -# response -# {"reqId":"1742576711388482249","msg":"success","id":"1742576711388482249","object":"completions","created":1742576712,"model":"lmsp:DeepSeek-R1-Distill-Qwen-7B","system_fingerprint":"lmsp:DeepSeek-R1-Distill-Qwen-7B_lmsp_1742576711388482249","choices":[{"index":0,"message":{"role":"assistant","content":"Okay, so I just saw this messagewhere someone"},"logprobs":null,"finish_reason":"length"}]} diff --git a/tests/call-api/delete-infer-task.sh b/tests/call-api/delete-infer-task.sh deleted file mode 100755 index 6ea90efe5416d365873987a07eed9644f7d9f3bc..0000000000000000000000000000000000000000 --- a/tests/call-api/delete-infer-task.sh +++ /dev/null @@ -1,3 +0,0 @@ -curl --location --request DELETE 'http://10.10.6.246:8901/v1/tasks/infer/b20cd5ef-312f-4240-b2cf-79432a2c38c3' - -# response diff --git a/tests/call-api/init-infer-task.sh b/tests/call-api/init-infer-task.sh deleted file mode 100755 index 5befe841dea29d3a210552e9151323ce89321c1e..0000000000000000000000000000000000000000 --- a/tests/call-api/init-infer-task.sh +++ /dev/null @@ -1,41 +0,0 @@ -curl --location 'http://10.10.6.246:8901/v1/tasks/infer' \ ---header 'Content-Type: application/json' \ ---data '{ - "clientId": "allinone", - "name": "DeepSeek-R1-Distill-Qwen-7B", - "modelId": "5b406c08-771a-4669-ab58-949c0c188c0c", - "modelSource": "public_models", - "modelRuntimeInfo": { - "contextLen": 16000, - "batchSize": 8 - }, - "modelPlacementPolicy": { - "policyName": "balanced", - "hardwareResourceList": [ - { - "hostname": "allinone-enflame-vllm", - "devicesInfo": [ - { - "deviceType": 0, - "deviceMemInfoList": [ - { - "deviceId": "0", - "requestDeviceMemory": 77701, - "allowedDeviceMemory": 44007424 - }, - { - "deviceId": "1", - "requestDeviceMemory": 77701, - "allowedDeviceMemory": 44007424 - } - ] - } - ] - } - ] - }, - "inferEngineType": 2 -}' - -# response -# {"reqId":"1742576571620007025","msg":"success","data":{"inferTaskId":"b20cd5ef-312f-4240-b2cf-79432a2c38c3"}} diff --git a/tools/addmdgxsudodocker.sh b/tools/addmdgxsudodocker.sh deleted file mode 100755 index b1b0f455ca91ffb4ccebb685742189ca02f87148..0000000000000000000000000000000000000000 --- a/tools/addmdgxsudodocker.sh +++ /dev/null @@ -1,5 +0,0 @@ -#!/bin/bash -#sudo useradd mdgx || echo "created" -sudo usermod -aG sudo $(whoami) #这个命令把mdgx 授予sudo执行权限 -sudo usermod -aG docker $(whoami) #这个命令把mdgx 授予docker命令执行权限 - diff --git a/tools/change-apt-mirror.sh b/tools/change-apt-mirror.sh deleted file mode 100755 index b1a93669ae9a2c65fe510ab4ab3917dc7f632894..0000000000000000000000000000000000000000 --- a/tools/change-apt-mirror.sh +++ /dev/null @@ -1,41 +0,0 @@ -#!/bin/bash -# Ubuntu 多版本 APT 源一键切换脚本（阿里云镜像） -# 支持版本：Ubuntu 22.04(jammy)/20.04(focal)/18.04(bionic) -# 原理：单条 sed 命令完成全量替换 - -# 获取系统代号 -CODENAME=$(lsb_release -cs) -MIRROR="mirrors.aliyun.com" - -# 合法性检查 -if ! grep -qP "jammy|focal|bionic" <<< "$CODENAME"; then - echo "错误：仅支持 Ubuntu 22.04/20.04/18.04" >&2 - exit 1 -fi - -# 备份原始源文件 -BACKUP="/etc/apt/sources.list.bak.$(date +%s)" -sudo cp /etc/apt/sources.list "$BACKUP" || { - echo "备份失败，请检查权限！" >&2 - exit 1 -} - -# 单条 sed 命令完成全量替换（兼容多版本） -sudo sed -i \ - -e "s|http://$\w\+\.$\?archive.ubuntu.com|http://$MIRROR|g" \ - -e "s|http://security.ubuntu.com|http://$MIRROR|g" \ - -e "s|http://$\w\+\.$\?ports.ubuntu.com|http://$MIRROR|g" \ - /etc/apt/sources.list - -# 更新缓存 -echo -e "\n正在更新软件列表..." -sudo apt-get update - -# 输出状态 -echo -e "\n操作完成！" -echo "----------------------------------" -echo "当前镜像源：阿里云 (http://$MIRROR)" -echo "系统版本：$CODENAME" -echo "备份文件：$BACKUP" -echo "配置文件：/etc/apt/sources.list" -echo "----------------------------------" \ No newline at end of file diff --git a/tools/change-docker-mirror.sh b/tools/change-docker-mirror.sh deleted file mode 100755 index e64ab704623e809427b1c7b8965495121e83b463..0000000000000000000000000000000000000000 --- a/tools/change-docker-mirror.sh +++ /dev/null @@ -1,68 +0,0 @@ -#!/bin/bash -# Ubuntu 22.04 Docker 镜像源一键切换脚本 (企业级生产环境适用) -# 更新时间：2025年3月24日 -# 支持镜像源：阿里云/腾讯云/中科大/网易云 - -# 镜像源选项配置（可修改） -REGISTRY_MIRRORS=( - "https://docker.m.daocloud.io" - "https://mirror.ccs.tencentyun.com" - "https://docker.nju.edu.cn" - "https://dockerhub.azk8s.cn" -) - -# 检查 Docker 是否安装 -if ! command -v docker &> /dev/null; then - echo "错误：Docker 未安装，请先执行 apt install docker.io" >&2 - exit 1 -fi - -# 创建配置文件目录 -sudo mkdir -p /etc/docker - -# 备份原始配置（带时间戳） -BACKUP_FILE="/etc/docker/daemon.json.bak.$(date +%Y%m%d%H%M%S)" -if [ -f /etc/docker/daemon.json ]; then - sudo cp /etc/docker/daemon.json "$BACKUP_FILE" - echo "[OK] 配置文件已备份至 $BACKUP_FILE" -fi - -# 生成新配置 -sudo tee /etc/docker/daemon.json < /dev/null; then - echo "错误：生成的 JSON 配置文件语法错误，自动恢复备份！" >&2 - sudo mv -f "$BACKUP_FILE" /etc/docker/daemon.json - exit 1 -fi - -# 重载服务配置 -sudo systemctl daemon-reload - -# 重启 Docker 服务 -if sudo systemctl restart docker.service; then - echo "[OK] Docker 服务重启成功" -else - echo "错误：Docker 服务重启失败，请检查 journalctl -u docker.service" >&2 - exit 1 -fi - -# 验证配置生效 -echo -e "\n当前生效的镜像仓库列表：" -docker info | grep -A 5 "Registry Mirrors" - -# 速度测试建议 -echo -e "\n建议执行以下命令测试镜像拉取速度：" -echo " docker pull registry.cn-hangzhou.aliyuncs.com/library/ubuntu:22.04" \ No newline at end of file diff --git a/tools/cleandocker.sh b/tools/cleandocker.sh deleted file mode 100755 index 5a38ede0710e9d12b12d5bddef411b11347424cf..0000000000000000000000000000000000000000 --- a/tools/cleandocker.sh +++ /dev/null @@ -1,10 +0,0 @@ -#/bin/bash - -echo "stop all containers......" -docker stop $(docker ps -a) - -echo "removing all containers......" -docker rm -f $(docker ps -a -q) - -echo "removing all docker images......" -docker rmi -f $(docker images -q) diff --git a/tools/docker-compose b/tools/docker-compose deleted file mode 100755 index 682a7a7a0bdcb3239694fc77f7134dfb762b56b2..0000000000000000000000000000000000000000 Binary files a/tools/docker-compose and /dev/null differ diff --git a/tools/docker-run.sh b/tools/docker-run.sh deleted file mode 100755 index 3b5968d8d0330b751f67e41ea582c32a940d5ebf..0000000000000000000000000000000000000000 --- a/tools/docker-run.sh +++ /dev/null @@ -1,9 +0,0 @@ -docker run -p 3306:3306 -v /home/user/s60/docker/mysql_data:/var/lib/mysql -v /home/user/s60/docker/mysql-init:/docker-entrypoint-initdb.d -e MYSQL_ROOT_PASSWORD=123456 harbor.metastonecorp.com/ai-dev/mysql_8_3:8.3 - -docker run --init --name allinone-api-server -h allinone-api-server --restart always -p 8900:8900 -p 8901:8901 -v /var/log/lmsp-api-server:/app/logs -v /home/user/s60/docker/datasets:/datasets --health-cmd "curl -fs http://localhost:8901/v1/ping || exit 1" --health-interval 30s --health-retries 2 --health-timeout 5s --add-host allinone-enflame-vllm:10.10.6.246 -e LMSP_API_SERVER_ENABLE_CONSOLE_LOG=1 -e LMSP_API_SERVER_LOG_FILE_PATH=/app/logs/lmsp-api-server.log -e LMSP_API_SERVER_LOG_PANIC_FILE_PATH=/app/logs/lmsp-api-server-panic.log -e LMSP_API_SERVER_LOG_LEVEL=0 -e LMSP_API_SERVER_GRPC_LISTEN_ADDRESS=:8900 -e LMSP_API_SERVER_HTTP_LISTEN_ADDRESS=:8901 -e LMSP_API_SERVER_INFER_SRV_ENDPOINT=http://10.10.6.246:8889/v1/chat/completions -e LMSP_API_SERVER_DATABASE_ADDRESS=10.10.6.246:3306 -e LMSP_API_SERVER_DATABASE_USERNAME=root -e LMSP_API_SERVER_DATABASE_PASSWORD=123456 -e LMSP_API_SERVER_DATASET_ROOT=/datasets -e LMSP_API_SERVER_DISABLE_AUTH=true -e LMSP_API_SERVER_ALL_IN_ONE=true harbor.metastonecorp.com/ai-dev/allinone-api-server:0.1.0 - -docker run --init --name allinone-inference-server -h allinone-inference-server --restart always -p 8889:8889 -v /var/log/inference-server:/app/logs --health-cmd "curl -fs http://localhost:8889/v1/ping || exit 1" --health-interval 30s --health-retries 2 --health-timeout 5s --add-host allinone-enflame-vllm:10.10.6.246 -e INFERENCE_SERVER_ENABLE_CONSOLE_LOG=1 -e INFERENCE_SERVER_LOG_FILE_PATH=/app/logs/inference-server.log -e INFERENCE_SERVER_LOG_PANIC_FILE_PATH=/app/logs/inference-server-panic.log -e INFERENCE_SERVER_LOG_LEVEL=1 -e INFERENCE_SERVER_LISTEN_ADDRESS=:8889 -e INFERENCE_SERVER_DATABASE_ADDRESS=10.10.6.246:3306 -e INFERENCE_SERVER_DATABASE_USERNAME=root -e INFERENCE_SERVER_DATABASE_PASSWORD=123456 -e INFERENCE_SERVER_TENSORRT_LLM_SRV_ADDR=http://10.10.6.246:8000 -e INFERENCE_SERVER_OLLAMA_SRV_ADDR=http://10.10.6.246:11434 -e INFERENCE_SERVER_API_GRPC_ADDR=10.10.6.246:8900 -e INFERENCE_SERVER_API_HTTP_ADDR=http://10.10.6.246:8901 -e INFERENCE_SERVER_DISABLE_AUTH=true harbor.metastonecorp.com/ai-dev/allinone-inference-server:0.0.4 - -docker run --net host -i -t --name allinone-enflame-vllm -h allinone-enflame-vllm --restart always -v /home/user/s60/docker/data/enflame-vllm/:/data/ -v /home/user/models/:/models/ -v /home/user/s60/docker/lmsp_finetuned_models:/lmsp_finetuned_models -v /var/run/docker.sock:/var/run/docker.sock -w /data/ --privileged -e LMSP_API_SERVER_ADDRESS=10.10.6.246:8900 -e LMSP_AGENT_HOSTNAME=allinone-enflame-vllm -e LMSP_AGENT_SERVER_CAPABILITIES=0,2 -e LMSP_AGENT_LOG_LEVEL=0 allinone-enflame-vllm-ubuntu2204:0.0.6 - -docker run -e HOST=10.10.6.246 -e NEXTAUTH_URL=http://10.10.6.246:18088 -e PORT=3000 -e AUTH_SECRET=lQjkRoFlrjfBkDeAaBkv+Hodt40wO1pBW29gwKKC0U8= -e API_SERVER_URL=http://10.10.6.246:8901/v1 -e INFERENCE_SERVER_URL=http://10.10.6.246:8889/v1 -e NEXT_PUBLIC_DEPLOY_TO_MSCLOUD=false -p 18088:3000 harbor.metastonecorp.com/ai-dev/aiallinone:v0.0.1 diff --git a/tools/download_models.py b/tools/download_models.py deleted file mode 100755 index 4fa3fd1a2a327511ebf3c9418bdc19c3883d7939..0000000000000000000000000000000000000000 --- a/tools/download_models.py +++ /dev/null @@ -1,78 +0,0 @@ -#!/usr/bin/env python3 - -import os -from modelscope.hub.snapshot_download import snapshot_download - -# 配置文件路径 -CONFIG_FILE = "models.ini" - -def read_config(file_path): - config = {} - try: - with open(file_path, 'r') as file: - for line in file: - line = line.strip() - if line and not line.startswith('#'): - key, value = line.split('=', 1) - config[key.strip()] = value.strip() - print(f"\e[1;34mConfig file read successfully: {config}\e[0m") - except Exception as e: - print(f"\e[1;31mError reading config file: {e}\e[0m") - return config - -def download_model(source, model_name, download_dir): - try: - if source == "ModelScope": - print(f"\e[1;34mDownloading model {model_name} from ModelScope...\e[0m") - model_dir = snapshot_download(model_name, cache_dir=download_dir) - print(f"\e[1;32mModel {model_name} downloaded successfully to {model_dir}.\e[0m") - else: - print(f"\e[1;31mUnsupported source: {source}. Please set source to ModelScope.\e[0m") - return False - return True - except Exception as e: - print(f"\e[1;31mFailed to download model {model_name}: {e}\e[0m") - return False - -def main(): - # 读取配置文件 - config = read_config(CONFIG_FILE) - - # 检查配置项 - required_keys = ["source", "models", "download_dir"] - for key in required_keys: - if key not in config: - print(f"\e[1;31mError: Config file missing key: {key}\e[0m") - exit(1) - - source = config["source"] - models = config["models"].split(',') - model_dir = config["download_dir"] - - # 构建完整的下载目录路径 - home_dir = os.path.expanduser('~') - download_dir = os.path.join(home_dir, 'mdgx') - download_dir = os.path.join(download_dir, model_dir) - - # 打印构建后的 download_dir 路径 - print(f"\e[1;34mConstructed download directory: {download_dir}\e[0m") - - # 创建下载目录（如果不存在） - os.makedirs(download_dir, exist_ok=True) - print(f"\e[1;34mDownload directory created: {download_dir}\e[0m") - - # 计算总模型数量 - total_models = len(models) - current_model = 0 - - - # 下载每个模型 - for model in models: - current_model += 1 - print(f"\e[1;34mDownloading model {current_model}/{total_models}: {model}\e[0m") - download_model(source, model, download_dir) - - print(f"\e[1;32mAll models downloaded successfully.\e[0m") - -if __name__ == "__main__": - main() diff --git a/tools/install-dockercompose.sh b/tools/install-dockercompose.sh deleted file mode 100755 index 79306cf37b7fc422425f1ca114e19550f17f1f2f..0000000000000000000000000000000000000000 --- a/tools/install-dockercompose.sh +++ /dev/null @@ -1,9 +0,0 @@ -#!/bin/bash -sudo cp $HOME/mdgx/tools/docker-compose /usr/local/bin/docker-compose -# 添加执行权限 -sudo chmod +x /usr/local/bin/docker-compose -sudo docker-compose --version - - - - diff --git a/tools/models.ini b/tools/models.ini deleted file mode 100755 index 619dc05e70b8b880866a9c80b9517ef7c0b2ae1e..0000000000000000000000000000000000000000 --- a/tools/models.ini +++ /dev/null @@ -1,3 +0,0 @@ -source=ModelScope -models=deepseek-ai/DeepSeek-R1-Distill-Qwen-14B,deepseek-ai/DeepSeek-R1-Distill-Qwen-32B,Qwen/Qwen2.5-32B-Instruct -download_dir=models diff --git a/tools/set-host-ip.sh b/tools/set-host-ip.sh deleted file mode 100755 index b10481ed68e9f2d29c4968afa3026e372dc9fa74..0000000000000000000000000000000000000000 --- a/tools/set-host-ip.sh +++ /dev/null @@ -1,22 +0,0 @@ -#!/bin/bash - -# 获取当前用户的主目录 -USER_HOME=$(eval echo ~$(whoami)) - -# 获取主机IP地址 -HOST_IP=$(hostname -I | awk '{print $1}') -if [ -z "$HOST_IP" ]; then - echo "Error: Unable to determine the host IP address." - exit 1 -fi - -echo "HOST_IP is $HOST_IP" -# 导出变量到当前脚本环境 -export HOST_IP - -# 将变量写入目标用户的 .bashrc 文件 -echo "export HOST_IP=\$(hostname -I | awk '{print \$1}')" >> $USER_HOME/.bashrc - -# 重新加载 .bashrc 文件 -source $USER_HOME/.bashrc - diff --git a/tools/swr_login.sh b/tools/swr_login.sh deleted file mode 100755 index 7537e85db2d559c1a49944bd16bab5350e46e82a..0000000000000000000000000000000000000000 --- a/tools/swr_login.sh +++ /dev/null @@ -1,3 +0,0 @@ -#!/bin/bash -sudo docker login -u cn-east-3@TCPETCYF3CQUNEKNKNGM -p b818bd25b93aeeb355cbedd95e40cf20f270ff599120fa1aa24eb473224943d4 swr.cn-east-3.myhuaweicloud.com -