Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

💄 style: add ModelFetcher for supported provider #5316

Merged
merged 28 commits into from
Jan 21, 2025
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
28 commits
Select commit Hold shift + click to select a range
c84e84f
💄 style: add ModelFetcher for supported provider
hezhijie0327 Jan 6, 2025
ec83218
💄 style: add Mistral ModelFetcher support
hezhijie0327 Jan 7, 2025
b7e67c8
💄 style: add `glm-zero-preview` model for Zhipu
hezhijie0327 Jan 7, 2025
e056ce7
💄 style: add more supported providers
hezhijie0327 Jan 7, 2025
4757282
💄 style: add SenseNova & Baichuan support with custom models endpoint
hezhijie0327 Jan 7, 2025
18e0395
🐛 fix: fix fc keyword typo, should lowercase
hezhijie0327 Jan 7, 2025
b429652
Merge branch 'lobehub:main' into fetchModels
hezhijie0327 Jan 7, 2025
7fb95a0
💄 style: update aiModels list
hezhijie0327 Jan 7, 2025
049662d
♻️ refactor: use aiModels sheme
hezhijie0327 Jan 7, 2025
f27b46b
Revert "♻️ refactor: use aiModels sheme" (#252)
hezhijie0327 Jan 7, 2025
7fb875d
🐛 fix: fix `firefunction` series not set as functionCall
hezhijie0327 Jan 7, 2025
80a1c4c
💄 style: add Zhipu FetchModel support
hezhijie0327 Jan 8, 2025
41eb03a
💄 style: add Anthropic FetchModel support
hezhijie0327 Jan 8, 2025
adb7234
💄 style: add Google ModelFetch support
hezhijie0327 Jan 8, 2025
de09627
Merge branch 'main' into fetchModels
hezhijie0327 Jan 8, 2025
98ed8d7
💄 style: add HuggingFace ModelFetch support
hezhijie0327 Jan 13, 2025
20b6324
🐛 fix: fix stepfun fc tag
hezhijie0327 Jan 13, 2025
10155f6
♻️ refactor: refactor `LOBE_DEFAULT_MODEL_LIST` import
hezhijie0327 Jan 14, 2025
3b9a390
Merge branch 'main' into fetchModels
hezhijie0327 Jan 14, 2025
8e73626
🔨 chore: revert some changes
hezhijie0327 Jan 15, 2025
ab0827f
Merge branch 'main' into fetchModels
hezhijie0327 Jan 15, 2025
a448c61
🐛 fix: fix aiModels import
hezhijie0327 Jan 15, 2025
48a70c5
🐛 fix: fix CI error
hezhijie0327 Jan 15, 2025
d6fd159
💄 style: add `MiniMax-Text-01`
hezhijie0327 Jan 15, 2025
c71e673
💄 style: add vision model support for Moonshot
hezhijie0327 Jan 15, 2025
f244702
Merge branch 'main' into fetchModels
hezhijie0327 Jan 16, 2025
2139cb6
Merge branch 'main' into fetchModels
hezhijie0327 Jan 17, 2025
b4501de
Merge branch 'main' into fetchModels
hezhijie0327 Jan 20, 2025
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
50 changes: 23 additions & 27 deletions src/config/aiModels/ai360.ts
Original file line number Diff line number Diff line change
Expand Up @@ -2,16 +2,29 @@ import { AIChatModelCard } from '@/types/aiModel';

const ai360ChatModels: AIChatModelCard[] = [
{
contextWindowTokens: 8192,
contextWindowTokens: 8000,
description:
'360GPT2 Pro 是 360 公司推出的高级自然语言处理模型,具备卓越的文本生成和理解能力,尤其在生成与创作领域表现出色,能够处理复杂的语言转换和角色演绎任务。',
'360gpt2-o1 使用树搜索构建思维链,并引入了反思机制,使用强化学习训练,模型具备自我反思与纠错的能力。',
displayName: '360GPT2 o1',
enabled: true,
id: '360gpt2-o1',
pricing: {
currency: 'CNY',
input: 20,
output: 50,
},
type: 'chat',
},
{
contextWindowTokens: 8000,
description:
'360智脑系列效果最好的主力千亿级大模型,广泛适用于各领域复杂任务场景。',
displayName: '360GPT2 Pro',
enabled: true,
id: '360gpt2-pro',
maxOutput: 7000,
pricing: {
currency: 'CNY',
input: 5,
input: 2,
output: 5,
},
type: 'chat',
Expand All @@ -20,46 +33,29 @@ const ai360ChatModels: AIChatModelCard[] = [
abilities: {
functionCall: true,
},
contextWindowTokens: 8192,
contextWindowTokens: 8000,
description:
'360GPT Pro 作为 360 AI 模型系列的重要成员,以高效的文本处理能力满足多样化的自然语言应用场景,支持长文本理解和多轮对话等功能。',
'360智脑系列效果最好的主力千亿级大模型,广泛适用于各领域复杂任务场景。',
displayName: '360GPT Pro',
enabled: true,
id: '360gpt-pro',
maxOutput: 7000,
pricing: {
currency: 'CNY',
input: 5,
input: 2,
output: 5,
},
type: 'chat',
},
{
contextWindowTokens: 8192,
contextWindowTokens: 7000,
description:
'360GPT Turbo 提供强大的计算和对话能力,具备出色的语义理解和生成效率,是企业和开发者理想的智能助理解决方案。',
'兼顾性能和效果的百亿级大模型,适合对性能/成本要求较高 的场景。',
displayName: '360GPT Turbo',
enabled: true,
id: '360gpt-turbo',
maxOutput: 7000,
pricing: {
currency: 'CNY',
input: 2,
output: 2,
},
type: 'chat',
},
{
contextWindowTokens: 8192,
description:
'360GPT Turbo Responsibility 8K 强调语义安全和责任导向,专为对内容安全有高度要求的应用场景设计,确保用户体验的准确性与稳健性。',
displayName: '360GPT Turbo Responsibility 8K',
enabled: true,
id: '360gpt-turbo-responsibility-8k',
maxOutput: 2048,
pricing: {
currency: 'CNY',
input: 2,
input: 1,
output: 2,
},
type: 'chat',
Expand Down
92 changes: 61 additions & 31 deletions src/config/aiModels/giteeai.ts
Original file line number Diff line number Diff line change
Expand Up @@ -7,24 +7,35 @@ const giteeaiChatModels: AIChatModelCard[] = [
},
contextWindowTokens: 16_000,
description:
'Qwen2.5-72B-Instruct 支持 16k 上下文, 生成长文本超过 8K 。支持 function call 与外部系统无缝交互,极大提升了灵活性和扩展性。模型知识明显增加,并且大大提高了编码和数学能力, 多语言支持超过 29 种',
'Qwen2.5-72B-Instruct 支持 16k 上下文, 生成长文本超过 8K 。支持 function call 与外部系统无缝交互,极大提升了灵活性和扩展性。模型知识明显增加,并且大大提高了编码和数学能力, 多语言支持超过 29 种',
displayName: 'Qwen2.5 72B Instruct',
enabled: true,
id: 'Qwen2.5-72B-Instruct',
type: 'chat',
},
{
contextWindowTokens: 32_000,
description:
'Qwen2.5-Coder-32B-Instruct 是一款专为代码生成、代码理解和高效开发场景设计的大型语言模型,采用了业界领先的32B参数规模,能够满足多样化的编程需求。',
displayName: 'Qwen2.5 Coder 32B Instruct',
'Qwen2.5-32B-Instruct 是一款 320 亿参数的大语言模型,性能表现均衡,优化中文和多语言场景,支持智能问答、内容生成等应用。',
displayName: 'Qwen2.5 32B Instruct',
enabled: true,
id: 'Qwen2.5-Coder-32B-Instruct',
id: 'Qwen2.5-32B-Instruct',
type: 'chat',
},
{
contextWindowTokens: 24_000,
description:
'Qwen2.5-14B-Instruct 是一款 140 亿参数的大语言模型,性能表现优秀,优化中文和多语言场景,支持智能问答、内容生成等应用。',
displayName: 'Qwen2.5 14B Instruct',
enabled: true,
id: 'Qwen2.5-14B-Instruct',
type: 'chat',
},
{
abilities: {
functionCall: true,
},
contextWindowTokens: 32_000,
description:
'Qwen2.5-7B-Instruct 是一款 70 亿参数的大语言模型,支持 function call 与外部系统无缝交互,极大提升了灵活性和扩展性。优化中文和多语言场景,支持智能问答、内容生成等应用。',
displayName: 'Qwen2.5 7B Instruct',
Expand All @@ -33,59 +44,75 @@ const giteeaiChatModels: AIChatModelCard[] = [
type: 'chat',
},
{
contextWindowTokens: 32_000,
description:
'Qwen2.5-32B-Instruct 是一款 320 亿参数的大语言模型,性能表现均衡,优化中文和多语言场景,支持智能问答、内容生成等应用。',
displayName: 'Qwen2.5 32B Instruct',
enabled: true,
id: 'Qwen2.5-32B-Instruct',
'Qwen2 是 Qwen 模型的最新系列,支持 128k 上下文,对比当前最优的开源模型,Qwen2-72B 在自然语言理解、知识、代码、数学及多语言等多项能力上均显著超越当前领先的模型。',
displayName: 'Qwen2 72B Instruct',
id: 'Qwen2-72B-Instruct',
type: 'chat',
},
{
contextWindowTokens: 24_000,
description:
'Qwen2.5-14B-Instruct 是一款 140 亿参数的大语言模型,性能表现优秀,优化中文和多语言场景,支持智能问答、内容生成等应用。',
displayName: 'Qwen2.5 14B Instruct',
enabled: true,
id: 'Qwen2.5-14B-Instruct',
'Qwen2 是 Qwen 模型的最新系列,能够超越同等规模的最优开源模型甚至更大规模的模型,Qwen2 7B 在多个评测上取得显著的优势,尤其是代码及中文理解上。',
displayName: 'Qwen2 7B Instruct',
id: 'Qwen2-7B-Instruct',
type: 'chat',
},
{
contextWindowTokens: 6000,
contextWindowTokens: 32_000,
description:
'Qwen2 是 Qwen 模型的最新系列,支持 128k 上下文,对比当前最优的开源模型,Qwen2-72B 在自然语言理解、知识、代码、数学及多语言等多项能力上均显著超越当前领先的模型。',
displayName: 'Qwen2 72B Instruct',
id: 'Qwen2-72B-Instruct',
'Qwen2.5-Coder-32B-Instruct 是一款专为代码生成、代码理解和高效开发场景设计的大型语言模型,采用了业界领先的32B参数规模,能够满足多样化的编程需求。',
displayName: 'Qwen2.5 Coder 32B Instruct',
enabled: true,
id: 'Qwen2.5-Coder-32B-Instruct',
type: 'chat',
},
{
contextWindowTokens: 32_000,
contextWindowTokens: 24_000,
description:
'Qwen2 是 Qwen 模型的最新系列,能够超越同等规模的最优开源模型甚至更大规模的模型,Qwen2 7B 在多个评测上取得显著的优势,尤其是代码及中文理解上。',
displayName: 'Qwen2 7B Instruct',
id: 'Qwen2-7B-Instruct',
'Qwen2.5-Coder-14B-Instruct 是一款基于大规模预训练的编程指令模型,具备强大的代码理解和生成能力,能够高效地处理各种编程任务,特别适合智能代码编写、自动化脚本生成和编程问题解答。',
displayName: 'Qwen2.5 Coder 14B Instruct',
enabled: true,
id: 'Qwen2.5-Coder-14B-Instruct',
type: 'chat',
},
{
abilities: {
vision: true,
},
contextWindowTokens: 32_000,
description:
'InternVL2-8B 是一款强大的视觉语言模型,支持图像与文本的多模态处理,能够精确识别图像内容并生成相关描述或回答。',
displayName: 'InternVL2 8B',
'Qwen2-VL-72B是一款强大的视觉语言模型,支持图像与文本的多模态处理,能够精确识别图像内容并生成相关描述或回答。',
displayName: 'Qwen2 VL 72B',
enabled: true,
id: 'InternVL2-8B',
id: 'Qwen2-VL-72B',
type: 'chat',
},
{
abilities: {
vision: true,
},
contextWindowTokens: 32_000,
description:
'InternVL2.5-26B 是一款强大的视觉语言模型,支持图像与文本的多模态处理,能够精确识别图像内容并生成相关描述或回答。',
displayName: 'InternVL2.5 26B',
enabled: true,
id: 'InternVL2.5-26B',
type: 'chat',
},
{
abilities: {
vision: true,
},
contextWindowTokens: 32_000,
description:
'InternVL2-8B 是一款强大的视觉语言模型,支持图像与文本的多模态处理,能够精确识别图像内容并生成相关描述或回答。',
displayName: 'InternVL2 8B',
enabled: true,
id: 'InternVL2-8B',
type: 'chat',
},
{
contextWindowTokens: 32_000,
description:
Expand All @@ -104,25 +131,28 @@ const giteeaiChatModels: AIChatModelCard[] = [
id: 'Yi-34B-Chat',
type: 'chat',
},
/*
// not compatible with OpenAI SDK
{
contextWindowTokens: 8000,
description:
'DeepSeek Coder 33B 是一个代码语言模型, 基于 2 万亿数据训练而成,其中 87% 为代码, 13% 为中英文语言。模型引入 16K 窗口大小和填空任务,提供项目级别的代码补全和片段填充功能。',
displayName: 'DeepSeek Coder 33B Instruct',
'代码小浣熊是基于商汤大语言模型的软件智能研发助手,覆盖软件需求分析、架构设计、代码编写、软件测试等环节,满足用户代码编写、编程学习等各类需求。代码小浣熊支持 Python、Java、JavaScript、C++、Go、SQL 等 90+主流编程语言和 VS Code、IntelliJ IDEA 等主流 IDE。在实际应用中,代码小浣熊可帮助开发者提升编程效率超 50%。',
displayName: 'code raccoon v1',
enabled: true,
id: 'deepseek-coder-33B-instruct',
id: 'code-raccoon-v1',
type: 'chat',
},
*/
{
contextWindowTokens: 8000,
description:
'代码小浣熊是基于商汤大语言模型的软件智能研发助手,覆盖软件需求分析、架构设计、代码编写、软件测试等环节,满足用户代码编写、编程学习等各类需求。代码小浣熊支持 Python、Java、JavaScript、C++、Go、SQL 等 90+主流编程语言和 VS Code、IntelliJ IDEA 等主流 IDE。在实际应用中,代码小浣熊可帮助开发者提升编程效率超 50%。',
displayName: 'code raccoon v1',
'DeepSeek Coder 33B 是一个代码语言模型, 基于 2 万亿数据训练而成,其中 87% 为代码, 13% 为中英文语言。模型引入 16K 窗口大小和填空任务,提供项目级别的代码补全和片段填充功能。',
displayName: 'DeepSeek Coder 33B Instruct',
enabled: true,
id: 'code-raccoon-v1',
id: 'deepseek-coder-33B-instruct',
type: 'chat',
},
{
contextWindowTokens: 40_000,
contextWindowTokens: 32_000,
description:
'CodeGeeX4-ALL-9B 是一个多语言代码生成模型,支持包括代码补全和生成、代码解释器、网络搜索、函数调用、仓库级代码问答在内的全面功能,覆盖软件开发的各种场景。是参数少于 10B 的顶尖代码生成模型。',
displayName: 'CodeGeeX4 All 9B',
Expand Down
19 changes: 19 additions & 0 deletions src/config/aiModels/minimax.ts
Original file line number Diff line number Diff line change
@@ -1,6 +1,25 @@
import { AIChatModelCard } from '@/types/aiModel';

const minimaxChatModels: AIChatModelCard[] = [
{
abilities: {
functionCall: true,
vision: true,
},
contextWindowTokens: 1_000_192,
description: '在 MiniMax-01系列模型中,我们做了大胆创新:首次大规模实现线性注意力机制,传统 Transformer架构不再是唯一的选择。这个模型的参数量高达4560亿,其中单次激活459亿。模型综合性能比肩海外顶尖模型,同时能够高效处理全球最长400万token的上下文,是GPT-4o的32倍,Claude-3.5-Sonnet的20倍。',
displayName: 'MiniMax-Text-01',
enabled: true,
id: 'MiniMax-Text-01',
maxOutput: 1_000_192,
pricing: {
currency: 'CNY',
input: 10,
output: 10,
},
releasedAt: '2025-01-15',
type: 'chat',
},
{
abilities: {
functionCall: true,
Expand Down
74 changes: 73 additions & 1 deletion src/config/aiModels/moonshot.ts
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,11 @@ const moonshotChatModels: AIChatModelCard[] = [
displayName: 'Moonshot V1 8K',
enabled: true,
id: 'moonshot-v1-8k',
pricing: {
currency: 'CNY',
input: 12,
output: 12,
},
type: 'chat',
},
{
Expand All @@ -23,18 +28,85 @@ const moonshotChatModels: AIChatModelCard[] = [
displayName: 'Moonshot V1 32K',
enabled: true,
id: 'moonshot-v1-32k',
pricing: {
currency: 'CNY',
input: 24,
output: 24,
},
type: 'chat',
},
{
abilities: {
functionCall: true,
},
contextWindowTokens: 128_000,
contextWindowTokens: 131_072,
description:
'Moonshot V1 128K 是一款拥有超长上下文处理能力的模型,适用于生成超长文本,满足复杂的生成任务需求,能够处理多达128,000个tokens的内容,非常适合科研、学术和大型文档生成等应用场景。',
displayName: 'Moonshot V1 128K',
enabled: true,
id: 'moonshot-v1-128k',
pricing: {
currency: 'CNY',
input: 60,
output: 60,
},
type: 'chat',
},
{
abilities: {
functionCall: true,
vision: true,
},
contextWindowTokens: 8192,
description:
'Kimi 视觉模型(包括 moonshot-v1-8k-vision-preview/moonshot-v1-32k-vision-preview/moonshot-v1-128k-vision-preview 等)能够理解图片内容,包括图片文字、图片颜色和物体形状等内容。',
displayName: 'Moonshot V1 8K Vision Preview',
enabled: true,
id: 'moonshot-v1-8k-vision-preview',
pricing: {
currency: 'CNY',
input: 12,
output: 12,
},
releasedAt: '2025-01-14',
type: 'chat',
},
{
abilities: {
functionCall: true,
vision: true,
},
contextWindowTokens: 32_768,
description:
'Kimi 视觉模型(包括 moonshot-v1-8k-vision-preview/moonshot-v1-32k-vision-preview/moonshot-v1-128k-vision-preview 等)能够理解图片内容,包括图片文字、图片颜色和物体形状等内容。',
displayName: 'Moonshot V1 32K Vision Preview',
enabled: true,
id: 'moonshot-v1-32k-vision-preview',
pricing: {
currency: 'CNY',
input: 24,
output: 24,
},
releasedAt: '2025-01-14',
type: 'chat',
},
{
abilities: {
functionCall: true,
vision: true,
},
contextWindowTokens: 131_072,
description:
'Kimi 视觉模型(包括 moonshot-v1-8k-vision-preview/moonshot-v1-32k-vision-preview/moonshot-v1-128k-vision-preview 等)能够理解图片内容,包括图片文字、图片颜色和物体形状等内容。',
displayName: 'Moonshot V1 128K Vision Preview',
enabled: true,
id: 'moonshot-v1-128k-vision-preview',
pricing: {
currency: 'CNY',
input: 60,
output: 60,
},
releasedAt: '2025-01-14',
type: 'chat',
},
];
Expand Down
Loading
Loading