From 043218937f84c7aa8cf49a48c3d9990de4390bba Mon Sep 17 00:00:00 2001 From: jialin Date: Sun, 1 Feb 2026 13:08:36 +0800 Subject: [PATCH] chore: benchmark detail locales --- src/components/icon-font/index.tsx | 2 +- src/locales/en-US/benchmark.ts | 55 +++++++- src/locales/en-US/resources.ts | 2 +- src/locales/ja-JP/benchmark.ts | 55 +++++++- src/locales/ja-JP/resources.ts | 2 +- src/locales/ru-RU/benchmark.ts | 55 +++++++- src/locales/ru-RU/resources.ts | 2 +- src/locales/zh-CN/benchmark.ts | 55 +++++++- src/locales/zh-CN/resources.ts | 2 +- .../components/environment/index.tsx | 39 +++--- .../benchmark/components/left-actions.tsx | 10 +- .../benchmark/components/right-actions.tsx | 25 +--- .../components/summary/benchmark.tsx | 17 ++- .../benchmark/components/summary/index.tsx | 9 +- .../benchmark/components/summary/instance.tsx | 45 ++++-- .../components/summary/metrics-result.tsx | 130 +++++------------- .../components/summary/percentile-result.tsx | 111 ++++++++------- src/pages/benchmark/config/index.ts | 17 ++- src/pages/benchmark/forms/dataset.tsx | 22 ++- src/pages/benchmark/index.tsx | 11 +- .../benchmark/services/use-query-profiles.ts | 10 +- .../llmodels/components/catalog-item.tsx | 14 +- src/pages/resources/hooks/use-gpu-columns.tsx | 2 +- 23 files changed, 450 insertions(+), 242 deletions(-) diff --git a/src/components/icon-font/index.tsx b/src/components/icon-font/index.tsx index 3f015b2a..48f1b99b 100644 --- a/src/components/icon-font/index.tsx +++ b/src/components/icon-font/index.tsx @@ -2,7 +2,7 @@ import { createFromIconfontCN } from '@ant-design/icons'; // import './iconfont/iconfont.js'; const IconFont = createFromIconfontCN({ - scriptUrl: '//at.alicdn.com/t/c/font_4613488_mis8x88kojd.js' + scriptUrl: '//at.alicdn.com/t/c/font_4613488_pxk0gz02dig.js' }); export default IconFont; diff --git a/src/locales/en-US/benchmark.ts b/src/locales/en-US/benchmark.ts index f515d968..73732095 100644 --- a/src/locales/en-US/benchmark.ts +++ b/src/locales/en-US/benchmark.ts @@ -1,6 +1,7 @@ export default { 'benchmark.title': 'Benchmarks', 'benchmark.button.add': 'Add Benchmark', + 'benchmark.button.edit': 'Edit Benchmark', 'benchmark.button.compare': 'Compare', 'benchmark.table.model': 'Model', 'benchmark.table.instance': 'Model Instance', @@ -26,5 +27,57 @@ export default { 'benchmark.detail.summary.title': 'Summary', 'benchmark.detail.configure.title': 'Configure', 'benchmark.detail.environment.title': 'Environment', - 'benchmark.detail.logs.title': 'Logs' + 'benchmark.detail.logs.title': 'Logs', + 'benchmark.form.totalRequests.tips': 'Total Requests', + 'benchmark.form.profile.latency.tips': + 'Single-request latency benchmark focusing on TTFT and tail latency. Suitable for API serving and chat scenarios.', + 'benchmark.form.profile.throughput.tips': + 'Max throughput with realistic prompt length. The most commonly used baseline for GPU and model comparison.', + 'benchmark.form.profile.longContext.tips': + 'Stress test for long-context handling. Evaluates KV cache behavior, memory usage, and backend stability.', + 'benchmark.form.profile.heavy.tips': + 'Decode-heavy generation benchmark. Measures sustained decoding speed and output token throughput.', + 'benchmark.table.filter.bygpu': 'Filter by GPU', + 'benchmark.table.filter.bymodel': 'Filter by Model', + 'benchmark.table.filter.bydataset': 'Filter by Dataset', + 'benchmark.table.avg': 'Average', + 'benchmark.table.columnSettings': 'Column Settings', + 'benchmark.detail.summary.results': 'Test Results', + 'benchmark.detail.summary.metrics': 'Test Metrics', + 'benchmark.detail.summary.latency': 'Latency', + 'benchmark.detail.summary.throughput': 'Throughput', + 'benchmark.detail.summary.request': 'Requests', + 'benchmark.detail.avg.reqLatency': 'Average Request Latency', + 'benchmark.detail.avg.ttft': 'Average TTFT', + 'benchmark.detail.avg.tpot': 'Average TPOT', + 'benchmark.detail.avg.itl': 'Average ITL', + 'benchmark.detail.throughput.totalToken': 'Total Token Throughput', + 'benchmark.detail.throughput.inputToken': 'Input Token Throughput', + 'benchmark.detail.throughput.outputToken': 'Output Token Throughput', + 'benchmark.detail.requests.total': 'Total Requests', + 'benchmark.detail.requests.success': 'Successful Requests', + 'benchmark.detail.requests.failed': 'Failed Requests', + 'benchmark.detail.requests.concurrency': 'Concurrency', + 'benchmark.detail.percentile.latency': 'Latency', + 'benchmark.detail.percentile.inputTokens': 'Input Tokens', + 'benchmark.detail.percentile.outputTokens': 'Output Tokens', + 'benchmark.detail.percentile.input': 'Input', + 'benchmark.detail.percentile.output': 'Output', + 'benchmark.detail.percentile.total': 'Total', + 'benchmark.detail.configure.tokenLength': 'Token Length (Input/Output)', + 'benchmark.detail.percentile.title': 'Percentile', + 'benchmark.detail.modelName': 'Model Name', + 'benchmark.detail.instanceName': 'Instance Name', + 'benchmark.detail.configure': 'Configuration', + 'benchmark.detail.modelFile': 'Model File', + 'benchmark.detail.kvCache': 'Extended KV Cache', + 'benchmark.detail.speculativeDecoding': 'Speculative Decoding', + 'benchmark.detail.inputOutputTokenLength': 'Token Length (Input/Output)', + 'benchmark.env.gpuName': 'GPU Name', + 'benchmark.env.workerName': 'Worker Name', + 'benchmark.env.index': 'Index', + 'benchmark.env.system': 'System', + 'benchmark.env.runtimeVersion': 'Runtime Version', + 'benchmark.env.driverVersion': 'Driver Version', + 'benchmark.env.cpuCounts': 'CPU Counts' }; diff --git a/src/locales/en-US/resources.ts b/src/locales/en-US/resources.ts index cf21cf56..2eb633e7 100644 --- a/src/locales/en-US/resources.ts +++ b/src/locales/en-US/resources.ts @@ -30,7 +30,7 @@ export default { 'resources.table.vram': 'VRAM', 'resources.table.index': 'Index', 'resources.table.workername': 'Worker Name', - 'resources.table.vender': 'Vendor', + 'resources.table.vendor': 'Vendor', 'resources.table.temperature': 'Temperature', 'resources.table.core': 'Cores', 'resources.table.utilization': 'Utilization', diff --git a/src/locales/ja-JP/benchmark.ts b/src/locales/ja-JP/benchmark.ts index f515d968..73732095 100644 --- a/src/locales/ja-JP/benchmark.ts +++ b/src/locales/ja-JP/benchmark.ts @@ -1,6 +1,7 @@ export default { 'benchmark.title': 'Benchmarks', 'benchmark.button.add': 'Add Benchmark', + 'benchmark.button.edit': 'Edit Benchmark', 'benchmark.button.compare': 'Compare', 'benchmark.table.model': 'Model', 'benchmark.table.instance': 'Model Instance', @@ -26,5 +27,57 @@ export default { 'benchmark.detail.summary.title': 'Summary', 'benchmark.detail.configure.title': 'Configure', 'benchmark.detail.environment.title': 'Environment', - 'benchmark.detail.logs.title': 'Logs' + 'benchmark.detail.logs.title': 'Logs', + 'benchmark.form.totalRequests.tips': 'Total Requests', + 'benchmark.form.profile.latency.tips': + 'Single-request latency benchmark focusing on TTFT and tail latency. Suitable for API serving and chat scenarios.', + 'benchmark.form.profile.throughput.tips': + 'Max throughput with realistic prompt length. The most commonly used baseline for GPU and model comparison.', + 'benchmark.form.profile.longContext.tips': + 'Stress test for long-context handling. Evaluates KV cache behavior, memory usage, and backend stability.', + 'benchmark.form.profile.heavy.tips': + 'Decode-heavy generation benchmark. Measures sustained decoding speed and output token throughput.', + 'benchmark.table.filter.bygpu': 'Filter by GPU', + 'benchmark.table.filter.bymodel': 'Filter by Model', + 'benchmark.table.filter.bydataset': 'Filter by Dataset', + 'benchmark.table.avg': 'Average', + 'benchmark.table.columnSettings': 'Column Settings', + 'benchmark.detail.summary.results': 'Test Results', + 'benchmark.detail.summary.metrics': 'Test Metrics', + 'benchmark.detail.summary.latency': 'Latency', + 'benchmark.detail.summary.throughput': 'Throughput', + 'benchmark.detail.summary.request': 'Requests', + 'benchmark.detail.avg.reqLatency': 'Average Request Latency', + 'benchmark.detail.avg.ttft': 'Average TTFT', + 'benchmark.detail.avg.tpot': 'Average TPOT', + 'benchmark.detail.avg.itl': 'Average ITL', + 'benchmark.detail.throughput.totalToken': 'Total Token Throughput', + 'benchmark.detail.throughput.inputToken': 'Input Token Throughput', + 'benchmark.detail.throughput.outputToken': 'Output Token Throughput', + 'benchmark.detail.requests.total': 'Total Requests', + 'benchmark.detail.requests.success': 'Successful Requests', + 'benchmark.detail.requests.failed': 'Failed Requests', + 'benchmark.detail.requests.concurrency': 'Concurrency', + 'benchmark.detail.percentile.latency': 'Latency', + 'benchmark.detail.percentile.inputTokens': 'Input Tokens', + 'benchmark.detail.percentile.outputTokens': 'Output Tokens', + 'benchmark.detail.percentile.input': 'Input', + 'benchmark.detail.percentile.output': 'Output', + 'benchmark.detail.percentile.total': 'Total', + 'benchmark.detail.configure.tokenLength': 'Token Length (Input/Output)', + 'benchmark.detail.percentile.title': 'Percentile', + 'benchmark.detail.modelName': 'Model Name', + 'benchmark.detail.instanceName': 'Instance Name', + 'benchmark.detail.configure': 'Configuration', + 'benchmark.detail.modelFile': 'Model File', + 'benchmark.detail.kvCache': 'Extended KV Cache', + 'benchmark.detail.speculativeDecoding': 'Speculative Decoding', + 'benchmark.detail.inputOutputTokenLength': 'Token Length (Input/Output)', + 'benchmark.env.gpuName': 'GPU Name', + 'benchmark.env.workerName': 'Worker Name', + 'benchmark.env.index': 'Index', + 'benchmark.env.system': 'System', + 'benchmark.env.runtimeVersion': 'Runtime Version', + 'benchmark.env.driverVersion': 'Driver Version', + 'benchmark.env.cpuCounts': 'CPU Counts' }; diff --git a/src/locales/ja-JP/resources.ts b/src/locales/ja-JP/resources.ts index 0abb26bc..3020e30c 100644 --- a/src/locales/ja-JP/resources.ts +++ b/src/locales/ja-JP/resources.ts @@ -29,7 +29,7 @@ export default { 'resources.table.vram': 'VRAM', 'resources.table.index': 'インデックス', 'resources.table.workername': 'ワーカー名', - 'resources.table.vender': 'ベンダー', + 'resources.table.vendor': 'ベンダー', 'resources.table.temperature': '温度', 'resources.table.core': 'コア数', 'resources.table.utilization': '利用率', diff --git a/src/locales/ru-RU/benchmark.ts b/src/locales/ru-RU/benchmark.ts index f515d968..73732095 100644 --- a/src/locales/ru-RU/benchmark.ts +++ b/src/locales/ru-RU/benchmark.ts @@ -1,6 +1,7 @@ export default { 'benchmark.title': 'Benchmarks', 'benchmark.button.add': 'Add Benchmark', + 'benchmark.button.edit': 'Edit Benchmark', 'benchmark.button.compare': 'Compare', 'benchmark.table.model': 'Model', 'benchmark.table.instance': 'Model Instance', @@ -26,5 +27,57 @@ export default { 'benchmark.detail.summary.title': 'Summary', 'benchmark.detail.configure.title': 'Configure', 'benchmark.detail.environment.title': 'Environment', - 'benchmark.detail.logs.title': 'Logs' + 'benchmark.detail.logs.title': 'Logs', + 'benchmark.form.totalRequests.tips': 'Total Requests', + 'benchmark.form.profile.latency.tips': + 'Single-request latency benchmark focusing on TTFT and tail latency. Suitable for API serving and chat scenarios.', + 'benchmark.form.profile.throughput.tips': + 'Max throughput with realistic prompt length. The most commonly used baseline for GPU and model comparison.', + 'benchmark.form.profile.longContext.tips': + 'Stress test for long-context handling. Evaluates KV cache behavior, memory usage, and backend stability.', + 'benchmark.form.profile.heavy.tips': + 'Decode-heavy generation benchmark. Measures sustained decoding speed and output token throughput.', + 'benchmark.table.filter.bygpu': 'Filter by GPU', + 'benchmark.table.filter.bymodel': 'Filter by Model', + 'benchmark.table.filter.bydataset': 'Filter by Dataset', + 'benchmark.table.avg': 'Average', + 'benchmark.table.columnSettings': 'Column Settings', + 'benchmark.detail.summary.results': 'Test Results', + 'benchmark.detail.summary.metrics': 'Test Metrics', + 'benchmark.detail.summary.latency': 'Latency', + 'benchmark.detail.summary.throughput': 'Throughput', + 'benchmark.detail.summary.request': 'Requests', + 'benchmark.detail.avg.reqLatency': 'Average Request Latency', + 'benchmark.detail.avg.ttft': 'Average TTFT', + 'benchmark.detail.avg.tpot': 'Average TPOT', + 'benchmark.detail.avg.itl': 'Average ITL', + 'benchmark.detail.throughput.totalToken': 'Total Token Throughput', + 'benchmark.detail.throughput.inputToken': 'Input Token Throughput', + 'benchmark.detail.throughput.outputToken': 'Output Token Throughput', + 'benchmark.detail.requests.total': 'Total Requests', + 'benchmark.detail.requests.success': 'Successful Requests', + 'benchmark.detail.requests.failed': 'Failed Requests', + 'benchmark.detail.requests.concurrency': 'Concurrency', + 'benchmark.detail.percentile.latency': 'Latency', + 'benchmark.detail.percentile.inputTokens': 'Input Tokens', + 'benchmark.detail.percentile.outputTokens': 'Output Tokens', + 'benchmark.detail.percentile.input': 'Input', + 'benchmark.detail.percentile.output': 'Output', + 'benchmark.detail.percentile.total': 'Total', + 'benchmark.detail.configure.tokenLength': 'Token Length (Input/Output)', + 'benchmark.detail.percentile.title': 'Percentile', + 'benchmark.detail.modelName': 'Model Name', + 'benchmark.detail.instanceName': 'Instance Name', + 'benchmark.detail.configure': 'Configuration', + 'benchmark.detail.modelFile': 'Model File', + 'benchmark.detail.kvCache': 'Extended KV Cache', + 'benchmark.detail.speculativeDecoding': 'Speculative Decoding', + 'benchmark.detail.inputOutputTokenLength': 'Token Length (Input/Output)', + 'benchmark.env.gpuName': 'GPU Name', + 'benchmark.env.workerName': 'Worker Name', + 'benchmark.env.index': 'Index', + 'benchmark.env.system': 'System', + 'benchmark.env.runtimeVersion': 'Runtime Version', + 'benchmark.env.driverVersion': 'Driver Version', + 'benchmark.env.cpuCounts': 'CPU Counts' }; diff --git a/src/locales/ru-RU/resources.ts b/src/locales/ru-RU/resources.ts index 253928a7..2a10de85 100644 --- a/src/locales/ru-RU/resources.ts +++ b/src/locales/ru-RU/resources.ts @@ -29,7 +29,7 @@ export default { 'resources.table.vram': 'VRAM', 'resources.table.index': 'Индекс', 'resources.table.workername': 'Имя воркера', - 'resources.table.vender': 'Производитель', + 'resources.table.vendor': 'Производитель', 'resources.table.temperature': 'Температура', 'resources.table.core': 'Ядра', 'resources.table.utilization': 'Использование', diff --git a/src/locales/zh-CN/benchmark.ts b/src/locales/zh-CN/benchmark.ts index 5860e71d..77c149fa 100644 --- a/src/locales/zh-CN/benchmark.ts +++ b/src/locales/zh-CN/benchmark.ts @@ -1,6 +1,7 @@ export default { 'benchmark.title': '基准测试', 'benchmark.button.add': '添加基准测试', + 'benchmark.button.edit': '编辑基准测试', 'benchmark.button.compare': '对比', 'benchmark.table.model': '模型', 'benchmark.table.instance': '模型实例', @@ -16,6 +17,7 @@ export default { 'benchmark.table.outputLen': '输出长度', 'benchmark.form.profile': '模式', 'benchmark.form.totalRequests': '总请求数', + 'benchmark.form.totalRequests.tips': '总请求数', 'benchmark.form.profile.latency': '延迟', 'benchmark.form.profile.throughput': '吞吐', 'benchmark.form.profile.longContext': '长上下文', @@ -26,5 +28,56 @@ export default { 'benchmark.detail.summary.title': '摘要', 'benchmark.detail.configure.title': '配置', 'benchmark.detail.environment.title': '环境', - 'benchmark.detail.logs.title': '日志' + 'benchmark.detail.logs.title': '日志', + 'benchmark.form.profile.latency.tips': + '单请求延迟测试，关注 TTFT 与尾延迟。适用于 API 服务和聊天场景。', + 'benchmark.form.profile.throughput.tips': + '在真实提示长度下测试最大吞吐量。常用于 GPU 与模型性能对比的基准。', + 'benchmark.form.profile.longContext.tips': + '长上下文压力测试。评估 KV Cache 行为、内存占用及后端稳定性。', + 'benchmark.form.profile.heavy.tips': + '以解码为主的生成测试。衡量持续解码速度与输出 Token 吞吐能力。', + 'benchmark.table.filter.bygpu': '按 GPU 过滤', + 'benchmark.table.filter.bymodel': '按模型过滤', + 'benchmark.table.filter.bydataset': '按数据集过滤', + 'benchmark.table.avg': '均值', + 'benchmark.table.columnSettings': '列设置', + 'benchmark.detail.summary.results': '测试结果', + 'benchmark.detail.summary.metrics': '测试指标', + 'benchmark.detail.summary.latency': '延迟', + 'benchmark.detail.summary.throughput': '吞吐', + 'benchmark.detail.summary.request': '请求数', + 'benchmark.detail.avg.reqLatency': '平均请求延迟', + 'benchmark.detail.avg.ttft': '平均 TTFT', + 'benchmark.detail.avg.tpot': '平均 TPOT', + 'benchmark.detail.avg.itl': '平均 ITL', + 'benchmark.detail.throughput.totalToken': '总 Token 吞吐量', + 'benchmark.detail.throughput.inputToken': '输入 Token 吞吐量', + 'benchmark.detail.throughput.outputToken': '输出 Token 吞吐量', + 'benchmark.detail.requests.total': '请求总数', + 'benchmark.detail.requests.success': '请求成功数', + 'benchmark.detail.requests.failed': '请求失败数', + 'benchmark.detail.requests.concurrency': '请求并发数', + 'benchmark.detail.percentile.latency': '延迟', + 'benchmark.detail.percentile.inputTokens': '输入 Token', + 'benchmark.detail.percentile.outputTokens': '输出 Token', + 'benchmark.detail.percentile.input': '输入', + 'benchmark.detail.percentile.output': '输出', + 'benchmark.detail.percentile.total': '总量', + 'benchmark.detail.configure.tokenLength': 'Token 长度 (输入/输出)', + 'benchmark.detail.percentile.title': '百分位数', + 'benchmark.detail.modelName': '模型名称', + 'benchmark.detail.instanceName': '实例名称', + 'benchmark.detail.configure': '测试配置', + 'benchmark.detail.modelFile': '模型文件', + 'benchmark.detail.kvCache': '扩展 KV 缓存', + 'benchmark.detail.speculativeDecoding': '推测解码', + 'benchmark.detail.inputOutputTokenLength': 'Token 长度 (输入/输出)', + 'benchmark.env.gpuName': 'GPU 名称', + 'benchmark.env.workerName': '节点名称', + 'benchmark.env.index': '序号', + 'benchmark.env.system': '系统', + 'benchmark.env.runtimeVersion': '运行时版本', + 'benchmark.env.driverVersion': '驱动版本', + 'benchmark.env.cpuCounts': 'CPU 核心数' }; diff --git a/src/locales/zh-CN/resources.ts b/src/locales/zh-CN/resources.ts index 777a0894..59425b90 100644 --- a/src/locales/zh-CN/resources.ts +++ b/src/locales/zh-CN/resources.ts @@ -30,7 +30,7 @@ export default { 'resources.table.vram': '显存', 'resources.table.index': '序号', 'resources.table.workername': '节点名称', - 'resources.table.vender': '厂商', + 'resources.table.vendor': '厂商', 'resources.table.temperature': '温度', 'resources.table.core': '核数', 'resources.table.gpuutilization': 'GPU 利用率', diff --git a/src/pages/benchmark/components/environment/index.tsx b/src/pages/benchmark/components/environment/index.tsx index c10886f8..bdedf762 100644 --- a/src/pages/benchmark/components/environment/index.tsx +++ b/src/pages/benchmark/components/environment/index.tsx @@ -3,6 +3,7 @@ import RowChildren from '@/components/seal-table/components/row-children'; import SealTable from '@/components/seal-table/index'; import useExpandedRowKeys from '@/hooks/use-expanded-row-keys'; import { convertFileSize } from '@/utils'; +import { useIntl } from '@umijs/max'; import useMemoizedFn from 'ahooks/lib/useMemoizedFn'; import { Col, Row, Tag } from 'antd'; import _ from 'lodash'; @@ -17,6 +18,7 @@ const Container = styled.div` `; const Environment: React.FC = () => { + const intl = useIntl(); const { detailData } = useDetailContext(); const { snapshot } = detailData; @@ -57,9 +59,16 @@ const Environment: React.FC = () => { }); }, [snapshot, mainWorker]); - const GPUColumns = [ + const GPUColumns: { + title: string; + dataIndex: string; + key: string; + span: number; + colStyle?: React.CSSProperties; + render?: (value: any, record: any) => React.ReactNode; + }[] = [ { - title: 'GPU Name', + title: intl.formatMessage({ id: 'benchmark.env.gpuName' }), dataIndex: 'name', key: 'name', span: 6, @@ -69,32 +78,30 @@ const Environment: React.FC = () => { ) }, { - title: 'Index', + title: intl.formatMessage({ id: 'benchmark.env.index' }), dataIndex: 'index', key: 'index', span: 4, colStyle: { paddingLeft: 48 } }, { - title: 'Vendor', + title: intl.formatMessage({ id: 'resources.table.vendor' }), dataIndex: 'vendor', key: 'vendor', span: 6, colStyle: { paddingLeft: 110 } }, { - title: 'VRAM', + title: intl.formatMessage({ id: 'resources.table.vram' }), dataIndex: 'memory_total', key: 'memory_total', - label: 'VRAM', span: 4, render: (value: number, record: any) => convertFileSize(value) }, { - title: 'Cores', + title: intl.formatMessage({ id: 'resources.table.core' }), dataIndex: 'core_total', key: 'core_total', - label: 'Cores', span: 4, colStyle: { paddingLeft: 36 } } @@ -102,7 +109,7 @@ const Environment: React.FC = () => { const columns = [ { - title: 'Worker Name', + title: intl.formatMessage({ id: 'benchmark.env.workerName' }), dataIndex: 'name', key: 'name', span: 6, @@ -120,7 +127,7 @@ const Environment: React.FC = () => { } }, { - title: 'System', + title: intl.formatMessage({ id: 'benchmark.env.system' }), dataIndex: 'os', key: 'system', span: 5, @@ -133,7 +140,7 @@ const Environment: React.FC = () => { } }, { - title: 'Runtime Version', + title: intl.formatMessage({ id: 'benchmark.env.runtimeVersion' }), dataIndex: 'runtime_version', key: 'runtime_version', span: 3, @@ -142,7 +149,7 @@ const Environment: React.FC = () => { } }, { - title: 'Driver Version', + title: intl.formatMessage({ id: 'benchmark.env.driverVersion' }), dataIndex: 'driver_version', key: 'driver_version', span: 3, @@ -151,13 +158,13 @@ const Environment: React.FC = () => { } }, { - title: 'CPU Count', + title: intl.formatMessage({ id: 'benchmark.env.cpuCounts' }), dataIndex: 'cpu_total', key: 'cpu_total', span: 3 }, { - title: 'Memory', + title: intl.formatMessage({ id: 'resources.table.memory' }), dataIndex: 'memory_total', key: 'memory_total', span: 4, @@ -218,8 +225,8 @@ const Environment: React.FC = () => { }} > {col.render - ? col.render((gpu as any)[col.dataIndex], gpu) - : (gpu as any)[col.dataIndex]} + ? col.render(gpu[col.dataIndex], gpu) + : gpu[col.dataIndex]} ))} diff --git a/src/pages/benchmark/components/left-actions.tsx b/src/pages/benchmark/components/left-actions.tsx index 348aeea6..75b83a92 100644 --- a/src/pages/benchmark/components/left-actions.tsx +++ b/src/pages/benchmark/components/left-actions.tsx @@ -52,14 +52,16 @@ const RightActions: React.FC = ({ style={{ color: 'var(--ant-color-text-placeholder)' }} > } - placeholder="Filter by GPU" + placeholder={intl.formatMessage({ id: 'benchmark.table.filter.bygpu' })} style={{ width: 180 }} allowClear onChange={handleGPUChange} > @@ -71,7 +73,9 @@ const RightActions: React.FC = ({ > ({ ...item, diff --git a/src/pages/benchmark/components/right-actions.tsx b/src/pages/benchmark/components/right-actions.tsx index d73a0574..8d54d37f 100644 --- a/src/pages/benchmark/components/right-actions.tsx +++ b/src/pages/benchmark/components/right-actions.tsx @@ -4,14 +4,12 @@ import { DownloadOutlined, PlusOutlined } from '@ant-design/icons'; -import { useIntl } from '@umijs/max'; import { Button, Space } from 'antd'; import React from 'react'; export interface RightActionsProps { handleDeleteByBatch: () => void; handleClickPrimary?: () => void; - handleSettingFields?: () => void; handleExport?: () => void; settingButton?: React.ReactNode; buttonText?: string; @@ -23,13 +21,11 @@ export interface RightActionsProps { const RightActions: React.FC = ({ handleDeleteByBatch, handleClickPrimary, - handleSettingFields, handleExport, settingButton, buttonText, rowSelection }) => { - const intl = useIntl(); const ButtonList = [ { label: 'common.button.export', @@ -56,13 +52,7 @@ const RightActions: React.FC = ({ return ( - {/* - - */} - {/* {settingButton} */} + {settingButton} - {/* */} { + const intl = useIntl(); const { detailData } = useDetailContext(); const items: DescriptionsProps['items'] = [ { key: '1', - label: 'Profile', + label: intl.formatMessage({ id: 'benchmark.form.profile' }), children: detailData?.profile || '-' }, { key: '2', - label: 'Dataset', + label: intl.formatMessage({ id: 'benchmark.table.dataset' }), children: detailData?.dataset_name || '-' }, { key: '3', - label: 'Token Length (Input/Output)', + label: intl.formatMessage({ + id: 'benchmark.detail.inputOutputTokenLength' + }), children: ( {detailData?.dataset_input_tokens || '-'} /{' '} @@ -28,24 +32,23 @@ const Benchmark: React.FC = () => { }, { key: '7', - label: 'Total Requests', + label: intl.formatMessage({ id: 'benchmark.form.totalRequests' }), children: detailData?.total_requests || '-' }, { key: '6', - label: 'Request Rate', + label: intl.formatMessage({ id: 'benchmark.table.requestRate' }), children: detailData?.request_rate || '-' }, { key: '5', - label: 'Seed', + label: intl.formatMessage({ id: 'playground.image.params.seed' }), children: detailData?.seed || '-' } ]; return (
- Parameters { + const intl = useIntl(); return ( -
+
-
+
diff --git a/src/pages/benchmark/components/summary/instance.tsx b/src/pages/benchmark/components/summary/instance.tsx index 9de05cc5..401d94b9 100644 --- a/src/pages/benchmark/components/summary/instance.tsx +++ b/src/pages/benchmark/components/summary/instance.tsx @@ -1,4 +1,5 @@ import AutoTooltip from '@/components/auto-tooltip'; +import { useIntl } from '@umijs/max'; import { Descriptions, Flex, Tag } from 'antd'; import _ from 'lodash'; import React, { useMemo } from 'react'; @@ -9,6 +10,7 @@ const calcTotalVram = (vram: Record) => { }; const Instance: React.FC = () => { + const intl = useIntl(); const { detailData } = useDetailContext(); const [, instanceData] = Object.entries(detailData?.snapshot?.instances || {})[0] || []; @@ -20,14 +22,14 @@ const Instance: React.FC = () => { return [ { key: '1', - label: 'Model Name', + label: intl.formatMessage({ id: 'benchmark.detail.modelName' }), children: ( {detailData?.model_name || '-'} ) }, { key: '2', - label: 'Instance Name', + label: intl.formatMessage({ id: 'benchmark.detail.instanceName' }), children: ( {detailData?.model_instance_name || '-'} @@ -37,7 +39,7 @@ const Instance: React.FC = () => { { key: '5', - label: 'Backend', + label: intl.formatMessage({ id: 'models.form.backend' }), children: `${instanceData?.backend || '-'} ${ instanceData?.backend_version ? `(${instanceData?.backend_version})` @@ -46,7 +48,7 @@ const Instance: React.FC = () => { }, { key: '6', - label: 'Model File', + label: intl.formatMessage({ id: 'benchmark.detail.modelFile' }), children: ( {instanceData?.resolved_path || '-'} ) @@ -61,7 +63,7 @@ const Instance: React.FC = () => { return [ { key: '1', - label: 'Backend Parameters', + label: intl.formatMessage({ id: 'models.form.backend_parameters' }), children: ( { }, { key: '3', - label: 'Extended KV Cache', + label: intl.formatMessage({ id: 'benchmark.detail.kvCache' }), children: ( {instanceData?.extended_kv_cache?.enabled ? ( <> - RAM-to-VRAM Ratio: + + {intl.formatMessage({ id: 'models.form.ramRatio' })}: + {instanceData?.extended_kv_cache?.ram_ratio} - Maximum RAM Size (GiB): + + {intl.formatMessage({ id: 'models.form.ramSize' })}: + {instanceData?.extended_kv_cache?.ram_size} - Size of Cache Chunks: + + {intl.formatMessage({ id: 'models.form.chunkSize' })}: + {instanceData?.extended_kv_cache?.chunk_size} @@ -110,21 +118,29 @@ const Instance: React.FC = () => { }, { key: '4', - label: 'Speculative Decoding', + label: intl.formatMessage({ + id: 'benchmark.detail.speculativeDecoding' + }), children: ( {instanceData?.speculative_config?.enabled ? ( <> - Algorithm: + + {intl.formatMessage({ id: 'models.form.algorithm' })}: + {instanceData?.speculative_config?.algorithm} - Draft Model: + + {intl.formatMessage({ id: 'models.form.draftModel' })}: + {instanceData?.speculative_config?.draft_model} - Number of Draft Tokens: + + {intl.formatMessage({ id: 'models.form.numDraftTokens' })}: + {instanceData?.speculative_config?.num_draft_tokens} @@ -138,7 +154,7 @@ const Instance: React.FC = () => { }, { key: '2', - label: 'Environment Variables', + label: intl.formatMessage({ id: 'models.form.env' }), children: ( {instanceData?.env @@ -158,7 +174,6 @@ const Instance: React.FC = () => { return (
- Instance round(value, 2) - // }, - - // { - // title: 'Total Requests', - // dataIndex: 'total_requests', - // path: 'total_requests', - // unit: '', - // render: (value: number) => round(value, 0) - // }, +const throughputColumns = [ { - title: 'Total token throughput', + title: 'benchmark.detail.throughput.totalToken', dataIndex: 'tokens_per_second_mean', path: 'tokens_per_second_mean', unit: 'Tokens/s', render: (value: number) => round(value, 2) }, { - title: 'Input token throughput ', + title: 'benchmark.detail.throughput.inputToken', dataIndex: 'input_tokens_per_second_mean', path: 'input_tokens_per_second_mean', unit: 'Tokens/s', render: (value: number) => round(value, 2) }, - // { - // title: 'Success Requests', - // dataIndex: 'successful_requests', - // path: ['raw_metrics', 'benchmarks', '0'], - // unit: '', - // render: (value: number) => { - // return ( - // - // {_.get(value, 'metrics.request_totals.successful')} - // - // ); - // } - // }, - { - title: 'Output token throughput', + title: 'benchmark.detail.throughput.outputToken', dataIndex: 'output_tokens_per_second_mean', path: 'output_tokens_per_second_mean', unit: 'Tokens/s', @@ -77,56 +35,30 @@ const columns = [ } ]; -const columnsSub = [ - // { - // title: 'Failed Requests', - // dataIndex: 'failed_requests', - // path: ['raw_metrics', 'benchmarks', '0'], - // unit: '', - // render: (value: number) => { - // return ( - // - // {_.get(value, 'metrics.request_totals.errored')} - // - // ); - // } - // }, - // { - // title: 'Concurrency', - // dataIndex: 'request_concurrency', - // path: ['raw_metrics', 'benchmarks', '0'], - // unit: '', - // render: (value: number) => { - // return round( - // _.get(value, 'metrics.request_concurrency.successful.mean'), - // 0 - // ); - // } - // }, - +const latencyColumns = [ { - title: 'Average Request Latency', + title: 'benchmark.detail.avg.reqLatency', dataIndex: 'request_latency_mean', path: 'request_latency_mean', unit: 'ms', render: (value: number) => round(value, 2) }, { - title: 'Average Time To First Token', + title: 'benchmark.detail.avg.ttft', dataIndex: 'time_to_first_token_mean', path: 'time_to_first_token_mean', unit: 'ms', render: (value: number) => round(value, 2) }, { - title: 'Average Time Per Output Token', + title: 'benchmark.detail.avg.tpot', dataIndex: 'time_per_output_token_mean', path: 'time_per_output_token_mean', unit: 'ms', render: (value: number) => round(value, 2) }, { - title: 'Average Inter Token Latency', + title: 'benchmark.detail.avg.itl', dataIndex: 'inter_token_latency_mean', path: 'inter_token_latency_mean', unit: 'ms', @@ -134,9 +66,9 @@ const columnsSub = [ } ]; -const requestFields = [ +const requestColumns = [ { - label: 'Total Requests', + label: 'benchmark.detail.requests.total', key: 'total_requests', dataIndex: 'total_requests', path: 'total_requests', @@ -145,7 +77,7 @@ const requestFields = [ unit: '' }, { - label: 'Success Requests', + label: 'benchmark.detail.requests.success', key: 'total_requests', dataIndex: 'successful_requests', path: ['raw_metrics', 'benchmarks', '0'], @@ -156,7 +88,7 @@ const requestFields = [ unit: '' }, { - label: 'Failed Requests', + label: 'benchmark.detail.requests.failed', key: 'total_requests', dataIndex: 'failed_requests', path: ['raw_metrics', 'benchmarks', '0'], @@ -167,7 +99,7 @@ const requestFields = [ unit: '' }, { - label: 'Concurrency', + label: 'benchmark.detail.requests.concurrency', key: 'request_concurrency', dataIndex: 'request_concurrency', path: ['raw_metrics', 'benchmarks', '0'], @@ -185,21 +117,24 @@ const PERCENTILES = [ const PercentileResult: React.FC = () => { const { detailData } = useDetailContext(); + const intl = useIntl(); const metrics = detailData?.raw_metrics?.benchmarks?.[0]?.metrics || {}; const buildPercentileTable = (metrics: any) => { return PERCENTILES.map(({ key, title }) => { const row: any = { metrics: title }; - [...columns, ...columnsSub].forEach(({ dataIndex, path }) => { - row[dataIndex] = _.get(detailData, path) ?? 0; - }); + [...throughputColumns, ...latencyColumns].forEach( + ({ dataIndex, path }) => { + row[dataIndex] = _.get(detailData, path) ?? 0; + } + ); return row; }); }; - const items = [...columns, ...columnsSub].map( + const items = [...throughputColumns, ...latencyColumns].map( ({ title, dataIndex, path, render, unit }) => ({ key: dataIndex, label: title, @@ -214,10 +149,10 @@ const PercentileResult: React.FC = () => { }) ); - const throughputItems = columns.map( + const throughputItems = throughputColumns.map( ({ title, dataIndex, path, render, unit }) => ({ key: dataIndex, - label: title, + label: intl.formatMessage({ id: title }), children: unit ? ( {render(_.get(detailData, path) ?? 0)}{' '} @@ -229,10 +164,10 @@ const PercentileResult: React.FC = () => { }) ); - const latencyItems = columnsSub.map( + const latencyItems = latencyColumns.map( ({ title, dataIndex, path, render, unit }) => ({ key: dataIndex, - label: title, + label: intl.formatMessage({ id: title }), children: unit ? ( {render(_.get(detailData, path) ?? 0)}{' '} @@ -244,10 +179,10 @@ const PercentileResult: React.FC = () => { }) ); - const requestItems = requestFields.map( + const requestItems = requestColumns.map( ({ label, dataIndex, path, render, unit, color }) => ({ key: dataIndex, - label: label, + label: intl.formatMessage({ id: label }), children: unit ? ( {render(_.get(detailData, path) ?? 0)}{' '} @@ -277,25 +212,26 @@ const PercentileResult: React.FC = () => { return (
- Metrics round(value, 2) - }, - { - title: 'ITL (ms)', - dataIndex: 'inter_token_latency_ms', - render: (value: number) => round(value, 2) - }, - { - title: 'TPOT (ms)', - dataIndex: 'time_per_output_token_ms', - render: (value: number) => round(value, 2) - }, - { - title: 'Latency (s)', - dataIndex: 'request_latency', - render: (value: number) => round(value, 2) - }, - { - title: 'Input tokens', - dataIndex: 'prompt_token_count', - render: (value: number) => round(value, 0) - }, - { - title: 'Output tokens', - dataIndex: 'output_token_count', - render: (value: number) => round(value, 0) - }, - { - title: 'Input (Tokens/s)', - dataIndex: 'prompt_tokens_per_second', - render: (value: number) => round(value, 2) - }, - { - title: 'Output (Tokens/s)', - dataIndex: 'output_tokens_per_second', - render: (value: number) => round(value, 2) - }, - { - title: 'Total (Tokens/s)', - dataIndex: 'tokens_per_second', - render: (value: number) => round(value, 2) - } -]; - const PERCENTILES = [ { key: 'p50', label: '50%' }, { key: 'p90', label: '90%' }, @@ -58,9 +11,62 @@ const PERCENTILES = [ ] as const; const PercentileResult: React.FC = () => { + const intl = useIntl(); const { detailData } = useDetailContext(); const metrics = detailData?.raw_metrics?.benchmarks?.[0]?.metrics || {}; + const columns = [ + { + title: 'TTFT (ms)', + dataIndex: 'time_to_first_token_ms', + render: (value: number) => round(value, 2) + }, + { + title: 'ITL (ms)', + dataIndex: 'inter_token_latency_ms', + render: (value: number) => round(value, 2) + }, + { + title: 'TPOT (ms)', + dataIndex: 'time_per_output_token_ms', + render: (value: number) => round(value, 2) + }, + { + title: `${intl.formatMessage({ id: 'benchmark.detail.percentile.latency' })} (s)`, + dataIndex: 'request_latency', + render: (value: number) => round(value, 2) + }, + { + title: intl.formatMessage({ + id: 'benchmark.detail.percentile.inputTokens' + }), + dataIndex: 'prompt_token_count', + render: (value: number) => round(value, 0) + }, + { + title: intl.formatMessage({ + id: 'benchmark.detail.percentile.outputTokens' + }), + dataIndex: 'output_token_count', + render: (value: number) => round(value, 0) + }, + { + title: `${intl.formatMessage({ id: 'benchmark.detail.percentile.input' })} (Tokens/s)`, + dataIndex: 'prompt_tokens_per_second', + render: (value: number) => round(value, 2) + }, + { + title: `${intl.formatMessage({ id: 'benchmark.detail.percentile.output' })} (Tokens/s)`, + dataIndex: 'output_tokens_per_second', + render: (value: number) => round(value, 2) + }, + { + title: `${intl.formatMessage({ id: 'benchmark.detail.percentile.total' })} (Tokens/s)`, + dataIndex: 'tokens_per_second', + render: (value: number) => round(value, 2) + } + ]; + const buildPercentileTable = (metrics: any) => { return PERCENTILES.map(({ key, label }) => { const row: any = { percentile: label }; @@ -76,12 +82,17 @@ const PercentileResult: React.FC = () => { return (
- Percentile Percentile, + title: ( + + {intl.formatMessage({ + id: 'benchmark.detail.percentile.title' + })} + + ), dataIndex: 'percentile', render: (value: string) => ( {value} diff --git a/src/pages/benchmark/config/index.ts b/src/pages/benchmark/config/index.ts index 677cae59..7a627d3e 100644 --- a/src/pages/benchmark/config/index.ts +++ b/src/pages/benchmark/config/index.ts @@ -32,10 +32,10 @@ export const BenchmarkStatus: Record = { }; export const ProfileValueMap = { - LatencyShort: 'latency_short', - ThroughputMedium: 'throughput_medium', - LongContextStress: 'long_context_stress', - GenerationHeavy: 'generation_heavy', + LatencyShort: 'Latency', + ThroughputMedium: 'Throughput', + LongContextStress: 'Long Context', + GenerationHeavy: 'Generation Heavy', Custom: 'Custom' }; @@ -50,28 +50,27 @@ export const ProfileLabelMap = { export const profileOptions = [ { label: 'benchmark.form.profile.latency', + tips: 'benchmark.form.profile.latency.tips', value: ProfileValueMap.LatencyShort, locale: true }, { label: 'benchmark.form.profile.throughput', + tips: 'benchmark.form.profile.throughput.tips', value: ProfileValueMap.ThroughputMedium, locale: true }, { label: 'benchmark.form.profile.longContext', + tips: 'benchmark.form.profile.longContext.tips', value: ProfileValueMap.LongContextStress, locale: true }, { label: 'benchmark.form.profile.heavy', + tips: 'benchmark.form.profile.heavy.tips', value: ProfileValueMap.GenerationHeavy, locale: true - }, - { - label: 'benchmark.form.profile.custom', - value: ProfileValueMap.Custom, - locale: true } ]; diff --git a/src/pages/benchmark/forms/dataset.tsx b/src/pages/benchmark/forms/dataset.tsx index 11899b01..78bb4d09 100644 --- a/src/pages/benchmark/forms/dataset.tsx +++ b/src/pages/benchmark/forms/dataset.tsx @@ -1,8 +1,9 @@ +import AutoTooltip from '@/components/auto-tooltip'; import SealSelect from '@/components/seal-form/seal-select'; import { PageAction } from '@/config'; import useAppUtils from '@/hooks/use-app-utils'; import { useIntl } from '@umijs/max'; -import { Form } from 'antd'; +import { Form, Select } from 'antd'; import _ from 'lodash'; import React, { useEffect } from 'react'; import { ProfileValueMap } from '../config'; @@ -68,10 +69,25 @@ const DatasetForm: React.FC = () => { + > + {profilesOptions?.map((item: any) => ( + + + {intl.formatMessage({ id: item?.label || '' })} + + + ))} + { }, []); const handleAddBenchmark = () => { - openBenchmarkModal(PageAction.CREATE, 'Add Benchmark'); + openBenchmarkModal( + PageAction.CREATE, + intl.formatMessage({ id: 'benchmark.button.add' }) + ); }; const handleModalOk = async (data: FormData) => { @@ -108,7 +111,11 @@ const Benchmark: React.FC = () => { }; const handleEditUser = (row: ListItem) => { - openBenchmarkModal(PageAction.EDIT, 'Edit Benchmark', row); + openBenchmarkModal( + PageAction.EDIT, + intl.formatMessage({ id: 'benchmark.button.edit' }), + row + ); }; const handleSelect = useMemoizedFn((val: any, row: ListItem) => { diff --git a/src/pages/benchmark/services/use-query-profiles.ts b/src/pages/benchmark/services/use-query-profiles.ts index dc81e869..442c410a 100644 --- a/src/pages/benchmark/services/use-query-profiles.ts +++ b/src/pages/benchmark/services/use-query-profiles.ts @@ -3,6 +3,7 @@ import { useIntl } from '@umijs/max'; import _ from 'lodash'; import { useState } from 'react'; import { queryProfiles } from '../apis'; +import { profileOptions } from '../config'; import { ProfileOption } from '../config/types'; export default function useQueryProfiles() { @@ -23,10 +24,16 @@ export default function useQueryProfiles() { const fetchProfilesData = async () => { const res = await fetchData({}); + const profileMap = profileOptions.reduce((map, obj) => { + map.set(obj.value, obj); + return map; + }, new Map()); + const list = res?.profiles?.map((item) => { return { - label: item.name, + label: profileMap.get(item.name)?.label || item.name, + tips: profileMap.get(item.name)?.tips || '', value: item.name, config: { ..._.omit(item, 'name') @@ -38,6 +45,7 @@ export default function useQueryProfiles() { ...list, { label: intl.formatMessage({ id: 'backend.custom' }), + tips: '', value: 'Custom', config: { dataset_name: '', diff --git a/src/pages/llmodels/components/catalog-item.tsx b/src/pages/llmodels/components/catalog-item.tsx index fd9a537e..452e4e33 100644 --- a/src/pages/llmodels/components/catalog-item.tsx +++ b/src/pages/llmodels/components/catalog-item.tsx @@ -81,8 +81,10 @@ const CatalogItem: React.FC = (props) => { > {data.release_date} @@ -90,7 +92,13 @@ const CatalogItem: React.FC = (props) => { {_.map(data.licenses, (license: string, index: number) => { return ( - + {license} ); diff --git a/src/pages/resources/hooks/use-gpu-columns.tsx b/src/pages/resources/hooks/use-gpu-columns.tsx index b8285a0b..a1eff26a 100644 --- a/src/pages/resources/hooks/use-gpu-columns.tsx +++ b/src/pages/resources/hooks/use-gpu-columns.tsx @@ -84,7 +84,7 @@ const useGPUColumns = (props: { ) }, { - title: intl.formatMessage({ id: 'resources.table.vender' }), + title: intl.formatMessage({ id: 'resources.table.vendor' }), dataIndex: 'vendor', sorter: tableSorter(5) },