chore: update models locales

src/components/auto-tooltip/index.tsx

@@ -121,9 +121,9 @@ const AutoTooltip: React.FC<AutoTooltipProps> = ({
           className={`${tagProps.className || ''} ${filled ? 'tag-filled' : ''}`}
           ref={contentRef}
           style={{
-            ...tagStyle,
             paddingInline: tagProps.closable ? '8px 22px' : 8,
-            borderRadius: radius
+            borderRadius: radius,
+            ...tagStyle
           }}
           closeIcon={
             tagProps.closable ? (

src/locales/en-US/models.ts

@@ -219,10 +219,19 @@ export default {
     'Ratio of system RAM to GPU VRAM used for KV cache. For example, 2.0 means the cache in RAM can be twice as large as the GPU VRAM.',
   'models.form.ramSize.tips': `Maximum size of the KV cache stored in system memory (GiB). If set, this value overrides "{content}".`,
   'models.form.chunkSize.tips': 'Number of tokens per KV cache chunk.',
-  'models.form.flavor.throughput.tips':
+  'models.form.mode': 'Mode',
+  'models.form.algorithm': 'Algorithm',
+  'models.form.draftModel': 'Draft Model',
+  'models.form.numDraftTokens': 'Number of Draft Tokens',
+  'models.form.ngramMinMatchLength': 'N-gram Minimum Match Length',
+  'models.form.ngramMaxMatchLength': 'N-gram Maximum Match Length',
+  'models.form.mode.throughput': 'Throughput',
+  'models.form.mode.latency': 'Latency',
+  'models.form.mode.reference': 'Reference',
+  'models.form.mode.throughput.tips':
     'optimized for high throughput under high request concurrency.',
-  'models.form.flavor.latency.tips':
+  'models.form.mode.latency.tips':
     'optimized for low latency under low request concurrency.',
-  'models.form.flavor.reference.tips':
+  'models.form.mode.reference.tips':
     'the most compatible option with full precision.'
 };

src/locales/ja-JP/models.ts

@@ -219,11 +219,20 @@ export default {
     'Ratio of system RAM to GPU VRAM used for KV cache. For example, 2.0 means the cache in RAM can be twice as large as the GPU VRAM.',
   'models.form.ramSize.tips': `Maximum size of the KV cache stored in system memory (GiB). If set, this value overrides "{content}".`,
   'models.form.chunkSize.tips': 'Number of tokens per KV cache chunk.',
-  'models.form.flavor.throughput.tips':
+  'models.form.mode': 'Mode',
+  'models.form.algorithm': 'Algorithm',
+  'models.form.draftModel': 'Draft Model',
+  'models.form.numDraftTokens': 'Number of Draft Tokens',
+  'models.form.ngramMinMatchLength': 'N-gram Minimum Match Length',
+  'models.form.ngramMaxMatchLength': 'N-gram Maximum Match Length',
+  'models.form.mode.throughput': 'Throughput',
+  'models.form.mode.latency': 'Latency',
+  'models.form.mode.reference': 'Reference',
+  'models.form.mode.throughput.tips':
     'optimized for high throughput under high request concurrency.',
-  'models.form.flavor.latency.tips':
+  'models.form.mode.latency.tips':
     'optimized for low latency under low request concurrency.',
-  'models.form.flavor.reference.tips':
+  'models.form.mode.reference.tips':
     'the most compatible option with full precision.'
 };
 
@@ -277,4 +286,17 @@ export default {
 // 49. 'models.form.ramRatio.tips': 'Ratio of system RAM to GPU VRAM used for KV cache. For example, 2.0 means the cache in RAM can be twice as large as the GPU VRAM.',
 // 50. 'models.form.ramSize.tips': `Maximum size of the KV cache stored in system memory (GiB). If set, this value overrides "{content}".`,
 // 51. 'models.form.chunkSize.tips': 'Number of tokens per KV cache chunk.'
+// 33. 'models.form.mode': 'Mode',
+// 34. 'models.form.algorithm': 'Algorithm',
+// 35. 'models.form.draftModel': 'Draft Model',
+// 36. 'models.form.numDraftTokens': 'Number of Draft Tokens',
+// 37. 'models.form.ngramMinMatchLength': 'N-gram Minimum Match Length',
+// 38. 'models.form.ngramMaxMatchLength': 'N-gram Maximum Match Length',
+// 39. 'models.form.mode.throughput': 'Throughput',
+// 40. 'models.form.mode.latency': 'Latency',
+// 41. 'models.form.mode.reference': 'Reference',
+// 42. 'models.form.mode.throughput.tips': 'optimized for high throughput under high request concurrency.',
+// 43. 'models.form.mode.latency.tips': 'optimized for low latency under low request concurrency.',
+// 44. 'models.form.mode.reference.tips': 'the most compatible option with full precision
+
 // ========== End of To-Do List ==========

src/locales/ru-RU/models.ts

@@ -218,7 +218,22 @@ export default {
   'models.form.ramRatio.tips':
     'Ratio of system RAM to GPU VRAM used for KV cache. For example, 2.0 means the cache in RAM can be twice as large as the GPU VRAM.',
   'models.form.ramSize.tips': `Maximum size of the KV cache stored in system memory (GiB). If set, this value overrides "{content}".`,
-  'models.form.chunkSize.tips': 'Number of tokens per KV cache chunk.'
+  'models.form.chunkSize.tips': 'Number of tokens per KV cache chunk.',
+  'models.form.mode': 'Mode',
+  'models.form.algorithm': 'Algorithm',
+  'models.form.draftModel': 'Draft Model',
+  'models.form.numDraftTokens': 'Number of Draft Tokens',
+  'models.form.ngramMinMatchLength': 'N-gram Minimum Match Length',
+  'models.form.ngramMaxMatchLength': 'N-gram Maximum Match Length',
+  'models.form.mode.throughput': 'Throughput',
+  'models.form.mode.latency': 'Latency',
+  'models.form.mode.reference': 'Reference',
+  'models.form.mode.throughput.tips':
+    'optimized for high throughput under high request concurrency.',
+  'models.form.mode.latency.tips':
+    'optimized for low latency under low request concurrency.',
+  'models.form.mode.reference.tips':
+    'the most compatible option with full precision.'
 };
 
 // ========== To-Do: Translate Keys (Remove After Translation) ==========
@@ -234,4 +249,16 @@ export default {
 // 11. 'models.form.ramRatio.tips': 'Ratio of system RAM to GPU VRAM used for KV cache. For example, 2.0 means the cache in RAM can be twice as large as the GPU VRAM.',
 // 12. 'models.form.ramSize.tips': `Maximum size of the KV cache stored in system memory (GiB). If set, this value overrides "{content}".`,
 // 13. 'models.form.chunkSize.tips': 'Number of tokens per KV cache chunk.'
+// 25. 'models.form.mode': 'Mode',
+// 26. 'models.form.algorithm': 'Algorithm',
+// 27. 'models.form.draftModel': 'Draft Model',
+// 28. 'models.form.numDraftTokens': 'Number of Draft Tokens',
+// 29. 'models.form.ngramMinMatchLength': 'N-gram Minimum Match Length',
+// 30. 'models.form.ngramMaxMatchLength': 'N-gram Maximum Match Length',
+// 31. 'models.form.mode.throughput': 'Throughput',
+// 32. 'models.form.mode.latency': 'Latency',
+// 33. 'models.form.mode.reference': 'Reference',
+// 34. 'models.form.mode.throughput.tips': 'optimized for high throughput under high request concurrency.',
+// 35. 'models.form.mode.latency.tips': 'optimized for low latency under low request concurrency.',
+// 36. 'models.form.mode.reference.tips': 'the most compatible option with full precision.'
 // ========== End of To-Do List ==========

src/locales/zh-CN/models.ts

@@ -208,5 +208,17 @@ export default {
     'KV 缓存在系统内存与 GPU 显存之间的比例。例如设置为 2.0 表示系统内存中可缓存的数据量是显存的两倍。',
   'models.form.ramSize.tips': `KV 缓存在系统内存中的最大值。当设置该值时，将覆盖 "{content}" 的配置。`,
   'models.form.chunkSize.tips':
-    '每个 KV 缓存块包含的 token 数量。数值越大可提升吞吐量，但也会增加内存占用。'
+    '每个 KV 缓存块包含的 token 数量。数值越大可提升吞吐量，但也会增加内存占用。',
+  'models.form.mode': '模式',
+  'models.form.algorithm': '算法',
+  'models.form.draftModel': '草稿模型',
+  'models.form.numDraftTokens': '草稿生成 Token 数',
+  'models.form.ngramMinMatchLength': 'N-gram 最小匹配长度',
+  'models.form.ngramMaxMatchLength': 'N-gram 最大匹配长度',
+  'models.form.mode.throughput': '高吞吐',
+  'models.form.mode.latency': '低延迟',
+  'models.form.mode.reference': '参考',
+  'models.form.mode.throughput.tips': '在高并发请求下优化吞吐性能。',
+  'models.form.mode.latency.tips': '在低并发请求下优化响应延迟。',
+  'models.form.mode.reference.tips': '最兼容且保留完整精度的选项。'
 };

src/pages/llmodels/components/catalog-item.tsx

@@ -152,11 +152,15 @@ const CatalogItem: React.FC<CatalogItemProps> = (props) => {
               );
             })}
           <span className="dot"></span>
-          <ThemeTag>
+          <AutoTooltip
+            style={{
+              borderRadius: 4
+            }}
+          >
             {data.activated_size
               ? `${data.size}B-A${data.activated_size}B`
               : `${data.size}B`}
-          </ThemeTag>
+          </AutoTooltip>
         </div>
       </div>
     </div>

src/pages/llmodels/forms/performance.tsx

@@ -8,18 +8,27 @@ import { useFormContext } from '../config/form-context';
 import KVCacheForm from './kv-cache';
 import SpeculativeDecode from './speculative-decode';
 
-const flavorTipsList = [
+const modeTipsList = [
   {
-    title: 'Throughput',
+    title: {
-    tips: 'models.form.flavor.throughput.tips'
+      text: 'models.form.mode.throughput',
+      locale: true
+    },
+    tips: 'models.form.mode.throughput.tips'
   },
   {
-    title: 'Latency',
+    title: {
-    tips: 'models.form.flavor.latency.tips'
+      text: 'models.form.mode.latency',
+      locale: true
+    },
+    tips: 'models.form.mode.latency.tips'
   },
   {
-    title: 'Reference',
+    title: {
-    tips: 'models.form.flavor.reference.tips'
+      text: 'models.form.mode.reference',
+      locale: true
+    },
+    tips: 'models.form.mode.reference.tips'
   }
 ];
 
@@ -34,12 +43,23 @@ const Performance: React.FC = () => {
       {formKey === deployFormKeyMap.catalog && (
         <Form.Item name="mode">
           <SealSelect
-            description={<TooltipList list={flavorTipsList}></TooltipList>}
+            description={<TooltipList list={modeTipsList}></TooltipList>}
-            label="Mode"
+            label={intl.formatMessage({ id: 'models.form.mode' })}
             options={[
-              { label: 'Throughput', value: 'throughput' },
+              {
-              { label: 'Latency', value: 'latency' },
+                label: intl.formatMessage({
-              { label: 'Reference', value: 'reference' }
+                  id: 'models.form.mode.throughput'
+                }),
+                value: 'throughput'
+              },
+              {
+                label: intl.formatMessage({ id: 'models.form.mode.latency' }),
+                value: 'latency'
+              },
+              {
+                label: intl.formatMessage({ id: 'models.form.mode.reference' }),
+                value: 'reference'
+              }
             ]}
           ></SealSelect>
         </Form.Item>

src/pages/llmodels/forms/speculative-decode.tsx

@@ -67,7 +67,9 @@ const SpeculativeDecode = () => {
         style={{ marginBottom: 8 }}
       >
         <CheckboxField
-          label={'Enable Speculative Decoding'}
+          label={intl.formatMessage({
+            id: 'models.form.enableSpeculativeDecoding'
+          })}
           onChange={handleSpeculativeEnabledChange}
         ></CheckboxField>
       </Form.Item>
@@ -78,13 +80,17 @@ const SpeculativeDecode = () => {
             rules={[
               {
                 required: true,
-                message: getRuleMessage('select', 'Algorithm', false)
+                message: getRuleMessage(
+                  'select',
+                  'models.form.algorithm',
+                  false
+                )
               }
             ]}
           >
             <SealSelect
               required
-              label={'Algorithm'}
+              label={intl.formatMessage({ id: 'models.form.algorithm' })}
               options={[
                 { label: 'Eagle3', value: AlgorithmMap.Eagle3 },
                 { label: 'MTP', value: AlgorithmMap.MTP },
@@ -98,13 +104,13 @@ const SpeculativeDecode = () => {
               rules={[
                 {
                   required: true,
-                  message: getRuleMessage('select', 'Draft Model', false)
+                  message: getRuleMessage('select', 'models.form.draftModel')
                 }
               ]}
             >
               <SealSelect
                 required
-                label={'Draft Model'}
+                label={intl.formatMessage({ id: 'models.form.draftModel' })}
                 options={draftModelList}
               ></SealSelect>
             </Form.Item>
@@ -113,7 +119,7 @@ const SpeculativeDecode = () => {
             name={['speculative_config', 'num_draft_tokens']}
           >
             <SealInputNumber
-              label={'Number of Draft Tokens'}
+              label={intl.formatMessage({ id: 'models.form.numDraftTokens' })}
               min={1}
               step={1}
               precision={0}
@@ -125,7 +131,9 @@ const SpeculativeDecode = () => {
                 name={['speculative_config', 'ngram_min_match_length']}
               >
                 <SealInputNumber
-                  label={'Minimum N-gram Match Length'}
+                  label={intl.formatMessage({
+                    id: 'models.form.ngramMinMatchLength'
+                  })}
                   min={1}
                   step={1}
                 />
@@ -134,7 +142,9 @@ const SpeculativeDecode = () => {
                 name={['speculative_config', 'ngram_max_match_length']}
               >
                 <SealInput.Input
-                  label={'Maximum N-gram Match Length'}
+                  label={intl.formatMessage({
+                    id: 'models.form.ngramMaxMatchLength'
+                  })}
                   min={2}
                   step={1}
                 />