chore: update extended_kv_cache config

src/locales/en-US/models.ts

@@ -212,5 +212,12 @@ export default {
     'Available only with built-in backends (vLLM / SGLang) — switch backend in <span class="bold-text">Advanced</span> to enable.',
   'models.form.kvCache.tips2':
     'KV cache is only supported when using built-in inference backends (vLLM or SGLang).',
-  'models.form.scheduling': 'Scheduling'
+  'models.form.scheduling': 'Scheduling',
+  'models.form.ramRatio': 'RAM-to-VRAM Ratio',
+  'models.form.ramSize': 'Maximum RAM Size (GiB)',
+  '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. A larger chunk size may improve throughput but increase memory usage.'
 };

src/locales/ja-JP/models.ts

@@ -212,7 +212,14 @@ export default {
     'Available only with built-in backends (vLLM / SGLang) — switch backend in <span class="bold-text">Advanced</span> to enable.',
   'models.form.kvCache.tips2':
     'KV cache is only supported when using built-in inference backends (vLLM or SGLang).',
-  'models.form.scheduling': 'Scheduling'
+  'models.form.scheduling': 'Scheduling',
+  'models.form.ramRatio': 'RAM-to-VRAM Ratio',
+  'models.form.ramSize': 'Maximum RAM Size (GiB)',
+  '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. A larger chunk size may improve throughput but increase memory usage.'
 };
 
 // ========== To-Do: Translate Keys (Remove After Translation) ==========
@@ -259,5 +266,10 @@ export default {
 // 43. 'models.form.remoteURL.tips': 'Refer to the <a href="https://docs.lmcache.ai/api_reference/configurations.html" target="_blank">configuration documentation</a> for details.',
 // 44. 'models.form.kvCache.tips': 'Available only with built-in backends (vLLM / SGLang) — switch backend in <span class="bold-text">Advanced</span> to enable.'
 // 45. 'models.form.kvCache.tips2': 'KV cache is only supported when using built-in inference backends (vLLM or SGLang).',
-// 46. 'models.form.scheduling': 'Scheduling'
+// 46. 'models.form.scheduling': 'Scheduling',
+// 47. 'models.form.ramRatio': 'RAM-to-VRAM Ratio',
+// 48. 'models.form.ramSize': 'Maximum RAM Size (GiB)',
+// 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. A larger chunk size may improve throughput but increase memory usage.'
 // ========== End of To-Do List ==========

src/locales/ru-RU/models.ts

@@ -212,7 +212,14 @@ export default {
     'Available only with built-in backends (vLLM / SGLang) — switch backend in <span class="bold-text">Advanced</span> to enable.',
   'models.form.kvCache.tips2':
     'KV cache is only supported when using built-in inference backends (vLLM or SGLang).',
-  'models.form.scheduling': 'Scheduling'
+  'models.form.scheduling': 'Scheduling',
+  'models.form.ramRatio': 'RAM-to-VRAM Ratio',
+  'models.form.ramSize': 'Maximum RAM Size (GiB)',
+  '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. A larger chunk size may improve throughput but increase memory usage.'
 };
 
 // ========== To-Do: Translate Keys (Remove After Translation) ==========
@@ -222,5 +229,10 @@ export default {
 // 5. 'models.form.remoteURL.tips': 'Refer to the <a href="https://docs.lmcache.ai/api_reference/configurations.html" target="_blank">configuration documentation</a> for details.',
 // 6. 'models.form.kvCache.tips': 'Available only with built-in backends (vLLM / SGLang) — switch backend in <span class="bold-text">Advanced</span> to enable.'
 // 7. 'models.form.kvCache.tips2': 'KV cache is only supported when using built-in inference backends (vLLM or SGLang).';
-// 8. 'models.form.scheduling': 'Scheduling'
+// 8. 'models.form.scheduling': 'Scheduling',
+// 9. 'models.form.ramRatio': 'RAM-to-VRAM Ratio',
+// 10. 'models.form.ramSize': 'Maximum RAM Size (GiB)',
+// 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. A larger chunk size may improve throughput but increase memory usage.'
 // ========== End of To-Do List ==========

src/locales/zh-CN/models.ts

@@ -201,5 +201,12 @@ export default {
     '仅在内置后端（vLLM / SGLang）可用 —— 请在<span class="bold-text">高级</span>配置中切换后端以启用。',
   'models.form.kvCache.tips2':
     '仅在使用内置推理后端（vLLM 或 SGLang）时支持 KV 缓存。',
-  'models.form.scheduling': '调度'
+  'models.form.scheduling': '调度',
+  'models.form.ramRatio': '内存与显存比例',
+  'models.form.ramSize': '内存最大占用 (GiB)',
+  'models.form.ramRatio.tips':
+    '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。'
 };