@qvac/sdk 0.14.0 → 0.15.0

package/dist/examples/embed-multi-gpu.js

@@ -1,4 +1,4 @@
-import { loadModel, embed, unloadModel, EMBEDDINGGEMMA_300M_Q8_0, } from "@qvac/sdk";
+import { loadModel, embed, unloadModel, EMBEDDINGGEMMA_300M_Q8_0 } from '@qvac/sdk';
 // Multi-GPU embedding distributes model layers across multiple GPUs using
 // llama.cpp's built-in split modes. Two strategies are available:
 //
@@ -16,17 +16,17 @@ try {
     const modelId = await loadModel({
         modelSrc: EMBEDDINGGEMMA_300M_Q8_0,
         modelConfig: {
-            splitMode: "layer",
-            tensorSplit: "1,1",
+            splitMode: 'layer',
+            tensorSplit: '1,1',
             mainGpu: 0,
             gpuLayers: 99,
-            verbosity: 0,
-        },
+            verbosity: 0
+        }
     });
     const texts = [
-        "Multi-GPU embedding distributes layer computation across GPUs.",
-        "Each GPU handles a subset of layers, improving throughput for large models.",
-        "The tensor-split ratio controls how much work each GPU receives.",
+        'Multi-GPU embedding distributes layer computation across GPUs.',
+        'Each GPU handles a subset of layers, improving throughput for large models.',
+        'The tensor-split ratio controls how much work each GPU receives.'
     ];
     for (const text of texts) {
         const { embedding, stats } = await embed({ modelId, text });
@@ -39,6 +39,6 @@ try {
     process.exit(0);
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }

package/dist/examples/embed-p2p.js

@@ -1,4 +1,4 @@
-import { embed, GTE_LARGE_FP16, loadModel, unloadModel } from "@qvac/sdk";
+import { embed, GTE_LARGE_FP16, loadModel, unloadModel } from '@qvac/sdk';
 function cosineSimilarity(vecA, vecB) {
     let dotProduct = 0;
     for (let i = 0; i < vecA.length; i++) {
@@ -14,47 +14,47 @@ try {
             const line = `▸ Downloading ${p.percentage.toFixed(0)}% (${mb(p.downloaded)}/${mb(p.total)} MB)`;
             process.stderr.write(process.stderr.isTTY ? `\r${line}` : `${line}\n`);
             if (p.percentage >= 100)
-                process.stderr.write("\n");
+                process.stderr.write('\n');
         },
         modelConfig: {
             gpuLayers: 99,
-            device: "gpu",
-        },
+            device: 'gpu'
+        }
     });
-    console.log("\n▸ Example 1: Single Text Embedding");
-    console.log("=".repeat(50));
+    console.log('\n▸ Example 1: Single Text Embedding');
+    console.log('='.repeat(50));
     const { embedding: singleEmbedding } = await embed({
         modelId,
-        text: "Hello, world!",
+        text: 'Hello, world!'
     });
     console.log("Input: 'Hello, world!'");
-    console.log("Embedding dimensions:", singleEmbedding.length);
-    console.log("First 10 values:", singleEmbedding.slice(0, 10));
-    console.log("\n▸ Example 2: Batch Text Embeddings");
-    console.log("=".repeat(50));
+    console.log('Embedding dimensions:', singleEmbedding.length);
+    console.log('First 10 values:', singleEmbedding.slice(0, 10));
+    console.log('\n▸ Example 2: Batch Text Embeddings');
+    console.log('='.repeat(50));
     const texts = [
-        "The quick brown fox jumps over the lazy dog",
-        "A fast auburn fox leaps over a sleepy canine",
-        "Python is a programming language",
+        'The quick brown fox jumps over the lazy dog',
+        'A fast auburn fox leaps over a sleepy canine',
+        'Python is a programming language'
     ];
     const { embedding: batchEmbeddings } = await embed({ modelId, text: texts });
-    console.log("Input: Array of", texts.length, "texts");
-    console.log("Output: Array of", batchEmbeddings.length, "embeddings");
+    console.log('Input: Array of', texts.length, 'texts');
+    console.log('Output: Array of', batchEmbeddings.length, 'embeddings');
     const [emb1, emb2, emb3] = batchEmbeddings;
     if (!emb1 || !emb2 || !emb3) {
-        throw new Error("Expected 3 embeddings");
+        throw new Error('Expected 3 embeddings');
     }
-    console.log("Each embedding dimensions:", emb1.length);
-    console.log("\n▸ Similarity Analysis");
-    console.log("=".repeat(50));
+    console.log('Each embedding dimensions:', emb1.length);
+    console.log('\n▸ Similarity Analysis');
+    console.log('='.repeat(50));
     const similarity1 = cosineSimilarity(emb1, emb2);
     const similarity2 = cosineSimilarity(emb1, emb3);
-    console.log("Similarity between texts 1 and 2 (similar meaning):", similarity1.toFixed(4));
-    console.log("Similarity between texts 1 and 3 (different topics):", similarity2.toFixed(4));
-    console.log("\n▸ Higher values indicate more similar meanings");
+    console.log('Similarity between texts 1 and 2 (similar meaning):', similarity1.toFixed(4));
+    console.log('Similarity between texts 1 and 3 (different topics):', similarity2.toFixed(4));
+    console.log('\n▸ Higher values indicate more similar meanings');
     await unloadModel({ modelId, clearStorage: false });
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }

package/dist/examples/finetune/llamacpp-finetune.js

@@ -1,10 +1,10 @@
-import { finetune, loadModel, QWEN3_600M_INST_Q4, unloadModel, } from "@qvac/sdk";
-const pauseResumeEnabled = process.argv.includes("--pause-resume");
+import { finetune, loadModel, QWEN3_600M_INST_Q4, unloadModel } from '@qvac/sdk';
+const pauseResumeEnabled = process.argv.includes('--pause-resume');
 let modelId;
 let exitCode = 0;
 async function readProgress(handle, onTick) {
     for await (const tick of handle.progressStream) {
-        const phase = tick.is_train ? "train" : "val";
+        const phase = tick.is_train ? 'train' : 'val';
         console.log(`▸ epoch=${tick.current_epoch + 1} step=${tick.global_steps} batch=${tick.current_batch}/${tick.total_batches} ${phase} loss=${tick.loss?.toFixed(4)} acc=${tick.accuracy?.toFixed(4)} eta=${Math.round(tick.eta_ms / 1000)}s`);
         onTick(tick.global_steps);
     }
@@ -13,29 +13,29 @@ try {
     modelId = await loadModel({
         modelSrc: QWEN3_600M_INST_Q4,
         modelConfig: {
-            device: "gpu",
-            ctx_size: 512,
-        },
+            device: 'gpu',
+            ctx_size: 512
+        }
     });
     console.log(`▸ Model loaded with ID: ${modelId}`);
     const loadedModelId = modelId;
     const finetuneParams = {
         modelId: loadedModelId,
         options: {
-            trainDatasetDir: "./examples/finetune/input/small_train_HF.jsonl",
+            trainDatasetDir: './examples/finetune/input/small_train_HF.jsonl',
             validation: {
-                type: "dataset",
-                path: "./examples/finetune/input/small_eval_HF.jsonl",
+                type: 'dataset',
+                path: './examples/finetune/input/small_eval_HF.jsonl'
             },
             numberOfEpochs: 2,
             learningRate: 1e-4,
             lrMin: 1e-8,
-            loraModules: "attn_q,attn_k,attn_v,attn_o,ffn_gate,ffn_up,ffn_down",
+            loraModules: 'attn_q,attn_k,attn_v,attn_o,ffn_gate,ffn_up,ffn_down',
             assistantLossOnly: true,
             checkpointSaveSteps: 2,
-            checkpointSaveDir: "./examples/finetune/results/checkpoints",
-            outputParametersDir: "./examples/finetune/results",
-        },
+            checkpointSaveDir: './examples/finetune/results/checkpoints',
+            outputParametersDir: './examples/finetune/results'
+        }
     };
     const handle = finetune(finetuneParams);
     let pauseRequested = false;
@@ -43,10 +43,10 @@ try {
     const progressTask = readProgress(handle, (globalSteps) => {
         if (pauseResumeEnabled && !pauseRequested && globalSteps >= 10) {
             pauseRequested = true;
-            console.log("▸ Requesting a pause so the run can be resumed...");
+            console.log('▸ Requesting a pause so the run can be resumed...');
             pauseResultPromise = finetune({
                 modelId: loadedModelId,
-                operation: "pause",
+                operation: 'pause'
             });
         }
     });
@@ -55,21 +55,21 @@ try {
     if (pauseResultPromise) {
         await pauseResultPromise;
     }
-    console.log("▸ Initial finetune result:", initialResult);
-    if (pauseResumeEnabled && initialResult.status === "PAUSED") {
-        console.log("▸ Resuming from the saved checkpoint...");
+    console.log('▸ Initial finetune result:', initialResult);
+    if (pauseResumeEnabled && initialResult.status === 'PAUSED') {
+        console.log('▸ Resuming from the saved checkpoint...');
         const resumedHandle = finetune({
             ...finetuneParams,
-            operation: "resume",
+            operation: 'resume'
         });
         const resumedProgressTask = readProgress(resumedHandle, function () { });
         const resumedResult = await resumedHandle.result;
         await resumedProgressTask;
-        console.log("▸ Resumed finetune result:", resumedResult);
+        console.log('▸ Resumed finetune result:', resumedResult);
     }
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     exitCode = 1;
 }
 finally {

package/dist/examples/kv-cache-cleanup.js

@@ -1,43 +1,43 @@
-import { completion, deleteCache, LLAMA_3_2_1B_INST_Q4_0, loadModel, unloadModel, VERBOSITY, } from "@qvac/sdk";
+import { completion, deleteCache, LLAMA_3_2_1B_INST_Q4_0, loadModel, unloadModel, VERBOSITY } from '@qvac/sdk';
 try {
-    console.log("▸ KV Cache Cleanup Demo\n");
-    console.log("▸ Cache structure: {kvCacheKey}/{modelId}/{configHash}.bin\n");
-    console.log("▸ " + "=".repeat(70) + "\n");
+    console.log('▸ KV Cache Cleanup Demo\n');
+    console.log('▸ Cache structure: {kvCacheKey}/{modelId}/{configHash}.bin\n');
+    console.log('▸ ' + '='.repeat(70) + '\n');
     const modelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
         modelConfig: {
-            device: "gpu",
+            device: 'gpu',
             ctx_size: 2048,
-            verbosity: VERBOSITY.ERROR,
-        },
+            verbosity: VERBOSITY.ERROR
+        }
     });
     // Create some cached sessions
-    const cacheKey1 = "user-alice-session";
-    const cacheKey2 = "user-bob-session";
+    const cacheKey1 = 'user-alice-session';
+    const cacheKey2 = 'user-bob-session';
     console.log(`▸ Creating cache for key: "${cacheKey1}"`);
     const result1 = completion({
         modelId,
-        history: [{ role: "user", content: "Hello, I'm Alice!" }],
+        history: [{ role: 'user', content: "Hello, I'm Alice!" }],
         stream: true,
-        kvCache: cacheKey1,
+        kvCache: cacheKey1
     });
     for await (const token of result1.tokenStream) {
         process.stdout.write(token);
     }
-    console.log("\n▸ Cache created\n");
+    console.log('\n▸ Cache created\n');
     console.log(`▸ Creating cache for key: "${cacheKey2}"`);
     const result2 = completion({
         modelId,
-        history: [{ role: "user", content: "Hello, I'm Bob!" }],
+        history: [{ role: 'user', content: "Hello, I'm Bob!" }],
         stream: true,
-        kvCache: cacheKey2,
+        kvCache: cacheKey2
     });
     for await (const token of result2.tokenStream) {
         process.stdout.write(token);
     }
-    console.log("\n▸ Cache created\n");
+    console.log('\n▸ Cache created\n');
     // Now cleanup specific cache key
-    console.log("▸ " + "=".repeat(70));
+    console.log('▸ ' + '='.repeat(70));
     console.log(`\n▸ Deleting cache key: "${cacheKey1}"`);
     const deleteResult1 = await deleteCache({ kvCacheKey: cacheKey1 });
     if (deleteResult1.success) {
@@ -48,7 +48,7 @@ try {
     console.log(`\n▸ Deleting specific model in cache key: "${cacheKey2}"`);
     const deleteResult2 = await deleteCache({
         kvCacheKey: cacheKey2,
-        modelId: modelId,
+        modelId: modelId
     });
     if (deleteResult2.success) {
         console.log(`▸ Deleted model "${modelId}" from cache key "${cacheKey2}"`);
@@ -56,31 +56,31 @@ try {
     }
     // Try to delete non-existent cache key (should succeed silently)
     console.log(`\n▸ Deleting non-existent cache key: "non-existent"`);
-    const deleteResult3 = await deleteCache({ kvCacheKey: "non-existent" });
+    const deleteResult3 = await deleteCache({ kvCacheKey: 'non-existent' });
     if (deleteResult3.success) {
         console.log("▸ Delete succeeded (cache key didn't exist)");
     }
-    console.log("\n▸ " + "=".repeat(70));
-    console.log("\n▸ CLEANUP OPTIONS:");
-    console.log("▸ " + "-".repeat(70));
-    console.log("\n▸ 1. Delete entire cache key (all models):");
+    console.log('\n▸ ' + '='.repeat(70));
+    console.log('\n▸ CLEANUP OPTIONS:');
+    console.log('▸ ' + '-'.repeat(70));
+    console.log('\n▸ 1. Delete entire cache key (all models):');
     console.log('▸    await deleteCache({ kvCacheKey: "my-session" })');
-    console.log("▸    └─ Removes: my-session/ (entire directory)");
-    console.log("\n▸ 2. Delete specific model within cache key:");
+    console.log('▸    └─ Removes: my-session/ (entire directory)');
+    console.log('\n▸ 2. Delete specific model within cache key:');
     console.log('▸    await deleteCache({ kvCacheKey: "my-session", modelId: "model-abc" })');
-    console.log("▸    └─ Removes: my-session/model-abc/ (specific model directory)");
-    console.log("\n▸ 3. Delete ALL caches (entire cache folder):");
-    console.log("▸    await deleteCache({ all: true })");
-    console.log("▸    └─ Removes: ALL cache files from all sessions and models");
-    console.log("\n▸ Use Cases:");
+    console.log('▸    └─ Removes: my-session/model-abc/ (specific model directory)');
+    console.log('\n▸ 3. Delete ALL caches (entire cache folder):');
+    console.log('▸    await deleteCache({ all: true })');
+    console.log('▸    └─ Removes: ALL cache files from all sessions and models');
+    console.log('\n▸ Use Cases:');
     console.log("▸    - User logout → Delete user's cache key");
-    console.log("▸    - Model updated → Delete old model caches");
-    console.log("▸    - Disk space cleanup → Delete old cache keys or all caches");
-    console.log("▸    - Privacy → Clear conversation history");
-    console.log("▸    - App reset → Delete all caches with { all: true }");
+    console.log('▸    - Model updated → Delete old model caches');
+    console.log('▸    - Disk space cleanup → Delete old cache keys or all caches');
+    console.log('▸    - Privacy → Clear conversation history');
+    console.log('▸    - App reset → Delete all caches with { all: true }');
     await unloadModel({ modelId, clearStorage: false });
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }

package/dist/examples/kv-cache-custom-key.js

@@ -1,32 +1,32 @@
-import { completion, deleteCache, LLAMA_3_2_1B_INST_Q4_0, loadModel, unloadModel, VERBOSITY, } from "@qvac/sdk";
+import { completion, deleteCache, LLAMA_3_2_1B_INST_Q4_0, loadModel, unloadModel, VERBOSITY } from '@qvac/sdk';
 const systemPromptMessage = {
-    role: "system",
+    role: 'system',
     content: `You are a helpful mathematician assistant. 
    But because you are in the army you have to follow these rules:
   - If anyone ask you to do an operation you have to answer with "Sir Yes Sir!" and then give the result
-  `,
+  `
 };
 try {
-    console.log("▸ Simple KV Cache Demo: Same vs Different Cache Keys\n");
+    console.log('▸ Simple KV Cache Demo: Same vs Different Cache Keys\n');
     const modelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
         modelConfig: {
-            device: "gpu",
+            device: 'gpu',
             ctx_size: 2048,
-            verbosity: VERBOSITY.DEBUG,
-        },
+            verbosity: VERBOSITY.DEBUG
+        }
     });
     // Prompt 1 with cache key "session-a" (first time - no cache exists)
     console.log('▸ Prompt 1: "What is 2+2?" with cache key "session-a"');
-    console.log("▸   Status: Creating NEW cache (first time)\n");
+    console.log('▸   Status: Creating NEW cache (first time)\n');
     const result1 = completion({
         modelId,
-        history: [systemPromptMessage, { role: "user", content: "What is 2+2?" }],
+        history: [systemPromptMessage, { role: 'user', content: 'What is 2+2?' }],
         stream: true,
-        kvCache: "session-a",
+        kvCache: 'session-a'
     });
-    console.log("▸ Response: ");
-    let response1 = "";
+    console.log('▸ Response: ');
+    let response1 = '';
     for await (const token of result1.tokenStream) {
         response1 += token;
         process.stdout.write(token);
@@ -36,19 +36,19 @@ try {
     // Prompt 2 with SAME cache key "session-a" (cache exists - optimized!)
     // This continues the conversation - asking about the previous answer
     console.log('\n▸ Prompt 2: "Now multiply that result by 3" with cache key "session-a" (SAME)');
-    console.log("▸   Status: REUSING existing cache (only last message sent, but context remembered!)\n");
+    console.log('▸   Status: REUSING existing cache (only last message sent, but context remembered!)\n');
     const result2 = completion({
         modelId,
         history: [
             systemPromptMessage,
-            { role: "user", content: "What is 2+2?" },
-            { role: "assistant", content: response1 },
-            { role: "user", content: "Now multiply that result by 3" }, // Refers to previous answer!
+            { role: 'user', content: 'What is 2+2?' },
+            { role: 'assistant', content: response1 },
+            { role: 'user', content: 'Now multiply that result by 3' } // Refers to previous answer!
         ],
         stream: true,
-        kvCache: "session-a", // Same cache key!
+        kvCache: 'session-a' // Same cache key!
     });
-    console.log("▸ Response: ");
+    console.log('▸ Response: ');
     for await (const token of result2.tokenStream) {
         process.stdout.write(token);
     }
@@ -56,14 +56,14 @@ try {
     console.log(`\n▸ Stats: TTFT=${stats2?.timeToFirstToken}ms, CacheTokens=${stats2?.cacheTokens}\n`);
     // Prompt 3 with DIFFERENT cache key "session-b" (new cache)
     console.log('\n▸ Prompt 3: "What is 10+10?" with cache key "session-b" (DIFFERENT)');
-    console.log("▸   Status: Creating NEW cache (different key = isolated cache)\n");
+    console.log('▸   Status: Creating NEW cache (different key = isolated cache)\n');
     const result3 = completion({
         modelId,
-        history: [systemPromptMessage, { role: "user", content: "What is 10+10?" }],
+        history: [systemPromptMessage, { role: 'user', content: 'What is 10+10?' }],
         stream: true,
-        kvCache: "session-b", // Different cache key!
+        kvCache: 'session-b' // Different cache key!
     });
-    console.log("▸ Response: ");
+    console.log('▸ Response: ');
     for await (const token of result3.tokenStream) {
         process.stdout.write(token);
     }
@@ -71,30 +71,30 @@ try {
     console.log(`\n▸ Stats: TTFT=${stats3?.timeToFirstToken}ms, CacheTokens=${stats3?.cacheTokens}\n`);
     // Prompt 4: Go back to session-a to trigger flush of session-b
     console.log('\n▸ Prompt 4: "What is 5+5?" with cache key "session-a" (BACK TO FIRST)');
-    console.log("▸   Status: This should trigger session-b to be flushed to disk\n");
+    console.log('▸   Status: This should trigger session-b to be flushed to disk\n');
     const result4 = completion({
         modelId,
-        history: [systemPromptMessage, { role: "user", content: "What is 5+5?" }],
+        history: [systemPromptMessage, { role: 'user', content: 'What is 5+5?' }],
         stream: true,
-        kvCache: "session-a", // Back to session-a!
+        kvCache: 'session-a' // Back to session-a!
     });
-    console.log("▸ Response: ");
+    console.log('▸ Response: ');
     for await (const token of result4.tokenStream) {
         process.stdout.write(token);
     }
     const stats4 = await result4.stats;
     console.log(`\n▸ Stats: TTFT=${stats4?.timeToFirstToken}ms, CacheTokens=${stats4?.cacheTokens}\n`);
     // Summary
-    console.log("\n▸ Summary:");
+    console.log('\n▸ Summary:');
     console.log('▸   - Prompts 1 & 2 share cache key "session-a" → same cache file');
     console.log("▸   - Prompt 2 built on Prompt 1's context (model remembered the answer!)");
-    console.log("▸   - Only the last message was sent, but full conversation context maintained");
+    console.log('▸   - Only the last message was sent, but full conversation context maintained');
     console.log('▸   - Prompt 3 uses cache key "session-b" → separate cache file (isolated)');
-    console.log("▸   - Different cache keys provide complete isolation between sessions");
+    console.log('▸   - Different cache keys provide complete isolation between sessions');
     await deleteCache({ all: true });
     await unloadModel({ modelId, clearStorage: false });
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }

package/dist/examples/kv-cache-example.js

@@ -1,25 +1,23 @@
-import { completion, LLAMA_3_2_1B_INST_Q4_0, loadModel, unloadModel, VERBOSITY, } from "@qvac/sdk";
+import { completion, LLAMA_3_2_1B_INST_Q4_0, loadModel, unloadModel, VERBOSITY } from '@qvac/sdk';
 try {
     // Load the model
     const modelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
         modelConfig: {
-            device: "gpu",
+            device: 'gpu',
             ctx_size: 2048,
-            verbosity: VERBOSITY.ERROR,
-        },
+            verbosity: VERBOSITY.ERROR
+        }
     });
-    console.log("▸ Testing KV Cache functionality...\n");
+    console.log('▸ Testing KV Cache functionality...\n');
     // First conversation with auto-keyed cache enabled
-    console.log("▸ First conversation (building cache for the next turn):");
-    const history1 = [
-        { role: "user", content: "What is the capital of France?" },
-    ];
+    console.log('▸ First conversation (building cache for the next turn):');
+    const history1 = [{ role: 'user', content: 'What is the capital of France?' }];
     const result1 = completion({
         modelId,
         history: history1,
         stream: true,
-        kvCache: true,
+        kvCache: true
     }); // kvCache = true
     for await (const token of result1.tokenStream) {
         process.stdout.write(token);
@@ -28,14 +26,14 @@ try {
     const stats1 = final1.stats;
     console.log(`\n▸ First completion stats: ${JSON.stringify(stats1)}\n`);
     // Continue conversation (should reuse the completed first-turn cache).
-    console.log("▸ Continuing conversation (reusing previous turn cache):");
+    console.log('▸ Continuing conversation (reusing previous turn cache):');
     const history2 = [
-        { role: "user", content: "What is the capital of France?" },
+        { role: 'user', content: 'What is the capital of France?' },
         {
-            role: "assistant",
-            content: final1.cacheableAssistantContent ?? final1.contentText,
+            role: 'assistant',
+            content: final1.cacheableAssistantContent ?? final1.contentText
         },
-        { role: "user", content: "What about Germany?" },
+        { role: 'user', content: 'What about Germany?' }
     ];
     // Auto-keyed caching should:
     // 1. Find the cache saved after turn 1 under [user, assistant]
@@ -45,7 +43,7 @@ try {
         modelId,
         history: history2,
         stream: true,
-        kvCache: true,
+        kvCache: true
     }); // kvCache = true
     for await (const token of result2.tokenStream) {
         process.stdout.write(token);
@@ -53,22 +51,22 @@ try {
     const stats2 = await result2.stats;
     console.log(`\n▸ Second completion stats: ${JSON.stringify(stats2)}\n`);
     // Compare with non-cached version
-    console.log("▸ Same conversation without cache:");
+    console.log('▸ Same conversation without cache:');
     const result3 = completion({
         modelId,
         history: history2,
         stream: true,
-        kvCache: false,
+        kvCache: false
     }); // kvCache = false
     for await (const token of result3.tokenStream) {
         process.stdout.write(token);
     }
     const stats3 = await result3.stats;
     console.log(`\n▸ Non-cached completion stats: ${JSON.stringify(stats3)}\n`);
-    console.log("▸ KV Cache test completed!");
+    console.log('▸ KV Cache test completed!');
     await unloadModel({ modelId, clearStorage: false });
 }
 catch (error) {
-    console.error("✖", error);
+    console.error('✖', error);
     process.exit(1);
 }