@qvac/sdk 0.13.4 → 0.14.0

package/dist/examples/download-with-cancel.js

@@ -1,6 +1,6 @@
 import { cancel, close, downloadAsset, LLAMA_3_2_1B_INST_Q4_0, } from "@qvac/sdk";
-console.log(`🚀 Starting download with pause/resume example`);
-console.log(`\n💡 Press Ctrl+C to pause the download (it will resume on restart)\n`);
+console.log(`▸ Starting download with pause/resume example`);
+console.log(`\n▸ Press Ctrl+C to pause the download (it will resume on restart)\n`);
 let modelId;
 let cancelled = false;
 try {
@@ -10,16 +10,15 @@ try {
     // synchronous `requestId` field so we can cancel before it settles.
     const download = downloadAsset({
         assetSrc: LLAMA_3_2_1B_INST_Q4_0,
-        onProgress: (progress) => {
-            const downloadedMB = (progress.downloaded / 1024 / 1024).toFixed(2);
-            const totalMB = (progress.total / 1024 / 1024).toFixed(2);
-            const percentage = progress.percentage.toFixed(1);
-            console.log(`📊 Progress: ${percentage}% (${downloadedMB}MB / ${totalMB}MB)`);
+        onProgress: (p) => {
+            const mb = (n) => (n / 1e6).toFixed(1);
+            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");
             // Example: Stops at 10% (or use Ctrl+C for manual stop)
-            if (parseFloat(percentage) >= 10 && !cancelled) {
-                console.log("\n🚫 Auto-cancelling at 10% for demo purposes...,");
-                console.log(`📊 Progress: ${percentage}% (${downloadedMB}MB / ${totalMB}MB)`);
-                console.log(progress);
+            if (p.percentage >= 10 && !cancelled) {
+                console.log("\n▸ Auto-cancelling at 10% for demo purposes...");
                 cancelled = true;
                 void cancel({
                     requestId: download.requestId,
@@ -28,18 +27,18 @@ try {
             }
         },
     });
-    await download;
-    console.log(`\n✅ Model downloaded successfully! Model ID: ${modelId}`);
-    console.log("🎯 Download completed without interruption");
+    modelId = await download;
+    console.log(`\n▸ Model downloaded successfully! Model ID: ${modelId}`);
+    console.log("▸ Download completed without interruption");
     void close();
 }
 catch (error) {
     if (error instanceof Error && error.message.includes("cancelled")) {
-        console.log("✅ Download was successfully cancelled");
+        console.log("▸ Download was successfully cancelled");
         void close();
     }
     else {
-        console.error("❌ Error:", error);
+        console.error("✖", error);
         process.exit(1);
     }
 }

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

@@ -30,16 +30,15 @@ try {
     ];
     for (const text of texts) {
         const { embedding, stats } = await embed({ modelId, text });
-        console.log(`Embedded ${text.slice(0, 50)}...`);
-        console.log(`  Dimensions: ${embedding.length}`);
+        console.log(`${embedding.length} dims  ${text.slice(0, 50)}...`);
         if (stats) {
-            console.log(`  Backend: ${stats.backendDevice}, TPS: ${stats.tokensPerSecond?.toFixed(1)}`);
+            console.log(`▸ ${stats.backendDevice}  ${stats.tokensPerSecond?.toFixed(1)} tok/s`);
         }
     }
     await unloadModel({ modelId, clearStorage: false });
     process.exit(0);
 }
 catch (error) {
-    console.error("Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

package/dist/examples/embed-p2p.js

@@ -9,15 +9,19 @@ function cosineSimilarity(vecA, vecB) {
 try {
     const modelId = await loadModel({
         modelSrc: GTE_LARGE_FP16,
-        onProgress: (progress) => {
-            console.log(progress);
+        onProgress: (p) => {
+            const mb = (n) => (n / 1e6).toFixed(1);
+            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");
         },
         modelConfig: {
             gpuLayers: 99,
             device: "gpu",
         },
     });
-    console.log("\n📝 Example 1: Single Text Embedding");
+    console.log("\n▸ Example 1: Single Text Embedding");
     console.log("=".repeat(50));
     const { embedding: singleEmbedding } = await embed({
         modelId,
@@ -26,7 +30,7 @@ try {
     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("\n▸ Example 2: Batch Text Embeddings");
     console.log("=".repeat(50));
     const texts = [
         "The quick brown fox jumps over the lazy dog",
@@ -41,16 +45,16 @@ try {
         throw new Error("Expected 3 embeddings");
     }
     console.log("Each embedding dimensions:", emb1.length);
-    console.log("\n🔍 Similarity Analysis");
+    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("\n▸ Higher values indicate more similar meanings");
     await unloadModel({ modelId, clearStorage: false });
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

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

@@ -5,7 +5,7 @@ let exitCode = 0;
 async function readProgress(handle, onTick) {
     for await (const tick of handle.progressStream) {
         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`);
+        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);
     }
 }
@@ -17,7 +17,7 @@ try {
             ctx_size: 512,
         },
     });
-    console.log(`✅ Model loaded with ID: ${modelId}`);
+    console.log(`▸ Model loaded with ID: ${modelId}`);
     const loadedModelId = modelId;
     const finetuneParams = {
         modelId: loadedModelId,
@@ -43,7 +43,7 @@ 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",
@@ -55,9 +55,9 @@ try {
     if (pauseResultPromise) {
         await pauseResultPromise;
     }
-    console.log("📦 Initial finetune result:", initialResult);
+    console.log("▸ Initial finetune result:", initialResult);
     if (pauseResumeEnabled && initialResult.status === "PAUSED") {
-        console.log("▶️ Resuming from the saved checkpoint...");
+        console.log("▸ Resuming from the saved checkpoint...");
         const resumedHandle = finetune({
             ...finetuneParams,
             operation: "resume",
@@ -65,11 +65,11 @@ try {
         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:", error);
+    console.error("✖", error);
     exitCode = 1;
 }
 finally {

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

@@ -1,8 +1,8 @@
 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: {
@@ -14,7 +14,7 @@ try {
     // Create some cached sessions
     const cacheKey1 = "user-alice-session";
     const cacheKey2 = "user-bob-session";
-    console.log(`📝 Creating cache for key: "${cacheKey1}"`);
+    console.log(`▸ Creating cache for key: "${cacheKey1}"`);
     const result1 = completion({
         modelId,
         history: [{ role: "user", content: "Hello, I'm Alice!" }],
@@ -24,8 +24,8 @@ try {
     for await (const token of result1.tokenStream) {
         process.stdout.write(token);
     }
-    console.log("\n✅ Cache created\n");
-    console.log(`📝 Creating cache for key: "${cacheKey2}"`);
+    console.log("\n▸ Cache created\n");
+    console.log(`▸ Creating cache for key: "${cacheKey2}"`);
     const result2 = completion({
         modelId,
         history: [{ role: "user", content: "Hello, I'm Bob!" }],
@@ -35,52 +35,52 @@ try {
     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(`\n🗑️  Deleting cache key: "${cacheKey1}"`);
+    console.log("▸ " + "=".repeat(70));
+    console.log(`\n▸ Deleting cache key: "${cacheKey1}"`);
     const deleteResult1 = await deleteCache({ kvCacheKey: cacheKey1 });
     if (deleteResult1.success) {
-        console.log(`✅ Deleted all caches for key "${cacheKey1}"`);
-        console.log(`   (Removed: ${cacheKey1}/{modelId}/{configHash}.bin)`);
+        console.log(`▸ Deleted all caches for key "${cacheKey1}"`);
+        console.log(`▸   (Removed: ${cacheKey1}/{modelId}/{configHash}.bin)`);
     }
     // Delete specific model within a cache key
-    console.log(`\n🗑️  Deleting specific model in cache key: "${cacheKey2}"`);
+    console.log(`\n▸ Deleting specific model in cache key: "${cacheKey2}"`);
     const deleteResult2 = await deleteCache({
         kvCacheKey: cacheKey2,
         modelId: modelId,
     });
     if (deleteResult2.success) {
-        console.log(`✅ Deleted model "${modelId}" from cache key "${cacheKey2}"`);
-        console.log(`   (Removed: ${cacheKey2}/${modelId}/)`);
+        console.log(`▸ Deleted model "${modelId}" from cache key "${cacheKey2}"`);
+        console.log(`▸   (Removed: ${cacheKey2}/${modelId}/)`);
     }
     // Try to delete non-existent cache key (should succeed silently)
-    console.log(`\n🗑️  Deleting non-existent cache key: "non-existent"`);
+    console.log(`\n▸ Deleting non-existent cache key: "non-existent"`);
     const deleteResult3 = await deleteCache({ kvCacheKey: "non-existent" });
     if (deleteResult3.success) {
-        console.log("✅ Delete succeeded (cache key didn't exist)");
+        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("\n1️⃣  Delete entire cache key (all models):");
-    console.log('   await deleteCache({ kvCacheKey: "my-session" })');
-    console.log("   └─ Removes: my-session/ (entire directory)");
-    console.log("\n2️⃣  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("\n3️⃣  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("\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('▸    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("▸    - 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 }");
     await unloadModel({ modelId, clearStorage: false });
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

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

@@ -7,7 +7,7 @@ const systemPromptMessage = {
   `,
 };
 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: {
@@ -17,26 +17,26 @@ try {
         },
     });
     // 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('▸ Prompt 1: "What is 2+2?" with cache key "session-a"');
+    console.log("▸   Status: Creating NEW cache (first time)\n");
     const result1 = completion({
         modelId,
         history: [systemPromptMessage, { role: "user", content: "What is 2+2?" }],
         stream: true,
         kvCache: "session-a",
     });
-    console.log("   Response: ");
+    console.log("▸ Response: ");
     let response1 = "";
     for await (const token of result1.tokenStream) {
         response1 += token;
         process.stdout.write(token);
     }
     const stats1 = await result1.stats;
-    console.log(`\n   Stats: TTFT=${stats1?.timeToFirstToken}ms, CacheTokens=${stats1?.cacheTokens}\n`);
+    console.log(`\n▸ Stats: TTFT=${stats1?.timeToFirstToken}ms, CacheTokens=${stats1?.cacheTokens}\n`);
     // 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('\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");
     const result2 = completion({
         modelId,
         history: [
@@ -48,53 +48,53 @@ try {
         stream: true,
         kvCache: "session-a", // Same cache key!
     });
-    console.log("   Response: ");
+    console.log("▸ Response: ");
     for await (const token of result2.tokenStream) {
         process.stdout.write(token);
     }
     const stats2 = await result2.stats;
-    console.log(`\n   Stats: TTFT=${stats2?.timeToFirstToken}ms, CacheTokens=${stats2?.cacheTokens}\n`);
+    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('\n▸ Prompt 3: "What is 10+10?" with cache key "session-b" (DIFFERENT)');
+    console.log("▸   Status: Creating NEW cache (different key = isolated cache)\n");
     const result3 = completion({
         modelId,
         history: [systemPromptMessage, { role: "user", content: "What is 10+10?" }],
         stream: true,
         kvCache: "session-b", // Different cache key!
     });
-    console.log("   Response: ");
+    console.log("▸ Response: ");
     for await (const token of result3.tokenStream) {
         process.stdout.write(token);
     }
     const stats3 = await result3.stats;
-    console.log(`\n   Stats: TTFT=${stats3?.timeToFirstToken}ms, CacheTokens=${stats3?.cacheTokens}\n`);
+    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('\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");
     const result4 = completion({
         modelId,
         history: [systemPromptMessage, { role: "user", content: "What is 5+5?" }],
         stream: true,
         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`);
+    console.log(`\n▸ Stats: TTFT=${stats4?.timeToFirstToken}ms, CacheTokens=${stats4?.cacheTokens}\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('   - Prompt 3 uses cache key "session-b" → separate cache file (isolated)');
-    console.log("   - Different cache keys provide complete isolation between sessions");
+    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('▸   - Prompt 3 uses cache key "session-b" → separate cache file (isolated)');
+    console.log("▸   - Different cache keys provide complete isolation between sessions");
     await deleteCache({ all: true });
     await unloadModel({ modelId, clearStorage: false });
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

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

@@ -9,9 +9,9 @@ try {
             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):");
+    console.log("▸ First conversation (building cache for the next turn):");
     const history1 = [
         { role: "user", content: "What is the capital of France?" },
     ];
@@ -26,9 +26,9 @@ try {
     }
     const final1 = await result1.final;
     const stats1 = final1.stats;
-    console.log(`\n⏱️  First completion stats: ${JSON.stringify(stats1)}\n`);
+    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?" },
         {
@@ -51,9 +51,9 @@ try {
         process.stdout.write(token);
     }
     const stats2 = await result2.stats;
-    console.log(`\n⏱️  Second completion stats: ${JSON.stringify(stats2)}\n`);
+    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,
@@ -64,11 +64,11 @@ try {
         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(`\n▸ Non-cached completion stats: ${JSON.stringify(stats3)}\n`);
+    console.log("▸ KV Cache test completed!");
     await unloadModel({ modelId, clearStorage: false });
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

package/dist/examples/llamacpp-cache.js

@@ -6,16 +6,16 @@ const systemPrompt = {
     content: "You are a concise travel assistant. Keep track of the user's preferences across turns.",
 };
 function printStats(stats) {
-    console.log("\nPerformance Stats:", {
+    console.log("\n▸ Performance Stats:", {
         timeToFirstToken: stats?.timeToFirstToken,
         tokensPerSecond: stats?.tokensPerSecond,
         cacheTokens: stats?.cacheTokens,
     });
 }
 async function runCachedCompletion(params) {
-    console.log(`\n=== ${params.label} ===`);
-    console.log(`Cache key: ${params.kvCache || "disabled"}`);
-    console.log("AI Response:");
+    console.log(`\n▸ ${params.label}`);
+    console.log(`▸ Cache key: ${params.kvCache || "disabled"}`);
+    console.log("▸ AI Response:");
     const result = completion({
         modelId: params.modelId,
         history: params.history,
@@ -33,18 +33,22 @@ async function runCachedCompletion(params) {
 }
 let modelId;
 try {
-    console.log("Loading llama.cpp model with cache support...");
+    console.log("▸ Loading llama.cpp model with cache support...");
     modelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
         modelConfig: {
             ctx_size: 4096,
             verbosity: VERBOSITY.ERROR,
         },
-        onProgress: (progress) => {
-            console.log(`Loading: ${progress.percentage.toFixed(1)}%`);
+        onProgress: (p) => {
+            const mb = (n) => (n / 1e6).toFixed(1);
+            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");
         },
     });
-    console.log(`Model loaded successfully. Model ID: ${modelId}`);
+    console.log(`▸ Model loaded successfully. Model ID: ${modelId}`);
     await deleteCache({ kvCacheKey: sharedCacheKey });
     await deleteCache({ kvCacheKey: isolatedCacheKey });
     const firstUserMessage = {
@@ -83,10 +87,10 @@ try {
             },
         ],
     });
-    console.log("\nSummary:");
-    console.log(`- Reusing "${sharedCacheKey}" keeps conversation context across turns.`);
-    console.log(`- Switching to "${isolatedCacheKey}" starts a separate cache session.`);
-    console.log('- Use deleteCache({ kvCacheKey: "your-session" }) to clear a saved session.');
+    console.log("\n▸ Summary:");
+    console.log(`▸ - Reusing "${sharedCacheKey}" keeps conversation context across turns.`);
+    console.log(`▸ - Switching to "${isolatedCacheKey}" starts a separate cache session.`);
+    console.log('▸ - Use deleteCache({ kvCacheKey: "your-session" }) to clear a saved session.');
     await deleteCache({ kvCacheKey: sharedCacheKey });
     await deleteCache({ kvCacheKey: isolatedCacheKey });
     await unloadModel({ modelId, clearStorage: false });
@@ -98,9 +102,9 @@ catch (error) {
             await unloadModel({ modelId, clearStorage: false });
         }
         catch (cleanupError) {
-            console.warn("⚠️ cleanup failed:", cleanupError);
+            console.warn("✖ cleanup failed:", cleanupError);
         }
     }
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

package/dist/examples/llamacpp-dynamic-tools.js

@@ -76,7 +76,7 @@ async function chatTurn({ modelId, kvCache, history, tools }) {
     })();
     const toolEventsTask = (async () => {
         for await (const evt of result.toolCallStream) {
-            console.log(`\n→ tool call: ${evt.call.name}(${JSON.stringify(evt.call.arguments)})`);
+            console.log(`\n▸ tool call: ${evt.call.name}(${JSON.stringify(evt.call.arguments)})`);
         }
     })();
     await Promise.all([tokensTask, toolEventsTask]);
@@ -91,7 +91,7 @@ async function chatTurn({ modelId, kvCache, history, tools }) {
         if (schema) {
             const parsed = schema.safeParse(call.arguments);
             if (!parsed.success) {
-                console.warn(`   ✗ validation failed for ${call.name}:`, parsed.error);
+                console.log(`✖ validation failed for ${call.name}:`, parsed.error);
             }
         }
     }
@@ -110,9 +110,15 @@ async function main() {
             tools: true,
             toolsMode: "dynamic",
         },
-        onProgress: (progress) => console.log(`Loading: ${progress.percentage.toFixed(1)}%`),
+        onProgress: (p) => {
+            const mb = (n) => (n / 1e6).toFixed(1);
+            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");
+        },
     });
-    console.log(`✅ Model loaded: ${modelId}`);
+    console.log(`▸ Model loaded: ${modelId}`);
     const kvCache = `dynamic-tools-${Date.now()}`;
     const history = [
         {
@@ -123,7 +129,7 @@ async function main() {
     ];
     // Turn 1 — only weather tools available.
     history.push({ role: "user", content: "What's the weather in Tokyo?" });
-    console.log("\n🤖 Turn 1 (tools=weather):\n");
+    console.log("\n▸ Turn 1 (tools=weather):\n");
     await chatTurn({ modelId, kvCache, history, tools: weatherTools });
     // Turn 2 — same session, swap to horoscope tools. Dynamic mode lets the
     // model see a different tool set without invalidating the kv-cache.
@@ -131,16 +137,16 @@ async function main() {
         role: "user",
         content: "Now check my horoscope for Aquarius.",
     });
-    console.log("\n\n🤖 Turn 2 (tools=horoscope):\n");
+    console.log("\n\n▸ Turn 2 (tools=horoscope):\n");
     await chatTurn({ modelId, kvCache, history, tools: horoscopeTools });
     // Turn 3 — swap to a parameterless tool to confirm empty-arg flows work.
     history.push({ role: "user", content: "What's today's date?" });
-    console.log("\n\n🤖 Turn 3 (tools=date):\n");
+    console.log("\n\n▸ Turn 3 (tools=date):\n");
     await chatTurn({ modelId, kvCache, history, tools: dateTools });
-    console.log("\n\n🎉 Done.");
+    console.log("\n\n▸ Done.");
     await unloadModel({ modelId, clearStorage: false });
 }
 main().catch((err) => {
-    console.error("❌ Error:", err);
+    console.error("✖", err);
     process.exit(1);
 });