Merge 48fc5eb30e into a0e917e6f5

nvidia/llama-cpp/README.md

@@ -1,7 +1,6 @@
 # Run models with llama.cpp on DGX Spark
 
-> Build llama.cpp with CUDA and serve models via an OpenAI-compatible API (Nemotron 3 Nano Omni as example)
-
+> Build llama.cpp with CUDA and serve models via an OpenAI-compatible API
 
 ## Table of Contents
 

nvidia/sglang/README.md

@@ -39,9 +39,9 @@ vision-language tasks using models like DeepSeek-V2-Lite.
 - NVIDIA Spark device with Blackwell architecture
 - Docker Engine installed and running: `docker --version`
 - NVIDIA GPU drivers installed: `nvidia-smi`
-- NVIDIA Container Toolkit configured: `docker run --rm --gpus all lmsysorg/sglang@sha256:ceaf8b16e02d165143633ac228bbb994a05fe77d7e0526cf035ae4bbf4eacc36 nvidia-smi`
+- NVIDIA Container Toolkit configured: `docker run --rm --gpus all lmsysorg/sglang:latest-cu130 nvidia-smi`
 - Sufficient disk space (>20GB available): `df -h`
-- Network connectivity for pulling containers: `docker pull lmsysorg/sglang@sha256:ceaf8b16e02d165143633ac228bbb994a05fe77d7e0526cf035ae4bbf4eacc36`
+- Network connectivity for pulling containers: `docker pull lmsysorg/sglang:latest-cu130`
 
 ## Ancillary files
 
@@ -103,7 +103,7 @@ docker --version
 nvidia-smi
 
 ## Verify Docker GPU support
-docker run --rm --gpus all lmsysorg/sglang@sha256:ceaf8b16e02d165143633ac228bbb994a05fe77d7e0526cf035ae4bbf4eacc36 nvidia-smi
+docker run --rm --gpus all lmsysorg/sglang:latest-cu130 nvidia-smi
 
 ## Check available disk space
 df -h /
@@ -124,7 +124,7 @@ several minutes depending on your network connection.
 
 ```bash
 ## Pull the SGLang container
-docker pull lmsysorg/sglang@sha256:ceaf8b16e02d165143633ac228bbb994a05fe77d7e0526cf035ae4bbf4eacc36
+docker pull lmsysorg/sglang:latest-cu130
 
 ## Verify the image was downloaded
 docker images | grep sglang
@@ -140,7 +140,7 @@ server inside the container, exposing it on port 30000 for client connections.
 docker run --gpus all -it --rm \
   -p 30000:30000 \
   -v /tmp:/tmp \
-  lmsysorg/sglang@sha256:ceaf8b16e02d165143633ac228bbb994a05fe77d7e0526cf035ae4bbf4eacc36 \
+  lmsysorg/sglang:latest-cu130 \
   bash
 ```
 
@@ -237,7 +237,7 @@ docker ps | grep sglang | awk '{print $1}' | xargs docker stop
 docker container prune -f
 
 ## Remove SGLang images (optional)
-docker rmi lmsysorg/sglang@sha256:ceaf8b16e02d165143633ac228bbb994a05fe77d7e0526cf035ae4bbf4eacc36
+docker rmi lmsysorg/sglang:latest-cu130
 ```
 
 ## Step 10. Next steps

nvidia/station-brev/README.md

@@ -52,21 +52,18 @@ You will also need the following:
 
 ## Step 1. Log in to Brev
 
-Go to the [Brev UI](https://brev.nvidia.com), log in, and confirm you’re in the correct org (by clicking the org button on the top right-hand side of the page). Once logged in, go to the [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute) section under the "GPU" tab in the main navigation.
+Go to the [Brev UI](https://brev.nvidia.com), log in, and confirm you’re in the correct org (by clicking the org button on the top right hand side of the page). Once logged in, go to the [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute) section under the "GPU" tab in the main navigation.
 
 Click the “Register Compute” button and follow the instructions in the pop-up window.
 
-## Step 2. Complete Pop-up Instructions
+## Step 2. Complete Popup Instructions
 
 * Install the Brev CLI
 * Configure your compute
     * Add a name for compute
-    * To configure SSH, ensure the “Enable SSH access” toggle is on
+    * To configure ssh, ensure the “Enable SSH access” toggle is on
 * Run the registration command
 
-> [!IMPORTANT]
-> Run the Brev CLI install command **without `sudo`**. Prefixing the installer with `sudo` writes the `brev` binary into root's home directory, which is not on your user shell's `PATH` — the next command will fail with `brev: command not found`. Copy the install command from the pop-up and run it as your normal user.
-
 ## Step 3. Follow Registration Flow
 
 In the CLI, you’ll be walked through registration. Go through the flow until registration is complete.
@@ -83,14 +80,10 @@ Your DGX Station is now integrated into Brev as a secure, remotely accessible GP
 
 Now that your hardware is connected, you can:
 
-* **Access your machine from anywhere:** Open the [Brev UI](https://brev.nvidia.com) and launch a session from [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute).
-* **Share access with others:** Invite teammates to your DGX Station from the Brev UI:
-    * Go to the [Brev UI](https://brev.nvidia.com) and open [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute).
-    * Find your DGX Station in the list and open the row's three-dot (⋯) menu.
-    * Select **Share Access**.
-    * Enter the email address of the person you want to share with.
-    * Choose their role / permission level.
-    * Confirm to send the invitation.
+* **Share Access Anywhere:** Access your machine from anywhere and share access with others through the Brev UI by:
+    * Adding the user to your [Team](https://brev.nvidia.com/org/team)
+    * Navigating to your instance in the [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute) section
+    * In **SSH Access** section of the instance, search for the user you wish to add and click **Modify Access** to enable access
 
 ## Step 6. Cleanup
 
@@ -105,7 +98,7 @@ brev deregister
 In the UI:
 * Go to the [Brev UI](https://brev.nvidia.com)
 * Navigate to the section listing “GPU Environments” and look under “Registered Compute”
-* Click the “Remove” menu item on the device you wish to delete from Brev.
+* Click the “Remove” menu item on the DGX Station you wish to delete from Brev.
 * Confirm your selection.
 
 ## Troubleshooting

nvidia/station-brev/endpoint-test.yaml

@@ -82,21 +82,18 @@ spec:
       content: |
         # Step 1. Log in to Brev
         
-        Go to the [Brev UI](https://brev.nvidia.com), log in, and confirm you’re in the correct org (by clicking the org button on the top right-hand side of the page). Once logged in, go to the [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute) section under the "GPU" tab in the main navigation.
+        Go to the [Brev UI](https://brev.nvidia.com), log in, and confirm you’re in the correct org (by clicking the org button on the top right hand side of the page). Once logged in, go to the [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute) section under the "GPU" tab in the main navigation.
         
         Click the “Register Compute” button and follow the instructions in the pop-up window.
         
-        # Step 2. Complete Pop-up Instructions
+        # Step 2. Complete Popup Instructions
         
         * Install the Brev CLI
         * Configure your compute
             * Add a name for compute
-            * To configure SSH, ensure the “Enable SSH access” toggle is on
+            * To configure ssh, ensure the “Enable SSH access” toggle is on
         * Run the registration command
         
-        > [!IMPORTANT]
-        > Run the Brev CLI install command **without `sudo`**. Prefixing the installer with `sudo` writes the `brev` binary into root's home directory, which is not on your user shell's `PATH` — the next command will fail with `brev: command not found`. Copy the install command from the pop-up and run it as your normal user.
-        
         # Step 3. Follow Registration Flow
         
         In the CLI, you’ll be walked through registration. Go through the flow until registration is complete.
@@ -113,14 +110,10 @@ spec:
         
         Now that your hardware is connected, you can:
         
-        * **Access your machine from anywhere:** Open the [Brev UI](https://brev.nvidia.com) and launch a session from [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute).
-        * **Share access with others:** Invite teammates to your DGX Station from the Brev UI:
-            * Go to the [Brev UI](https://brev.nvidia.com) and open [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute).
-            * Find your DGX Station in the list and open the row's three-dot (⋯) menu.
-            * Select **Share Access**.
-            * Enter the email address of the person you want to share with.
-            * Choose their role / permission level.
-            * Confirm to send the invitation.
+        * **Share Access Anywhere:** Access your machine from anywhere and share access with others through the Brev UI by:
+            * Adding the user to your [Team](https://brev.nvidia.com/org/team)
+            * Navigating to your instance in the [Registered Compute](https://brev.nvidia.com/org/environments?tab=registered-compute) section
+            * In **SSH Access** section of the instance, search for the user you wish to add and click **Modify Access** to enable access
         
         # Step 6. Cleanup
         
@@ -135,7 +128,7 @@ spec:
         In the UI:
         * Go to the [Brev UI](https://brev.nvidia.com)
         * Navigate to the section listing “GPU Environments” and look under “Registered Compute”
-        * Click the “Remove” menu item on the device you wish to delete from Brev.
+        * Click the “Remove” menu item on the DGX Station you wish to delete from Brev.
         * Confirm your selection.
         
       

nvidia/station-nanochat/endpoint-production.yaml

@@ -107,7 +107,7 @@ spec:
         
         # Time & risk
         
-        - **Estimated time:** ~30 minutes for setup. Full d24 training takes on the order of 16+ hours on a single GB300 Ultra.
+        - **Estimated time:** ~30 minutes for setup. Full d24 training takes on the order of 12+ hours on a single GB300 Ultra.
         - **Risk level:** Medium
           - Large downloads (FineWeb) can be slow; ensure stable network and disk space.
           - API keys (W&B, HF) must be set or `launch.sh` will exit immediately.
@@ -184,7 +184,7 @@ spec:
         3. **SFT** — downloads synthetic identity conversations, fine-tunes for chat
         4. **Report generation** — produces `report.md` with metrics and samples
         
-        Training on a single GB300 Ultra takes on the order of 16+ hours for the full d24 run.
+        Training on a single GB300 Ultra takes on the order of 12+ hours for the full d24 run.
         
         # Step 4. Monitor training
         

nvidia/txt2kg/assets/frontend/components/embeddings-generator.tsx

@@ -226,6 +226,8 @@ export function EmbeddingsGenerator({ showTripleExtraction = false }: Embeddings
           const model = JSON.parse(selectedModel);
           if (model.provider === "ollama") {
             processingMethod = `Ollama ${model.model || 'qwen3:1.7b'}`;
+          } else if (model.provider === "vllm") {
+            processingMethod = `vLLM ${model.model || 'local model'}`;
           } else if (model.id?.startsWith("nvidia-")) {
             processingMethod = 'NVIDIA Nemotron';
           }
@@ -242,14 +244,36 @@ export function EmbeddingsGenerator({ showTripleExtraction = false }: Embeddings
       
       // Call processDocuments with the selected document IDs and processing options
       const useGraphTransformer = useLangChain && langChainMethod === 'graphtransformer';
-      await processDocuments(selectedDocs, {
+      const processingOptions: Parameters<typeof processDocuments>[1] = {
         useLangChain,
         useGraphTransformer,
         promptConfigs: options || undefined,
         chunkSize: options?.chunkSize,
         overlapSize: options?.overlapSize,
         chunkingMethod: options?.chunkingMethod
-      });
+      };
+
+      try {
+        const selectedModel = localStorage.getItem("selectedModel");
+        if (selectedModel) {
+          const model = JSON.parse(selectedModel);
+          if (model.provider === "ollama") {
+            processingOptions.llmProvider = "ollama";
+            processingOptions.ollamaModel = model.model || "qwen3:1.7b";
+            processingOptions.ollamaBaseUrl = model.baseURL || "http://localhost:11434/v1";
+          } else if (model.provider === "vllm") {
+            processingOptions.llmProvider = "vllm";
+            processingOptions.vllmModel = model.model;
+            processingOptions.vllmBaseUrl = model.baseURL || "http://localhost:8001/v1";
+          } else if (model.provider === "nvidia" || model.id?.startsWith("nvidia-")) {
+            processingOptions.llmProvider = "nvidia";
+          }
+        }
+      } catch (e) {
+        console.log("Could not parse selected model, using default extraction provider");
+      }
+
+      await processDocuments(selectedDocs, processingOptions);
       
       // Navigate to the edit tab after processing is complete
       setTimeout(() => {
@@ -1265,4 +1289,4 @@ function InfoIcon(props: React.SVGProps<SVGSVGElement>) {
       <path d="M12 8h.01" />
     </svg>
   )
-} 
+}

nvidia/txt2kg/assets/frontend/components/model-selector.tsx

@@ -151,7 +151,9 @@ export function ModelSelector() {
       
       // Default to first available local model (vLLM or Ollama)
       const localModel = availableModels.find(m => m.provider === "vllm" || m.provider === "ollama")
-      setSelectedModel(localModel || availableModels[0])
+      const defaultModel = localModel || availableModels[0]
+      setSelectedModel(defaultModel)
+      localStorage.setItem("selectedModel", JSON.stringify(defaultModel))
     }
     
     setIsLoading(false)

nvidia/txt2kg/assets/frontend/contexts/document-context.tsx

@@ -49,7 +49,7 @@ export type Document = {
   }
 }
 
-export type LLMProvider = 'nvidia' | 'ollama';
+export type LLMProvider = 'nvidia' | 'ollama' | 'vllm';
 
 export type ProcessingOptions = {
   useLangChain?: boolean;
@@ -58,6 +58,8 @@ export type ProcessingOptions = {
   llmProvider?: LLMProvider;
   ollamaModel?: string;
   ollamaBaseUrl?: string;
+  vllmModel?: string;
+  vllmBaseUrl?: string;
   chunkSize?: number;
   overlapSize?: number;
   chunkingMethod?: 'optimized' | 'pyg';
@@ -451,6 +453,8 @@ export function DocumentProvider({ children }: { children: React.ReactNode }) {
       llmProvider = 'ollama',
       ollamaModel = 'qwen3:1.7b',
       ollamaBaseUrl = 'http://localhost:11434/v1',
+      vllmModel,
+      vllmBaseUrl,
       chunkSize = 64000,
       overlapSize = 2000,
       chunkingMethod = 'optimized'
@@ -460,6 +464,8 @@ export function DocumentProvider({ children }: { children: React.ReactNode }) {
       llmProvider,
       ollamaModel,
       ollamaBaseUrl,
+      vllmModel,
+      vllmBaseUrl,
       chunkSize,
       overlapSize,
       chunkingMethod
@@ -485,6 +491,8 @@ export function DocumentProvider({ children }: { children: React.ReactNode }) {
       llmProvider?: LLMProvider;
       ollamaModel?: string;
       ollamaBaseUrl?: string;
+      vllmModel?: string;
+      vllmBaseUrl?: string;
       chunkSize?: number;
       overlapSize?: number;
       chunkingMethod?: 'optimized' | 'pyg';
@@ -673,6 +681,12 @@ export function DocumentProvider({ children }: { children: React.ReactNode }) {
                 if (llmOptions.ollamaBaseUrl) {
                   requestBody.ollamaBaseUrl = llmOptions.ollamaBaseUrl;
                 }
+                if (llmOptions.vllmModel) {
+                  requestBody.vllmModel = llmOptions.vllmModel;
+                }
+                if (llmOptions.vllmBaseUrl) {
+                  requestBody.vllmBaseUrl = llmOptions.vllmBaseUrl;
+                }
               }
               
               // Add prompt configs if available
@@ -1273,4 +1287,4 @@ export function useDocuments() {
     throw new Error("useDocuments must be used within a DocumentProvider")
   }
   return context
-}
+}

nvidia/txt2kg/assets/frontend/lib/langchain-service.ts

@@ -290,9 +290,6 @@ export class LangChainService {
         configuration: {
           baseURL: baseURL,
           timeout: 120000, // 2 minute timeout for vLLM inference
-        },
-        modelKwargs: {
-          "response_format": { "type": "text" }
         }
       });
       
@@ -320,4 +317,4 @@ export class LangChainService {
 }
 
 // Export a singleton instance for convenience
-export const langChainService = LangChainService.getInstance(); 
+export const langChainService = LangChainService.getInstance();

nvidia/txt2kg/assets/frontend/lib/qdrant.ts

@@ -153,6 +153,16 @@ export class QdrantService {
         return true;
       }
 
+      const collectionsResponse = await fetch(`${this.hostUrl}/collections`, {
+        method: 'GET'
+      });
+
+      if (collectionsResponse.ok) {
+        console.log(`Qdrant server is reachable`);
+        this.isQdrantRunningCheck = false;
+        return true;
+      }
+
       console.log('Qdrant health check failed - server might not be running');
       this.isQdrantRunningCheck = false;
       return false;
@@ -534,6 +544,21 @@ export class QdrantService {
   public async getStats(): Promise<any> {
     try {
       console.log('Getting stats from Qdrant...');
+      const isRunning = await this.isQdrantRunning();
+      if (!isRunning) {
+        return {
+          totalVectorCount: 0,
+          source: 'qdrant',
+          httpHealthy: false,
+          url: this.hostUrl,
+          error: `Qdrant is not reachable at ${this.hostUrl}. Start vector search with ./start.sh --vector-search if you need Vector DB features.`
+        };
+      }
+
+      if (!this.initialized) {
+        await this.initialize();
+      }
+
       const response = await this.makeRequest(`/collections/${this.collectionName}`, 'GET');
 
       if (response && response.result) {
@@ -554,17 +579,19 @@ export class QdrantService {
         console.log(`Qdrant stats request failed`);
         return {
           totalVectorCount: 0,
-          source: 'error',
-          httpHealthy: false,
-          error: 'Failed to get stats'
+          source: 'qdrant',
+          httpHealthy: true,
+          url: this.hostUrl,
+          error: `Qdrant is reachable, but collection '${this.collectionName}' is not available.`
         };
       }
     } catch (error) {
       console.log('Qdrant connection failed - server may not be running');
       return {
         totalVectorCount: 0,
-        source: 'error',
+        source: 'qdrant',
         httpHealthy: false,
+        url: this.hostUrl,
         error: error instanceof Error ? error.message : String(error)
       };
     }

nvidia/txt2kg/assets/start.sh

@@ -176,6 +176,10 @@ if [ "$USE_VECTOR_SEARCH" = true ]; then
   echo "  • Qdrant: http://localhost:6333"
   echo "  • Sentence Transformers: http://localhost:8000"
   echo ""
+else
+  echo "Vector Search Services: disabled"
+  echo "  • Start with --vector-search to enable Vector DB status and embedding search"
+  echo ""
 fi
 
 echo "Next steps:"

nvidia/vibe-coding/README.md

@@ -171,10 +171,12 @@ Add additional model entries for any other Ollama models you wish to host remote
 
 | Symptom | Cause | Fix |
 |---------|-------|-----|
-|Ollama not starting|GPU drivers may not be installed correctly|Run `nvidia-smi` in the terminal. If the command fails check DGX Dashboard for updates to your DGX Spark.|
-|Continue can't connect over the network|Port 11434 may not be open or accessible|Run command `ss -tuln \| grep 11434`. If the output does not reflect ` tcp   LISTEN 0      4096               *:11434            *:*  `, go back to step 2 and run the ufw command.|
-|Continue can't detect a locally running Ollama model|Configuration not properly set or detected|Check `OLLAMA_HOST` and `OLLAMA_ORIGINS` in `/etc/systemd/system/ollama.service.d/override.conf` file. If `OLLAMA_HOST` and `OLLAMA_ORIGINS` are set correctly, add these lines to your `~/.bashrc` file.|
-|High memory usage|Model size too big|Confirm no other large models or containers are running with `nvidia-smi`. Use smaller models such as `gpt-oss:20b` for lightweight usage.|
+| **WiFi connection drops or becomes unreachable** (especially in headless mode) | Aggressive WiFi power-saving settings in NetworkManager | Edit `/etc/NetworkManager/conf.d/default-wifi-powersave-on.conf`, set `wifi.powersave = 2`, and run `sudo systemctl restart NetworkManager`. |
+| **Random reboots and "00" error code on the display** | Watchdog timer module (`sbsa_gwdt`) not loaded | Add `sbsa_gwdt` to `/etc/modules-load.d/watchdog.conf` and reboot to ensure the hardware watchdog is correctly managed by the kernel. |
+| Ollama not starting | GPU drivers may not be installed correctly | Run `nvidia-smi` in the terminal. If the command fails check DGX Dashboard for updates to your DGX Spark. |
+| Continue can't connect over the network | Port 11434 may not be open or accessible | Run command `ss -tuln \| grep 11434`. If the output does not reflect `tcp LISTEN 0 4096 *:11434 *:*`, go back to step 2 and run the ufw command. |
+| Continue can't detect a locally running Ollama model | Configuration not properly set or detected | Check `OLLAMA_HOST` and `OLLAMA_ORIGINS` in `/etc/systemd/system/ollama.service.d/override.conf` file. If `OLLAMA_HOST` and `OLLAMA_ORIGINS` are set correctly, add these lines to your `~/.bashrc` file. |
+| High memory usage | Model size too big | Confirm no other large models or containers are running with `nvidia-smi`. Use smaller models such as `gpt-oss:20b` for lightweight usage. |
 
 > [!NOTE]
 > DGX Spark uses a Unified Memory Architecture (UMA), which enables dynamic memory sharing between the GPU and CPU.