sarman/dgx-spark-playbooks
1
0
Fork 0
mirror of https://github.com/NVIDIA/dgx-spark-playbooks.git synced 2026-04-25 03:13:53 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

feat(langchain): upgrade to Llama 3.3 Nemotron Super 49B

Browse Source 
- Update LangChain service to use Llama 3.3 Nemotron Super 49B v1.5
- Adjust temperature to 0.6 for better response quality
- Increase timeout to 120s for larger model
- Add top_p, frequency_penalty, and presence_penalty parameters
- Remove deprecated response_format configuration
This commit is contained in:
Santosh Bhavani 2025-10-19 20:57:03 -07:00
parent 5be2ad78bf
commit 12c4777eae
1 changed files with 6 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
nvidia/txt2kg/assets/frontend/lib/langchain-service.ts
View File

@ -29,7 +29,7 @@ export class LangChainService {
temperature?: number; temperature?: number;
maxTokens?: number; maxTokens?: number;
}): Promise<ChatOpenAI> { }): Promise<ChatOpenAI> {
const modelId = "nvdev/nvidia/llama-3.1-nemotron-70b-instruct"; const modelId = "nvidia/llama-3.3-nemotron-super-49b-v1.5";
const cacheKey = `nemotron-${options?.temperature || 0.7}-${options?.maxTokens || 8192}`; const cacheKey = `nemotron-${options?.temperature || 0.7}-${options?.maxTokens || 8192}`;
console.log(`Requesting Nemotron model (cacheKey: ${cacheKey})`); console.log(`Requesting Nemotron model (cacheKey: ${cacheKey})`);
@ -73,15 +73,17 @@ export class LangChainService {
// Create a new ChatOpenAI instance // Create a new ChatOpenAI instance
const model = new ChatOpenAI({ const model = new ChatOpenAI({
modelName: modelId, modelName: modelId,
temperature: options?.temperature || 0.7, temperature: options?.temperature || 0.6,
maxTokens: options?.maxTokens || 8192, maxTokens: options?.maxTokens || 8192,
openAIApiKey: apiKey, openAIApiKey: apiKey,
configuration: { configuration: {
baseURL: "https://integrate.api.nvidia.com/v1", baseURL: "https://integrate.api.nvidia.com/v1",
timeout: 60000, // 60 second timeout timeout: 120000, // 120 second timeout for larger model
}, },
modelKwargs: { modelKwargs: {
"response_format": { "type": "text" } top_p: 0.95,
frequency_penalty: 0,
presence_penalty: 0
} }
}); });