{"id":717,"date":"2025-07-08T13:35:51","date_gmt":"2025-07-08T13:35:51","guid":{"rendered":"https:\/\/blog.aetherix.com\/?p=717"},"modified":"2026-03-27T13:13:18","modified_gmt":"2026-03-27T13:13:18","slug":"jetson-generative-ai-llamaspeak","status":"publish","type":"post","link":"https:\/\/blog.openzeka.com\/en\/jetson-generative-ai-llamaspeak\/","title":{"rendered":"Jetson Generative AI \u2013 LlamaSpeak"},"content":{"rendered":"

Vision wasn’t the only modality transformed by Transformers\u2014**LlamaSpeak** brings the power of large language models to spoken conversations, streaming speech-to-text (ASR), intelligent response generation, and text-to-speech (TTS) back out in real time<\/strong>\u00a0on your Jetson device.<\/p>\n

In this article you’ll learn how to run NanoLLM’s WebChat agent<\/strong> (nicknamed LlamaSpeak<\/strong>) on Jetson using NVIDIA TensorRT\/MLC and Riva Speech Skills.<\/p>\n<\/div>

Requirements<\/h3><\/h2><\/div>\n
\n\n\n\n\n\n\n\n\n\n
\n
\n
Hardware \/ Software<\/div>\n<\/div>\n<\/th>\n
\n
\n
Notes<\/div>\n<\/div>\n<\/th>\n<\/tr>\n<\/thead>\n
\n
\n
Jetson AI Kit \/ Dev Kit<\/strong><\/div>\n<\/div>\n<\/td>\n
\n
\n
Orin AGX \/ Orin NX recommended for best latency<\/div>\n<\/div>\n<\/td>\n<\/tr>\n
\n
\n
JetPack 6 (L4T r36.x)<\/strong><\/div>\n<\/div>\n<\/td>\n
\n
\n
Needed for latest pre-built containers<\/div>\n<\/div>\n<\/td>\n<\/tr>\n
\n
\n
USB microphone & speakers \/ headset<\/strong><\/div>\n<\/div>\n<\/td>\n
\n
\n
Confirm with `arecord -l`<\/div>\n<\/div>\n<\/td>\n<\/tr>\n
\n
\n
Riva Speech Skills 2.15+<\/strong><\/div>\n<\/div>\n<\/td>\n
\n
\n
Provides low-latency ASR engine<\/div>\n<\/div>\n<\/td>\n<\/tr>\n
\n
\n
Hugging Face token<\/strong><\/div>\n<\/div>\n<\/td>\n
\n
\n
Needed for gated Meta-Llama weights<\/div>\n<\/div>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n

Obtaining Your Hugging Face Token<\/h2><\/div>

To download the gated Llama checkpoints you’ll need a personal access token (PAT) from Hugging Face:<\/p>\n

1. Create \/ sign into your account at https:\/\/huggingface.co<\/div>\n
2. Click your avatar \u25b8 Settings<\/strong> \u25b8 Access Tokens<\/strong><\/div>\n
3. Press New Token<\/strong>, select _”Read”_ scope, name it (e.g., `jetson-llamaspeak`), and click Generate<\/strong><\/div>\n
4. Copy the token string that starts with `hf_` \u2014 you’ll export it in Step 7 below:<\/div>\n<\/div>