cookbook: add Cosmos3-Edge TensorRT-Edge-LLM reasoner and policy notebooks

README.md

@@ -87,6 +87,7 @@ This repository is the home of the models: everything for exploring, running, an
   <tr><td><b>Explore runnable notebooks</b></td><td><a href="cookbooks/cosmos3/README.md">Cookbooks</a></td><td>browse</td></tr>
   <tr><td><b>Evaluate a model</b></td><td><a href="evaluation/">Evaluation suites</a> — PAIBench, Physics-IQ, VLMEvalKit</td><td>hours</td></tr>
   <tr><td><b>Check latency &amp; throughput</b></td><td><a href="inference_benchmarks.md">Benchmarks</a></td><td>browse</td></tr>
+  <tr><td><b>Run Cosmos3-Edge on device</b></td><td><a href="cookbooks/cosmos3/reasoner/run_with_trt_edge_llm.ipynb">Reasoner</a> and <a href="cookbooks/cosmos3/generator/action/run_policy_with_trt_edge_llm.ipynb">policy</a> with TensorRT-Edge-LLM</td><td>setup</td></tr>
 </table>
 
 ## Generate your first video
@@ -164,7 +165,7 @@ Training, optimization, and deployment tooling lives in [Cosmos Framework](https
 | [Cosmos Curator](https://github.com/NVIDIA/cosmos-curator) | Distributed data curation: processing, annotation, filtering, dedup |
 | [Cosmos Evaluator](https://github.com/NVIDIA/cosmos-evaluator) | Automated evaluation system for world generation & reasoning outputs |
 
-Cosmos 3 runs on Diffusers, Transformers, vLLM, vLLM-Omni, SGLang, TensorRT-LLM, and NIM — pick a backend in the [environment setup guide](cookbooks/cosmos3/README.md).
+Cosmos 3 runs on Diffusers, Transformers, vLLM, vLLM-Omni, SGLang, TensorRT-LLM, TensorRT-Edge-LLM, and NIM — pick a backend in the [environment setup guide](cookbooks/cosmos3/README.md).
 
 ## Community & contributing
 

cookbooks/cosmos3/README.md

@@ -10,6 +10,7 @@ backend you want to run and follow that one section.
 | [Diffusers](#diffusers) | Direct generation with `Cosmos3OmniPipeline` | Generator (Audiovisual) |
 | [TensorRT-LLM Generator](#tensorrt-llm-generator) | OpenAI-compatible VisualGen server (image/video/audio/action/transfer generation) | Generator (Audiovisual, Action, **Transfer**) |
 | [TensorRT-LLM Reasoner](#tensorrt-llm-reasoner) | OpenAI-compatible image/video reasoning server | Reasoner |
+| [TensorRT-Edge-LLM](#tensorrt-edge-llm) | On-device TensorRT runtime for Cosmos3-Edge reasoning and DROID policy | Reasoner (Edge), Generator (Action / Policy) |
 | [Transformers](#transformers) | Hugging Face Transformers inference | Reasoner |
 | [vLLM](#vllm) | OpenAI-compatible reasoning server (image/video understanding) | Reasoner |
 | [vLLM-Omni](#vllm-omni) | OpenAI-compatible generation server (image/video/audio/action/transfer) | Generator (Audiovisual, Action, **Transfer**) |
@@ -448,6 +449,42 @@ The server exposes `/health` and the OpenAI-compatible API at
 [Reasoner TensorRT-LLM notebook](reasoner/run_with_tensorrt_llm.ipynb) for image
 and video requests.
 
+## TensorRT-Edge-LLM
+
+On-device **TensorRT** path for **Cosmos3-Edge** multimodal reasoning and
+**Cosmos3-Edge-Policy-DROID** policy generation. This is a different stack from
+[TensorRT-LLM Reasoner](#tensorrt-llm-reasoner) / [Generator](#tensorrt-llm-generator):
+export ONNX on an x86 host, build TensorRT engines on the target (Jetson
+Thor/Orin, DRIVE, DGX Spark, or an x86 developer build), then run the C++
+inference binaries.
+
+Use [TensorRT-Edge-LLM](https://github.com/NVIDIA/TensorRT-Edge-LLM) **v0.11.0**.
+Complete the upstream
+[Installation](https://nvidia.github.io/TensorRT-Edge-LLM/user_guide/getting_started/installation.html)
+guide first (Python export package on the host; C++ runtime on the device).
+
+Policy examples also require building the C++ runtime with
+`-DBUILD_EXPERIMENTAL_MODELS=ON`.
+
+Set a workspace and the Edge checkpoints:
+
+```bash
+export EDGELLM_ROOT="${EDGELLM_ROOT:-$PWD/TensorRT-Edge-LLM}"
+export POLICY_CHECKPOINT=nvidia/Cosmos3-Edge-Policy-DROID
+export REASONING_CHECKPOINT=nvidia/Cosmos3-Edge
+export ONNX_DIR="${ONNX_DIR:-$HOME/tensorrt-edgellm-workspace/Cosmos3-Edge/onnx}"
+export ENGINE_DIR="${ENGINE_DIR:-$HOME/tensorrt-edgellm-workspace/Cosmos3-Edge/engines}"
+```
+
+Runnable cookbooks:
+
+- Reasoner: [`reasoner/run_with_trt_edge_llm.ipynb`](reasoner/run_with_trt_edge_llm.ipynb)
+- Policy: [`generator/action/run_policy_with_trt_edge_llm.ipynb`](generator/action/run_policy_with_trt_edge_llm.ipynb)
+
+Upstream reference:
+[Cosmos3-Edge VLA guide](https://nvidia.github.io/TensorRT-Edge-LLM/user_guide/examples/vla/cosmos3.html)
+(export → build → run for reasoning and policy).
+
 ## Transformers
 
 Local Python inference for the Cosmos3 Reasoner. This backend uses the

cookbooks/cosmos3/generator/action/README.md

@@ -24,13 +24,15 @@ The rest of this doc shows how to run these modes on selected embodiments direct
 - [Run with vLLM-Omni](#run-with-vllm-omni)
   - [Quickstart](#quickstart-3)
   - [Notebook walkthrough](#vllm-omni-notebook-walkthrough)
+- [Run with TensorRT-Edge-LLM](#run-with-tensorrt-edge-llm)
 - [Post-Train for Cosmos3-Nano-Policy-DROID](#post-train-for-cosmos3-nano-policy-droid)
 
 ## Overview
 
-Examples are shown across native PyTorch (Cosmos Framework), Diffusers,
-TensorRT-LLM, and vLLM-Omni. Every backend uses the sample
-assets under [`assets/`](./assets) and covers three tasks:
+Examples for native PyTorch (Cosmos Framework), Diffusers, TensorRT-LLM,
+and vLLM-Omni use the sample assets under [`assets/`](./assets) and cover
+the three tasks above. TensorRT-Edge-LLM covers the Edge policy task only,
+with the same sample assets.
 
 Environment setup for all backends is centralized in the shared
 [Cosmos3 cookbooks environment setup](../../README.md) guide; each backend below
@@ -292,6 +294,29 @@ write outputs under `outputs/cosmos3_action_vllm/`:
 - [`run_policy_with_vllm_omni.ipynb`](./run_policy_with_vllm_omni.ipynb) — policy
   inference for DROID through the async video API.
 
+## Run with TensorRT-Edge-LLM
+
+### Quickstart
+
+On-device TensorRT path for **Cosmos3-Edge-Policy-DROID**. Set up the shared
+[TensorRT-Edge-LLM](../../README.md#tensorrt-edge-llm) environment (upstream
+install; C++ runtime with `-DBUILD_EXPERIMENTAL_MODELS=ON`). Then follow
+[`run_policy_with_trt_edge_llm.ipynb`](./run_policy_with_trt_edge_llm.ipynb):
+export → `cosmos3_policy_build` → `cosmos3_policy_inference`.
+
+This is separate from workstation [vLLM-Omni](#run-with-vllm-omni) policy
+serving. Multimodal reasoning on the same Edge-LLM runtime is documented in
+[`../../reasoner/run_with_trt_edge_llm.ipynb`](../../reasoner/run_with_trt_edge_llm.ipynb).
+
+### Notebook walkthrough
+
+[`run_policy_with_trt_edge_llm.ipynb`](./run_policy_with_trt_edge_llm.ipynb)
+builds the same 640×540 DROID `concat_view` as the vLLM-Omni / SGLang policy
+notebooks (from `assets/droid_lerobot_example`), runs
+`cosmos3_policy_inference` with `--viewPoint concat_view`, and writes `action.json`
+(`[batch, chunk, action_dimension]`). Optional `--steps` / `--seed` match the
+upstream VLA guide.
+
 ## Post-Train for Cosmos3-Nano-Policy-DROID
 
 To reproduce our post-training recipe for [Cosmos3-Nano-Policy-DROID](https://huggingface.co/nvidia/Cosmos3-Nano-Policy-DROID), use the