Dynamo with TensorRT-LLM supports two methods for transferring KV cache in disaggregated serving: UCX (default) and NIXL (experimental). For detailed information and configuration instructions for each method, see the [KV cache transfer guide](./kv-cache-tranfer.md).
Dynamo with TensorRT-LLM supports two methods for transferring KV cache in disaggregated serving: UCX (default) and NIXL (experimental). For detailed information and configuration instructions for each method, see the [KV cache transfer guide](./kv-cache-tranfer.md).
## Request Migration
## Request Migration
You can enable [request migration](../../../docs/architecture/request_migration.md) to handle worker failures gracefully. Use the `--migration-limit` flag to specify how many times a request can be migrated to another worker:
You can enable [request migration](../../../docs/architecture/request_migration.md) to handle worker failures gracefully. Use the `--migration-limit` flag to specify how many times a request can be migrated to another worker:
...
@@ -281,3 +283,140 @@ NOTE: To send a request to a multi-node deployment, target the node which is run
...
@@ -281,3 +283,140 @@ NOTE: To send a request to a multi-node deployment, target the node which is run
To benchmark your deployment with GenAI-Perf, see this utility script, configuring the
To benchmark your deployment with GenAI-Perf, see this utility script, configuring the
`model` name and `host` based on your deployment: [perf.sh](../../../benchmarks/llm/perf.sh)
`model` name and `host` based on your deployment: [perf.sh](../../../benchmarks/llm/perf.sh)
## Multimodal support
TRTLLM supports multimodal models with dynamo. You can provide multimodal inputs in the following ways:
- By sending image URLs
- By providing paths to pre-computed embedding files
Please note that you should provide **either image URLs or embedding file paths** in a single request.
### Aggregated
Here are quick steps to launch Llama-4 Maverick BF16 in aggregated mode
{"id":"unknown-id","choices":[{"index":0,"message":{"content":"The image depicts a serene landscape featuring a large rock formation, likely El Capitan in Yosemite National Park, California. The scene is characterized by a winding road that curves from the bottom-right corner towards the center-left of the image, with a few rocks and trees lining its edge.\n\n**Key Features:**\n\n* **Rock Formation:** A prominent, tall, and flat-topped rock formation dominates the center of the image.\n* **Road:** A paved road winds its way through the landscape, curving from the bottom-right corner towards the center-left.\n* **Trees and Rocks:** Trees are visible on both sides of the road, with rocks scattered along the left side.\n* **Sky:** The sky above is blue, dotted with white clouds.\n* **Atmosphere:** The overall atmosphere of the","refusal":null,"tool_calls":null,"role":"assistant","function_call":null,"audio":null},"finish_reason":"stop","logprobs":null}],"created":1753322607,"model":"meta-llama/Llama-4-Maverick-17B-128E-Instruct","service_tier":null,"system_fingerprint":null,"object":"chat.completion","usage":null}
```
### Disaggregated
Here are quick steps to launch in disaggregated mode.
The following is an example of launching a model in disaggregated mode. While this example uses `Qwen/Qwen2-VL-7B-Instruct`, you can adapt it for other models by modifying the environment variables for the model path and engine configurations.
For a large model like `meta-llama/Llama-4-Maverick-17B-128E-Instruct`, a multi-node setup is required for disaggregated serving, while aggregated serving can run on a single node. This is because the model with a disaggregated configuration is too large to fit on a single node's GPUs. For instance, running this model in disaggregated mode requires a setup of 2 nodes with 8xH200 GPUs or 4 nodes with 4xGB200 GPUs.
In general, disaggregated serving can run on a single node, provided the model fits on the GPU. The multi-node requirement in this example is specific to the size and configuration of the `meta-llama/Llama-4-Maverick-17B-128E-Instruct` model.
To deploy `Llama-4-Maverick-17B-128E-Instruct` in disaggregated mode, you will need to follow the multi-node setup instructions, which can be found [here](multinode/multinode-multimodal-example.md).
### Using Pre-computed Embeddings (Experimental)
Dynamo with TensorRT-LLM supports providing pre-computed embeddings directly in an inference request. This bypasses the need for the model to process an image and generate embeddings itself, which is useful for performance optimization or when working with custom, pre-generated embeddings.
#### Enabling the Feature
This is an experimental feature that requires using a specific TensorRT-LLM commit.
To enable it build the dynamo container with the `--tensorrtllm-commit` flag, followed by the commit hash:
Once the container is built, you can send requests with paths to local embedding files.
-**Format:** Provide the embedding as part of the `messages` array, using the `image_url` content type.
-**URL:** The `url` field should contain the absolute or relative path to your embedding file on the local filesystem.
-**File Types:** Supported embedding file extensions are `.pt`, `.pth`, and `.bin`. Dynamo will automatically detect these extensions.
When a request with a supported embedding file is received, Dynamo will load the tensor from the file and pass it directly to the model for inference, skipping the image-to-embedding pipeline.
#### Example Request
Here is an example of how to send a request with a pre-computed embedding file.
SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
SPDX-License-Identifier: Apache-2.0
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->
# Example: Multi-node TRTLLM Workers with Dynamo on Slurm for multimodal models
This guide demonstrates how to deploy large multimodal models that require a multi-node setup. It builds on the general multi-node deployment process described in the main [multinode-examples.md](./multinode-examples.md) guide.
Before you begin, ensure you have completed the initial environment configuration by following the **Setup** section in that guide.
The following sections provide specific instructions for deploying `meta-llama/Llama-4-Maverick-17B-128E-Instruct`, including environment variable setup and launch commands. These steps can be adapted for other large multimodal models.
### Environment Variable Setup
Assuming you have already allocated your nodes via `salloc`, and are
inside an interactive shell on one of the allocated nodes, set the
To cleanup background `srun` processes launched by `srun_aggregated.sh` or
`srun_disaggregated.sh`, you can run:
```bash
pkill srun
```
## Known Issues
- Loading `meta-llama/Llama-4-Maverick-17B-128E-Instruct` with 8 nodes of H100 with TP=16 is not posssible due to Llama4 Maverick has a config `"num_attention_heads": 40` , trtllm engine asserts on assert `self.num_heads % tp_size == 0` causing the engine to crash on model loading.
