---
title: "llama-3.2-nemoretriever-1b-vlm-embed-v1"
publisher: "nvidia"
type: "endpoint"
updated: "2025-06-25T14:17:57.052Z"
description: "Multimodal question-answer retrieval representing user queries as text and documents as images."
canonical: "https://build.nvidia.com/nvidia/llama-3_2-nemoretriever-1b-vlm-embed-v1"
---

## **Model Overview**

### **Description**

The Llama 3.2 NeMo Retriever Multimodal Embedding 1B model is optimized for **multimodal** question-answering retrieval. The model can embed 'documents' in the form of image, text, or image and text combined. Documents can be retrieved given a user query in text form. The model supports images containing text, tables, charts, and infographics. This model was evaluated on [ViDoRe V1](https://huggingface.co/spaces/vidore/vidore-leaderboard) and two internal multimodal retrieval benchmarks.

An embedding model is a crucial component of a retrieval system, because it transforms information into dense vector representations. An embedding model is typically a transformer encoder that processes tokens of input (text or image) (for example: question, passage) to output an embedding. The Llama 3.2 NeMo Retriever Multimodal Embedding 1B model is a combined language model and vision model.

The Llama 3.2 NeMo Retriever Multimodal Embedding 1B model is a part of the NVIDIA NeMo Retriever collection of NIM, which provides state-of-the-art, commercially-ready models and microservices optimized for the lowest latency and highest throughput. It features a production-ready information retrieval pipeline with enterprise support. The models that form the core of this solution have been trained using responsibly selected, auditable data sources. With multiple pre-trained models available as starting points, developers can readily customize them for domain-specific use cases, such as information technology, human resource help assistants, and research & development research assistants.

This model is ready for commercial use.

### **License/Terms of use**

GOVERNING TERMS: Access to this trial is governed by the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf). Use of this model is governed by the [NVIDIA Community Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-community-models-license/). 

ADDITIONAL INFORMATION: [Llama 3.2 Community License Agreement](https://www.llama.com/llama3_2/license/). Built with Llama.

### Deployment Geography:
Global <br>

### Use Case: <br>
The Llama 3.2 NeMo Retriever Multimodal Embedding 1B model is most suitable for users who want to build a multimodal question-and-answer application over a large corpus, leveraging the latest dense retrieval technologies.

### Release Date:  <br>
**Build.Nvidia.com:** May 20, 2025 via https://build.nvidia.com/nvidia/llama-3_2-nemoretriever-1b-vlm-embed-v1 <br>
**NGC:** May 20, 2025 <br>

**You are responsible for ensuring that your use of NVIDIA AI Foundation Models complies with all applicable laws.**

### **Model Architecture**

**Architecture Type:** Transformer<br>
**Network Architecture:** Fine-tuned MultiModal Llama 3.2 1B Retriever<br>

This NeMo Retriever embedding model is a transformer encoder. It is a fine-tuned version of Llama 3.2 1B with SigLip2 400M, with 16 layers and an embedding size of 2048, which is trained on public datasets. Embedding models for text retrieval are typically trained using a bi-encoder architecture. This involves encoding a pair of query and document independently using the embedding model. Contrastive learning is used in this model to maximize the similarity between the query and the document that contains the answer, while minimizing the similarity between the query and sampled negative documents not useful to answer the question.

The vision-language model encoder incorporates key innovations from NVIDIA, including [Eagle 2 work](https://arxiv.org/abs/2501.14818) and [nemoretriever-parse](https://build.nvidia.com/nvidia/nemoretriever-parse), which use a tiling-based VLM architecture. This architecture, available on [Hugging Face](https://huggingface.co/collections/nvidia/eagle-2-6764ba887fa1ef387f7df067), significantly enhances multimodal understanding through its dynamic tiling and mixture of vision encoders design. It particularly improves performance on tasks that involve high-resolution images and complex visual content.

### **Input**

| Property | Query | Document |
|----------|-------|----------|
| Input Type | Text | Text \| Image |
| Input Format | List of strings | List of strings \| List of Images |
| Input Parameter | 1D | 1D |
| Other Properties | The model's maximum context length is 8192 tokens. Texts longer than maximum length must either be chunked or truncated. | The model's maximum context length is 8192 tokens. Texts longer than maximum length must either be chunked or truncated.<br>Images must be `8192 x 16384` or `16384 x 8192` and less than 25MB. They are resized automatically by the NIM. |

### **Output**

**Output Type:** Floats<br>
**Output Format:** List of float arrays<br>
**Output:** Model outputs embedding vectors of maximum dimension 2048 for each input.<br>
**Other Properties Related to Output:** N/A<br>

Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA's hardware (such as GPU cores) and software frameworks (such as CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions.

### **Software Integration**

**Runtime Engine:** NeMo Retriever Embedding NIM<br>
**Supported Hardware Microarchitecture Compatibility**: NVIDIA Ampere, NVIDIA Blackwell, NVIDIA Hopper, NVIDIA Lovelace<br>
**Supported Operating System(s):** Linux<br>

### **Model Version(s)**

Llama 3.2 NeMo Retriever Multimodal Embedding 1B<br>
Short Name: `llama-3.2-nemoretriever-1b-vlm-embed-v1`<br>

## **Training Dataset & Evaluation**

### **Training Dataset**

The development of large-scale, public, open-QA datasets has enabled tremendous progress in powerful embedding models. However, the following issues limit the use of these models in commercial settings.

- One popular dataset, named MS MARCO, restricts ‌commercial licensing.
- Many multimodal datasets use synthetic data generation with proprietary models.

To address these issues, NVIDIA created its own training dataset. NVIDIA's training dataset is based on public QA datasets, and only includes datasets that have a license for commercial applications.

**Data Collection Method by dataset**: Hybrid: Automated, Human, Synthetic<br>
**Labeling Method by dataset**: Hybrid: Automated, Human, Synthetic<br>
**Properties:** The text component is comprised of semi-supervised pre-training on 12M samples from public datasets and fine-tuning on 1.5M samples from public datasets. The VLM component uses only commercially-viable data from the [Eagle2](https://github.com/NVlabs/EAGLE) training data.<br>

### **Evaluation Datasets**

We evaluated the NeMo Retriever Multimodal Embedding Model against both published literature and existing open-source and commercial retriever models. Our evaluation used three benchmark datasets for question-answering tasks: the public [ViDoRe V1](https://huggingface.co/spaces/vidore/vidore-leaderboard) benchmark and two internal multimodal retrieval benchmarks. For those interested in reproducing our results, one of our internal datasets (DigitalCorpora-767) can be created by following instructions in [this notebook](https://github.com/NVIDIA/nv-ingest/blob/main/evaluation/digital_corpora_download.ipynb) from the NeMo Retriever Extraction GitHub repository.

**Data Collection Method by dataset**: Hybrid: Automated, Human, Synthetic<br>
**Labeling Method by dataset**: Hybrid: Automated, Human, Synthetic<br>
**Properties:** More details on [ViDoRe V1](https://huggingface.co/spaces/vidore/vidore-leaderboard) can be found on  their leaderboard. [DigitalCorpora-767](https://github.com/NVIDIA/nv-ingest/blob/main/evaluation/digital_corpora_download.ipynb) is a set of 767 PDFs that have a good mixture of text, tables, and charts.

### **Evaluation Results**

| Model                                          | # Params Vision (in M) | # Params LLM-backbone (in M) | Average Recall@5 on DigitalCorpora-767, Earnings, ViDoRe V1 |
|------------------------------------------------|------------------------|------------------------------|--------------|
| llama-3.2-nemoretriever-1b-vlm-embed-v1        |                    429 |                         1236 |        80.9% |
| llamaindex/vdr-2b-multi-v1                     |                    665 |                         1544 |        80.9% |
| MrLight/dse-qwen2-2b-mrl-v1                    |                    665 |                         1544 |        80.4% |
| Alibaba-NLP/gme-Qwen2-VL-2B-Instruct           |                    665 |                         1544 |        79.9% |

We do not compare to col-style embedding (late interaction) models because late interaction embeddings require a significant embedding store.

### Detailed Performance Analysis

The model's performance was evaluated across different modalities and compared with other models using various pipelines. The following table contains the detailed results for the DigitalCorpora-767 dataset:

| Modality | Queries | Text-based Pipeline<br>(llama-3.2-nv-embedqa-1b-v2) | VLM-based Pipeline<br>(llama-3.2-nemoretriever-1b-vlm-embed-v1) |
|----------|---------|--------------------------------------------------|--------------------------------------------------------------|
| Multimodal | 991 | 0.845 | 0.865 |
| Table | 235 | 0.753 | 0.838 |
| Chart | 268 | 0.881 | 0.881 |
| Text | 488 | 0.869 | 0.869 |

**Inference**<br>
**Engine:** TensorRT<br>
**Test Hardware:** H100 PCIe/SXM, A100 PCIe/SXM, L40s, L4, and A10G<br>

## **Ethical Considerations**

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their supporting model team to ensure this model meets requirements for the relevant industry and use case, and address unforeseen product misuse.

For more detailed information on ethical considerations for this model, see the Model Card++ tab for the Explainability, Bias, Safety & Security, and Privacy subcards.

Please report security vulnerabilities or NVIDIA AI Concerns [here](https://www.nvidia.com/en-us/support/submit-security-vulnerability/).

## Bias

| Field | Response |
| ----- | ----- |
| Participation considerations from adversely impacted groups [protected classes](https://www.senate.ca.gov/content/protected-classes) in model design and testing | None |
| Measures taken to mitigate against unwanted bias | None |

## Explainability

| Field | Response |
| ----- | ----- |
| Intended Application & Domain: | Document and query embedding for question and answer retrieval. |
| Model Type: | Transformer encoder. |
| Intended User: | Generative AI creators working with conversational AI models. Users who want to build a question and answer application over a large corpus, leveraging the latest dense retrieval technologies. The corpus can be images of PDFs, such as text, tables, charts or infographics. |
| Output: | Array of float numbers (Dense Vector Representation for the input text). |
| Describe how the model works: | Model transforms the input into a dense vector representation. |
| Performance Metrics: | Accuracy, Throughput, and Latency. |
| Potential Known Risks: | This model does not guarantee to always retrieve the correct passage(s) for a given query. |
| Licensing & Terms of Use: | GOVERNING TERMS: The trial service is governed by the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf); and the use of this model is governed by the [NVIDIA Community Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-community-models-license/).<br>ADDITIONAL INFORMATION: [Llama 3.2 Community License](https://www.llama.com/llama3_2/license/). Built with Llama. |
| Technical Limitations: | The model's max sequence length is 8192. Longer text inputs should be truncated. |
| Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of: | N/A |
| Verified to have met prescribed NVIDIA quality standards: | Yes |

## Privacy

| Field | Response |
| ----- | ----- |
| Generatable or reverse engineerable personal data? | None |
| Personal data used to create this model? | None |
| How often is dataset reviewed? | Dataset is initially reviewed upon addition, and subsequent reviews are conducted as needed or upon request for changes. |
| Is there provenance for all datasets used in training? | Yes |
| Does data labeling (annotation, metadata) comply with privacy laws? | Yes |
| Is data compliant with data subject requests for data correction or removal, if such a request was made? | No, not possible with externally-sourced data. |
| Applicable Privacy Policy | https://www.nvidia.com/en-us/about-nvidia/privacy-policy/ |

## Safety & Security

| Field | Response |
| ----- | ----- |
| Model Application(s): | Document Embedding for Retrieval. User queries can be text and documents can be images of text, charts, tables, and infographics. |
| Use Case Restrictions: | Abide by [NVIDIA AI Foundation Models Community License Agreement](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-community-models-license/).   |
| Model and dataset restrictions: | The Principle of least privilege (PoLP) is applied limiting access for dataset generation and model development. Restrictions enforce dataset access during training, and dataset license constraints adhered to. |
| Describe the life critical impact (if present) | Not applicable |

## Prototype

```bash
image_source="https://assets.ngc.nvidia.com/products/api-catalog/nemo-retriever/embedding/court-sizing-metrics.png"
if [[ $image_source == http* ]]; then
base64_image=$(curl -s "${image_source}" | base64 -w 0)
else
base64_image=$(base64 -w 0 < "${image_source}")
fi

json_payload='{
"input": ["data:image/png;base64,'"${base64_image}"'"],
"model": "nvidia/llama-3.2-nemoretriever-1b-vlm-embed-v1",
"modality": ["image"],
"input_type": "",
"encoding_format": "float",
"truncate": "NONE"
}'

echo "${json_payload}" | \
curl -X POST https://integrate.api.nvidia.com/v1/embeddings \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $NVIDIA_API_KEY" \
-d @-
```

```python
import base64
import requests
from openai import OpenAI

image_source = "https://assets.ngc.nvidia.com/products/api-catalog/nemo-retriever/embedding/court-sizing-metrics.png"

if image_source.startswith(('http://', 'https://')):
response = requests.get(image_source)
image_b64 = base64.b64encode(response.content).decode()
else:
with open(image_source, "rb") as image_file:
image_b64 = base64.b64encode(image_file.read()).decode()

client = OpenAI(
api_key="$NVIDIA_API_KEY",
base_url="https://integrate.api.nvidia.com/v1"
)

response = client.embeddings.create(
input=[f"data:image/png,{image_b64}"],
model="nvidia/llama-3.2-nemoretriever-1b-vlm-embed-v1",
encoding_format="float",
extra_body={"modality": ["image"], "input_type": "", "truncate": "NONE"}
)

print(response.data[0].embedding)
```

```python
import base64
import requests
from langchain_nvidia_ai_endpoints import NVIDIAEmbeddings

image_source = "https://assets.ngc.nvidia.com/products/api-catalog/nemo-retriever/embedding/court-sizing-metrics.png"

if image_source.startswith(('http://', 'https://')):
response = requests.get(image_source)
image_b64 = base64.b64encode(response.content).decode()
else:
with open(image_source, "rb") as image_file:
image_b64 = base64.b64encode(image_file.read()).decode()

client = NVIDIAEmbeddings(
model="nvidia/llama-3.2-nemoretriever-1b-vlm-embed-v1",
api_key="$NVIDIA_API_KEY",
truncate="NONE",
)

image_input = f"data:image/png;base64,{image_b64}"

embedding = client.embed_documents([image_input])
print(embedding)
```