Data & Analysis ✓ NVIDIA · Official video-embeddingvision-languagevideo-retrievalsemantic-deduplicationfine-tuningdockerhuggingface

Cosmos-Embed Video Retrieval & Fine-Tuning

Use Cosmos-Embed1 to embed, retrieve, deduplicate, and fine-tune video-text data.

FollowSkills review · FSRS-2.0
Not recommended
53/ 100 5-point scale 2.7 / 5
Trust16 / 25 · 3.2/5

The skill declares only Read and Bash, uses a container with read-only data/spec mounts, passes HF_TOKEN through the environment, and explicitly forbids writing the token to specs, logs, or reports. However, Docker uses --network=host and --gpus all, model/results/hf_cache are writable, runtime pip installation is not fully pinned, and user confirmation, isolation, rollback, and sensitive-video handling are not specified, so points are deducted.

Reliability9 / 20 · 2.3/5

SKILL.md gives commands for train/evaluate/inference/export, parameter constraints, checkpoint rules, and several troubleshooting paths, making the documented happy path fairly clear. No committed tests cover the skill's key actions; operation depends on the container, GPU, external model downloads, and versions.yaml, while runtime dependency repair can alter the environment, so the score is limited by static calibration and uncertainty.

Adaptability10 / 15 · 3.3/5

Activation phrases, use cases, MSR-VTT inputs, model variants, and the boundary excluding AutoML routing are relatively clear. The skill nevertheless targets NVIDIA GPU, Docker, and a specific video-data format; Chinese-language use is not addressed, and HF/model downloads depend on external services without evidence of mainland-China reachability, so points are deducted.

Convention8 / 15 · 2.7/5

The documentation is well organized across quick start, actions, parameters, outputs, data formats, weights, S3 staging, and known pitfalls. It declares Apache-2.0 and version 0.1.0, while repository materials provide licensing and NVIDIA synchronization/maintenance context. The BENCHMARK report notes missing recommended Instructions and Examples, and there is no skill-level changelog or clearly named maintenance owner/update path, so points are deducted.

Effectiveness6 / 15 · 2.0/5

The skill documents training, evaluation, inference, and export with directly reusable Docker commands, template parameters, and output paths. The BENCHMARK report records a pass on one task and strong Codex correctness/effectiveness results, but it contains only one positive task, no negative-trigger coverage, and this review did not execute the workflows; the static cap and limited evidence therefore require deductions.

Verifiability4 / 10 · 2.0/5

The pinned revision, structured skill_info, configuration templates, known-error guidance, and BENCHMARK report provide limited auditability. However, the benchmark covers only one plan-oriented task and there is no committed test suite for the key actions or independent corroboration; reported results cannot substitute for execution in this review, so only limited credit is warranted.

Evidence confidence:Low Reviewed Jul 20, 2026 Reviewed revision 55f18499943e
The upstream repository has new commits since this review. The score still applies to the reviewed revision shown and may not cover the latest changes.
Before you use it
  • Commands use host networking and all visible GPUs, and write to model, results, and HuggingFace cache mounts; confirm the intended data, network, and filesystem scope before use.
  • HF_TOKEN, Cosmos-Embed weights, and google-bert/bert-base-uncased depend on external downloads and possible gated-model agreements; verify credentials, model terms, and mainland-China reachability in advance.
  • Each action installs protobuf<7 at runtime, without a complete dependency lock or rollback procedure; validate in an isolated environment and preserve recoverable caches and results.
  • The BENCHMARK contains only one positive task and does not establish reliability across real training, inference, export, or negative-trigger cases.
See the full review method →

What does this skill do, and when should you use it?

This skill guides agents through NVIDIA Cosmos-Embed1 for joint video-text embeddings. It covers text-to-video retrieval, video-to-video search, zero-shot or kNN classification, semantic deduplication, training, evaluation, inference, and export. It requires Docker, nvidia-container-toolkit, the published Cosmos-Embed TAO container, and a HuggingFace token when gated model weights are downloaded. It is best suited to teams with local video metadata and an NVIDIA GPU container environment.

It resolves and pulls the published Cosmos-Embed TAO container, then runs the cosmos-embed1 train, evaluate, inference, and export actions. The actions read container-mounted videos, JSON metadata, model weights, and YAML specification files; the documented MSR-VTT format uses a local video glob and metadata rows containing video and caption. Training produces checkpoints, evaluation produces metrics.json, inference produces results.json, and export can produce ONNX or HuggingFace output. The instructions also cover LoRA, single-GPU checkpoint resumption, cached embeddings, and staging S3 data locally before execution.

  1. A video platform team needs natural-language search over a local video library and uses text-to-video retrieval.
  2. A media or security team needs to find videos similar to a reference clip and uses video-to-video search.
  3. A data-governance team needs to identify duplicate or near-duplicate videos and uses semantic deduplication.
  4. A researcher needs to adapt Cosmos-Embed1 to proprietary video-text data and uses full fine-tuning or LoRA.
  5. A deployment team needs to convert a trained checkpoint to ONNX or HuggingFace format and uses the export action.

What are this skill's strengths and limitations?

Pros
  • Covers the complete documented action path from training and evaluation through inference and ONNX/HuggingFace export.
  • Provides concrete MSR-VTT data requirements, checkpoint handling, LoRA settings, smoke overrides, and troubleshooting.
  • Documents 224p, 336p, and 448p model variants.
  • The repository README states that published skills carry signatures, Tier-3 evaluation datasets, and benchmark reports.
Limitations
  • Requires Docker, NVIDIA GPU container support, and the published container; it is not documented as a standalone local Python library.
  • The container's wandb and protobuf 7.x combination causes an import failure unless protobuf is pinned below 7 before each action.
  • S3 video globs cannot be passed directly to the CLI; data must first be staged on the execution host or container filesystem.
  • The source does not provide pricing, a GPU-model compatibility matrix, or a skill-specific standalone test report.

How do you install this skill?

Install the NVIDIA collection with: npx skills add nvidia/skills. Install only this skill with: npx skills add nvidia/skills --skill tao-finetune-cosmos-embed --yes. To target Codex, use: npx skills add nvidia/skills --skill tao-finetune-cosmos-embed --agent codex --yes. The source does not document a separate Python-package installation procedure.

How do you use this skill?

After installation, ask the agent to “fine-tune Cosmos-Embed1,” “run cosmos-embed inference,” “embed videos,” or “search videos with text.” Prepare data, model, specs, and results directories, mount them into the GPU-enabled Docker container, run python -m pip install "protobuf<7", and invoke cosmos-embed1 train, evaluate, inference, or export. The documented inference example overrides inference.query.input_texts=["a man is singing on stage"] and inference.k=5. Weights may come from /model/Cosmos-Embed1-224p or from nvidia/Cosmos-Embed1-224p with HF_TOKEN supplied as an environment variable.

FAQ

Is this suitable for an environment without an NVIDIA GPU?
No clear CPU path is documented. The standard Docker options use --gpus all and require nvidia-container-toolkit.
Is a HuggingFace token always required?
It is required when downloading gated nvidia/Cosmos-Embed1-* weights. Even with the model checkpoint disabled, the Q-Former may download google-bert/bert-base-uncased, so fresh containers should receive HF_TOKEN or use a persistent cache.
Can this skill be routed through AutoML?
Not under the documented policy. Because Cosmos-Embed schemas are absent from schemas/, use this model skill's direct train, evaluate, inference, and export actions.
What should be checked first when a run fails?
Check that protobuf is below 7, the video glob points to a container-local path such as /data/video/*.mp4, metadata filenames match the videos, and HF_TOKEN and model-access terms are configured.

More skills from this repository

All from NVIDIA/skills

Data & Analysis ✓ NVIDIA · Official

Cosmos Reason Video QA Fine-Tuning

A Cosmos-RL playbook for FSDP-based SFT, LoRA, evaluation, and AutoML of Cosmos3-Nano video QA models.

Data & Analysis ✓ NVIDIA · Official

Nemotron Retrieval Recipes

Plan, debug, tune, evaluate, export, and deploy Nemotron embedding and reranking recipes.

Data & Analysis ✓ NVIDIA · Official

TAO CLIP Fine-Tuning and Deployment

Train and deploy NVIDIA TAO CLIP for image-text retrieval, zero-shot classification, and embedding extraction.

Dev & Engineering ✓ NVIDIA · Official

TAO Hugging Face Model Integrator

Connect Hugging Face computer-vision models to NVIDIA TAO training, ONNX export, and TensorRT deployment workflows.

Data & Analysis ✓ NVIDIA · Official

Nemotron Customization Pipelines

Plan, configure, and chain existing Nemotron steps for data preparation, training, alignment, conversion, optimization, and evaluation.

Automation & Ops ✓ NVIDIA · Official

RTVI-CV 2D Detection & Tracking Operations

Deploy and operate the RTVI-CV 2D video detection and tracking microservice.

Automation & Ops ✓ NVIDIA · Official

VSS Video Analysis Reports

Generates structured reports from recorded video clips or incident time ranges.

Data & Analysis ✓ NVIDIA · Official

VSS Video Analytics Query

Read video-analytics incidents, metrics, alerts, and sensor data through VA-MCP.

Dev & Engineering ✓ NVIDIA · Official

NeMo MBridge Recipe Recommender

Recommends adjustable Megatron Bridge recipes for your model, GPU budget, and training goal.

Automation & Ops ✓ NVIDIA · Official

VIOS Video I/O and Storage Manager

Manage VIOS sensors, streams, recordings, snapshots, clips, and storage through its REST API.

Data & Analysis ✓ NVIDIA · Official

TAO SegFormer Semantic Segmentation

Guidance for training, evaluating, exporting, and deploying TAO SegFormer models.

Data & Analysis ✓ NVIDIA · Official

NV Brain MRI Diffusion Fine-Tuner

Fine-tune the NV-Generate-CTMR diffusion UNet from NIfTI brain MRI training volumes.

Data & Analysis ✓ NVIDIA · Official

TAO AutoML + DEFT Three-Phase Pipeline

Chains hyperparameter tuning, iterative data improvement, and final refinement for AOI and other DEFT applications.

Automation & Ops ✓ NVIDIA · Official

VSS Video Auto-Calibration

Run end-to-end AutoMagicCalib calibration for local videos, RTSP cameras, or the bundled sample dataset.

Automation & Ops ✓ NVIDIA · Official

VSS Alert Management

Operate NVIDIA VSS real-time alerts, subscriptions, incident queries, Slack notifications, and camera onboarding.

Automation & Ops ✓ NVIDIA · Official

VSS Archive Video Search

Search archived video with natural language and ingest files or RTSP streams into a searchable index.

Data & Analysis ✓ NVIDIA · Official

Clinical ASR Fine-Tuning

Fine-tune Parakeet TDT v2 with NeMo for clinical vocabulary and close the loop with offline KER evaluation.

Data & Analysis ✓ NVIDIA · Official

TAO Depth Anything v2

Train, evaluate, export, and deploy monocular depth models from single RGB images with NVIDIA TAO.

Data & Analysis ✓ NVIDIA · Official

TAO NVPanoptix3D

Train and run 3D panoptic reconstruction with occupancy completion from posed RGB images.

Data & Analysis ✓ NVIDIA · Official

Hugging Face GPU Fine-Tuning

Build reproducible Hugging Face fine-tuning pipelines on local NVIDIA GPUs.

Related skills