strands-for-cosmos 0.7.0

NVIDIA Cosmos provider for Strands Agents - physical AI reasoning, video understanding, and embodied intelligence

strands-cosmos

NVIDIA Cosmos for Strands Agents. Give your agent eyes that understand physics and hands that generate video, audio, and robot actions - on local compute.

4 model providers (Cosmos 3 omnimodal Reasoner & Generator + Cosmos-Reason2 VLM) and 45 tools spanning the full pipeline: reasoning, generation, curation, post-training, quantization, edge deployment, and evaluation.

---

⏱️ Learn it in 90 seconds

1. Install (we use uv everywhere):

uv pip install strands-cosmos

2. Understand video - the reasoner reads vision and reasons in text. It talks to a local vLLM server - see the Run the reasoner server dropdown just below to start one, then:

from strands import Agent
from strands_cosmos import Cosmos3ReasonerModel

agent = Agent(model=Cosmos3ReasonerModel(base_url="http://localhost:8000/v1"))
agent("Caption in detail: <video>scene.mp4</video>")
agent("List the notable events with timestamps: <video>scene.mp4</video>")

3. Generate video - the generator runs in-process (Diffusers, no server):

from strands_cosmos import Cosmos3GeneratorModel

gen = Cosmos3GeneratorModel(model_id="nvidia/Cosmos3-Nano")
gen.generate(mode="text2video",            prompt="A robot navigates a warehouse.", out_path="vid.mp4")
gen.generate(mode="text2video-with-sound", prompt="A robot pours water.",           out_path="av.mp4", enable_sound=True)
gen.generate(mode="image2video",           prompt="It moves forward.", image="frame.jpg", out_path="i2v.mp4")

That's the whole loop: understand → generate. Everything below is depth - expand only what you need.

You want to…	Provider	Where
Understand video/image (text+vision → text)	Cosmos3ReasonerModel	vLLM server
Generate image/video/audio/action	Cosmos3GeneratorModel	in-process Diffusers
Run a tiny VLM on Jetson edge	CosmosVisionModel, CosmosModel	Edge VLM
Drive the full pipeline from tools	45 cosmos* tools	Tools

---

🚀 Run the reasoner server (one-time setup + serve)

The reasoner needs a vLLM server. Build a CUDA-matched env and serve Cosmos3-Nano on :8000:

just c3-doctor          # check GPU / CUDA / uv / venvs / disk + recommended CUDA pairing
just c3-setup-reason    # build the reasoner venv: vllm + vllm-cosmos3 (uv-managed)
just c3-serve-reason    # serve Cosmos3-Nano on :8000 (--max-model-len 32768)

Verify it's up before pointing an agent at it:

curl -s http://localhost:8000/v1/models    # → {"data":[{"id":"nvidia/Cosmos3-Nano",...}]}

Or one-shot a caption straight from the justfile (no Python):

just c3-reason "Caption in detail." "" scene.mp4 caption

CUDA pairing: match torch to your driver - CUDA 13 → cu130 + vllm==0.21.0; CUDA 12.8 → cu128 + vllm==0.19.1. just c3-doctor reports your driver's recommendation.

Single-GPU note: reasoner (vLLM) and generator (Diffusers) each load a 16B model and won't co-fit on one ~46GB GPU - stop one before running the other, or use separate GPUs.

📦 Install matrix (pick the extra for your task)

uv pip install strands-cosmos                  # core: Reason2 VLM + all tools
uv pip install "strands-cosmos[cosmos3]"       # + Cosmos 3 reasoner client (vLLM server)
uv pip install "strands-cosmos[cosmos3-gen]"   # + Cosmos 3 generator (in-proc Diffusers: image/video/sound)
uv pip install "strands-cosmos[vllm]"          # + bundled vLLM + openai client
uv pip install "strands-cosmos[all]"           # everything (heavy)

The generator needs the diffusers dev build (Cosmos3OmniPipeline); PyPI forbids direct-URL deps, so pin it at install time (or just use just c3-setup-gen):

uv pip install -U "git+https://github.com/huggingface/diffusers.git"

Extra	Pulls in	For
(none)	transformers, torch, torchvision, torchcodec, av	Reason2 VLM + tools
cosmos3	openai	Cosmos 3 reasoner client
cosmos3-gen	diffusers, cosmos_guardrail, soundfile, imageio	Cosmos 3 generator
vllm	vllm, openai	self-hosting vLLM
jetson	torchcodec	Jetson companions (torch via JetPack)
all	all of the above + dev tools	kitchen sink

🛠️ Developer setup (clone + build everything)

git clone https://github.com/cagataycali/strands-cosmos && cd strands-cosmos
just setup-full    # system deps, Python deps, clones all Cosmos repos (uv-managed venvs)
just doctor        # verify everything

# dedicated, CUDA-matched envs (each is its own uv venv):
just c3-setup-reason      # reasoner: vllm + vllm-cosmos3
just c3-setup-gen         # generator: diffusers(main) + cosmos_guardrail
just c3-setup-framework   # action + training: Cosmos Framework

Run/lint/test against the dev env via uv:

uv pip install -e ".[dev]"
uv run pytest
uv run ruff check .

📜 Single-file script with inline deps (PEP 723 + uv run)

Drop dependencies directly into a script's header - uv run builds an ephemeral env, no manual install. Save as agent.py:

# /// script
# requires-python = ">=3.10"
# dependencies = [
#     "strands-agents[openai]",
#     "strands-cosmos",
# ]
# ///
import os, sys
from strands import Agent
from strands_cosmos import Cosmos3ReasonerModel, cosmos3_caption, cosmos3_temporal, video_probe

model = Cosmos3ReasonerModel(base_url=os.environ.get("COSMOS_BASE_URL", "http://localhost:8000/v1"))
agent = Agent(model=model, tools=[cosmos3_caption, cosmos3_temporal, video_probe])
agent(" ".join(sys.argv[1:]) or "Caption in detail: <video>scene.mp4</video>")

uv run agent.py "List the events with timestamps: <video>scene.mp4</video>"

---

🌌 Cosmos 3 - Omnimodal World Models

Cosmos 3 is NVIDIA's newest model family: a unified Mixture-of-Transformers that jointly understands and generates text, images, video, audio, and action. strands-cosmos exposes both runtime surfaces - the Reasoner (vLLM, text+vision → text) and the Generator (Diffusers, → image/video/audio/action).

See it end-to-end: Reason → Generate

Cosmos 3 watches a real construction-site clip, describes it, then generates new videos (one with synchronized audio) from its own description - all on a single local GPU.

① Input video

② Cosmos 3 understands it

"Two construction workers wearing yellow safety vests and helmets are walking away from the camera on a dirt path within a bustling construction site. The ground is covered in loose soil, with visible tire tracks crisscrossing the surface. In the background, a large yellow front-end loader moves slowly across the site, its bucket raised slightly as it navigates the terrain. Behind the loader, partially obscured by rebar and concrete slabs, an excavator operates near a foundation area. The scene is framed by urban buildings in the distance, including a distinctive church-like structure with a tall spire and modern glass-fronted buildings. The overall atmosphere suggests active progress on a significant infrastructure project under clear daylight conditions." Cosmos3ReasonerModel (caption in 5.2s)

The reasoner distills its own understanding into a generation prompt:

"Two construction workers in yellow safety vests and helmets walk across a dusty site, gesturing toward a yellow front loader and distant excavator as they converse."

Then Cosmos3GeneratorModel generates similar videos from that prompt (832×480, 49f):

text → video	text → video + 🔊 sound	image → video
55.5s	43.2s · AAC stereo 48kHz	42.1s · from a real frame

→ Full demo + MP4s + reasoning: demo/cosmos3_showcase/ · reproduce with uv run examples/09_cosmos3_showcase.py

Capabilities & tool map

Surface	Tools	Backend
Reasoner	cosmos3_reason, cosmos3_caption, cosmos3_temporal, cosmos3_embodied, cosmos3_ground, cosmos3_plausibility, cosmos3_situation, cosmos3_action_cot	vLLM
Generator	cosmos3_text2image, cosmos3_text2video, cosmos3_image2video, cosmos3_text2video_sound	Diffusers Cosmos3OmniPipeline (in-proc)
Video-to-video	cosmos3_video2video (transfer: day→night, recolor, restyle)	vLLM-Omni Docker (vllm/vllm-omni:cosmos3)
Action / World-Model	cosmos3_forward_dynamics, cosmos3_inverse_dynamics, cosmos3_policy	Cosmos Framework (torchrun)
Training (SFT)	cosmos3_train, cosmos3_train_convert, cosmos3_train_show, cosmos3_train_export, …	Cosmos Framework (torchrun)
Servers	cosmos3_serve	start / stop / status

just c3-setup-gen       # generator env: diffusers(main) + cosmos_guardrail
just c3-gen text2video "A robot in a warehouse." "" out.mp4

just c3-setup-framework # action + training env: Cosmos Framework
just c3-action spec.jsonl /tmp/out      # forward/inverse dynamics, policy
just c3-train-recipes                   # list SFT recipes
just c3-train vision_sft_nano           # fine-tune (8x H100); see the training guide

📖 Full guides: Cosmos 3 · Training/SFT

Models

Model	Size	Capability
Cosmos3-Nano	16B	Omnimodal (reasoner + generator + action) - fits a single ~46GB GPU
Cosmos3-Super	64B	Frontier-scale (multi-GPU / tensor-parallel)
Cosmos3-Nano-Policy-DROID	16B	VL robot policy (DROID)

---

🤖 Cosmos-Reason2 - Lightweight Edge VLM

For edge/Jetson deployments, the Cosmos-Reason2 VLM runs as a Strands model provider with a tiny footprint - verified on Jetson AGX Thor with Chain-of-Thought reasoning.

from strands import Agent
from strands_cosmos import CosmosVisionModel

agent = Agent(model=CosmosVisionModel(model_id="nvidia/Cosmos-Reason2-2B"))
agent("Caption in detail: <video>dashcam.mp4</video>")                  # video understanding
agent("<image>robot_view.jpg</image> What should the robot do next?")   # image reasoning
agent("What happens when a ball rolls off a table?")                    # text-only physics

Demo, models & performance

Dashcam safety analysis with Chain-of-Thought reasoning on Jetson AGX Thor

Model	GPU Memory	Use Case
Cosmos-Reason2-2B	24GB	Edge deployment (Jetson Thor/Orin)
Cosmos-Reason2-8B	32GB	Cloud/desktop high-accuracy

Performance (Jetson AGX Thor, Reason2-2B): text inference 1.4s (46 tokens) · video caption 2.2s (short clip @ 4fps), 7s load.

Jetson install:

uv pip install strands-cosmos
strands-cosmos-fix-cublas   # fix CUBLAS for Jetson GPU architecture

---

🧰 Tools

Use any of the 45 tools inside a Strands Agent for full-pipeline automation:

from strands import Agent
from strands_cosmos import cosmos_reason_hf, video_probe, cosmos_sysinfo

agent = Agent(tools=[cosmos_reason_hf, video_probe, cosmos_sysinfo])
agent("Check the system, then analyze the video at /tmp/scene.mp4")

Full tool catalog (Reason2 / Predict / Transfer / lifecycle / data / eval)

Category	Tools	Description
Reason2 VLM	cosmos_inference, cosmos_reason_hf, cosmos_serve	TRT server inference, HF direct inference, server lifecycle
Predict 2.5	cosmos_predict_generate	World-model video generation (future frame prediction)
Transfer 2.5	cosmos_transfer_generate	ControlNet video-to-video (depth/edge/sketch→video)
Model Lifecycle	cosmos_model_download, cosmos_quantize, cosmos_export_onnx, cosmos_build_engine	Download, FP8 quantize, ONNX export, TRT engine build
Training	cosmos_post_train, cosmos_distill	SFT/LoRA post-training, knowledge distillation
Data	cosmos_curate	Xenna data curation pipeline
Evaluation	cosmos_evaluate	FID/FVD/CSE/CLIP benchmark evaluation
I/O	rtp_capture_frame, nats_publish, video_probe, video_extract_frames, image_read	RTP capture, NATS messaging, video/image utilities
System	cosmos_sysinfo	GPU/platform diagnostics

---

⚙️ Configuration

# Cosmos 3 Reasoner
Cosmos3ReasonerModel(base_url="http://localhost:8000/v1", reasoning=True, max_tokens=4096)

# Cosmos 3 Generator
Cosmos3GeneratorModel(model_id="nvidia/Cosmos3-Nano", guardrails=True)

# Cosmos-Reason2 VLM
CosmosVisionModel(
    model_id="nvidia/Cosmos-Reason2-8B",
    reasoning=True,           # Chain-of-thought <think>...</think>
    fps=4,
    params={"max_tokens": 4096, "temperature": 0.6},
)

🏗️ Architecture

strands_cosmos/
├── cosmos3_reasoner_model.py    # Cosmos3ReasonerModel (vLLM, text+vision -> text)
├── cosmos3_generator_model.py   # Cosmos3GeneratorModel (Diffusers, -> image/video/sound)
├── cosmos_vision_model.py       # CosmosVisionModel (Reason2 VLM: video+image+text)
├── cosmos_model.py              # CosmosModel (Reason2 text-only)
├── fix_cublas.py                # Jetson CUBLAS compatibility fix
├── tools/
│   ├── cosmos3.py               # 16 Cosmos 3 tools (reason/generate/action/serve)
│   ├── inference.py · reason_hf.py · serve.py        # Reason2 VLM
│   ├── predict_generate.py · transfer_generate.py    # Predict2.5 / Transfer2.5
│   ├── model_download.py · quantize.py · export_onnx.py · build_engine.py
│   ├── post_train.py · distill.py · curate.py · evaluate.py
│   └── rtp.py · nats_pub.py · video_utils.py · image_read.py · sysinfo.py
└── justfile                     # Developer workflow + c3-* recipes

✅ Verified platforms

Platform	GPU	Status
Desktop / Cloud	NVIDIA L40S / A100 / H100 / RTX 4090	✅ Cosmos 3 + Reason2
Jetson AGX Thor	NVIDIA Thor 132GB	✅ Reason2 (with CUBLAS fix)
Jetson Orin	32/64GB	✅ Reason2 (may need CUBLAS fix)

🩺 Troubleshooting

CUBLAS_STATUS_INVALID_VALUE on Jetson

strands-cosmos-fix-cublas    # replaces torch's bundled CUBLAS with JetPack system CUBLAS

Cosmos 3 reasoner OOM on a single GPU - the default sequence length (262K) needs a huge KV cache. just c3-serve-reason caps it at --max-model-len 32768. Stop the generator before serving the reasoner (and vice versa).

StopIteration in get_rope_index during video (Reason2) - already handled; strands-cosmos pins a compatible transformers range. If you still see it:

uv pip install "transformers>=4.57.0,<5.3.0"

module 'torchvision.io' has no attribute 'read_video' - transformers 5.x decodes video with torchcodec and falls back to torchvision, which removed io.read_video in >=0.27:

uv pip install torchcodec

TRT tools return exit 127 - expected on workstations; those run on Jetson or in TRT Docker. Run just doctor.

---

Resources

• Changelog - Release history
• Cosmos 3 - Latest omnimodal world models
• Cosmos Cookbook - Official recipes
• Cosmos-Reason2 - VLM source
• Strands Agents - Agent framework

---

License

Licensed under the Apache License 2.0. See NOTICE for attribution and SECURITY.md for vulnerability reporting.

Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved. | Built with NVIDIA Cosmos and Strands Agents