tao-generate-image-grounding
Two-step image grounding pipeline: extracts referring expressions from (image, caption) pairs and grounds them to pixel-space bounding boxes via a VLM. Use when the user wants to ground captions to bboxes, generate phrase-grounded annotations, auto-label images for grounding, or run the image_grounding pipeline. Triggers include 'image grounding', 'phrase grounding', 'ground captions', 'auto-label image grounding', 'image_grounding'.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-02
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- name
- tao-generate-image-grounding
- tags
- [ "image", "grounding", "bounding-boxes", "auto-label", "vlm", "2d-grounding" ]
- license
- Apache-2.0
- metadata
- { "author": "NVIDIA Corporation", "version": "0.1.0" }
- description
- Two-step image grounding pipeline: extracts referring expressions from (image, caption) pairs and grounds them to pixel-space bounding boxes via a VLM. Use when the user wants to ground captions to bboxes, generate phrase-grounded annotations, auto-label images for grounding, or run the image_grounding pipeline. Triggers include 'image grounding', 'phrase grounding', 'ground captions', 'auto-label image grounding', 'image_grounding'.
- allowed-tools
- Read Bash Write
- compatibility
- Requires docker + nvidia-container-toolkit + at least one VLM endpoint (Gemini API key or OpenAI-compatible).
02 资源文件 7
- 参考文档BENCHMARK.md3.8 KB
- 参考文档evals/evals.json819 B
- 参考文档references/configuration.md9.3 KB
- 参考文档references/skill_info.yaml1015 B
- 参考文档references/vllm_server.md6.7 KB
- 参考文档skill-card.md3.8 KB
- 其他skill.oms.sig5.2 KB
03 SKILL.md 正文
Image Grounding Pipeline
Turn (image, caption) pairs into per-image grounded annotations: cleaned captions, referring expressions with character spans, and pixel-space bounding boxes for each expression. A single VLM (Gemini or any OpenAI-compatible endpoint) handles both steps.
Purpose
Generate phrase-grounded training data for referring-expression and grounding models. The VLM acts as a "teacher" annotator: Step 0 extracts referring expressions from the caption while looking at the image; Step 1 returns one bbox set per expression for each image.
Pipeline Architecture
Step 0: Expression extraction → VLM cleans caption, extracts referring expressions + char spans
Step 1: Phrase grounding → VLM returns pixel bboxes + scores per expression
Steps are individually selectable via workflow.steps. Each step writes a per-sample checkpoint to step_<N>_*/.ckpt/<sample_id>.json and skips already-processed records on re-run. Set workflow.force_reprocess: true to ignore checkpoints and reprocess from scratch.
Instructions
Initial setup
When a user wants to run this pipeline, walk through these steps:
-
Input JSONL: Ask for the JSONL path. Each line must be one object like
{"image_path": "...", "caption": "..."}.image_pathcan be absolute or relative. -
Image root: If any
image_pathvalues are relative, setdata.image_rootto the directory they should resolve from. -
API access: Ask the user which VLM endpoint they want to use. Present these five options and act on the choice:
- Gemini — set
vlm.backend: "gemini"; requireGOOGLE_API_KEY(env var orvlm.gemini.api_key). - NIM (e.g.
https://inference-api.nvidia.com/v1) — setvlm.backend: "openai"; collectbase_url,model_name, andapi_key. - TAO inference microservice (self-hosted, OpenAI-compatible). Confirm whether the server is already running:
- Running — collect
base_url,model_name, and (optionally)api_key; setvlm.backend: "openai". - Not running — guide the user through the
skills/applications/tao-run-inference-serviceskill, which stands up a local TAO inference microservice with an OpenAI-compatible API. Before promising a specific model, checkskills/applications/tao-run-inference-service/references/service.yamlforvalid_network_arch_config_basenames. Once the server is up, collectbase_url,model_name, and (optionally)api_key; setvlm.backend: "openai".
- Running — collect
- vLLM (self-hosted, OpenAI-compatible). Confirm whether the server is already running:
- Running — collect
base_url,model_name, and (optionally)api_key; setvlm.backend: "openai". - Not running — follow references/vllm_server.md to install and launch a vLLM server, then collect
base_url,model_name, and (optionally)api_key; setvlm.backend: "openai".
- Running — collect
- Custom (any other OpenAI-compatible endpoint) — set
vlm.backend: "openai"; collectbase_url,model_name, and (optionally)api_key.
If the user has no endpoint and does not want to set one up, stop and help resolve API access first.
- Gemini — set
-
Workflow steps: Choose one of:
- Full pipeline:
["0", "1"] - Expression extraction only:
["0"] - Grounding only:
["1"], which requires existing step-0 output atresults_dir/step_0_expression_extraction/annotations.jsonl
- Full pipeline:
-
Resume vs fresh run: By default, the workflow reuses checkpoints and skips completed records. To reprocess everything, set
image_grounding.workflow.force_reprocess=true.
Running the pipeline
The pipeline runs inside the TAO Toolkit container via the auto_label CLI:
auto_label generate -e /path/to/spec.yaml \
results_dir=/results \
image_grounding.data.input_jsonl=/data/captions.jsonl \
image_grounding.data.image_root=/data/images \
image_grounding.vlm.gemini.api_key=$GOOGLE_API_KEY
Generate a default spec: auto_label default_specs results_dir=/results module_name=auto_label, then set autolabel_type: "image_grounding". All fields support Hydra dot-notation overrides on the command line.
See references/configuration.md for the full YAML structure, all parameters, model/endpoint setup, and error patterns.
Recommended pilot workflow
- Run on 5-10 images with both steps
- Inspect
step_0_expression_extraction/annotations.jsonl— arecleaned_captionandexpressions[]accurate? Are the right noun phrases captured? - Inspect
step_1_grounding/annotations.jsonl— do the bboxes inexpressions[].instances[]look right? Are confidence scores reasonable? - If quality is insufficient, switch the VLM to a stronger model (e.g.
gemini-2.5-pro) or raisemedia_resolution/max_output_tokens, then re-run withforce_reprocess=true. - Scale to the full dataset once satisfied.
Configuration
Key configuration fields (full reference in references/configuration.md):
| Field | Default | Description |
|---|---|---|
workflow.steps |
["0","1"] |
Which pipeline steps to execute ("0" = expressions, "1" = grounding) |
workflow.max_workers |
4 |
Parallel threads per step (watch API rate limits) |
workflow.force_reprocess |
false |
Ignore per-sample checkpoints and reprocess from scratch |
vlm.backend |
"gemini" |
"gemini" or "openai" (OpenAI-compatible endpoint) |
data.input_jsonl |
required | Path to input JSONL with image_path + caption per line |
data.image_root |
"" |
Optional prefix for resolving relative image_path entries |
Inputs
A single JSONL file at data.input_jsonl. One JSON object per line:
| Field | Required | Description |
|---|---|---|
image_path |
yes | Absolute path, or relative path resolved against data.image_root |
caption |
yes | Free-text caption for the image |
image_id |
no | Stable identifier; auto-derived from the filename if missing |
width, height |
no | Image dimensions in pixels; default to 1920×1080 for bbox clamping if missing |
Outputs
All outputs go to results_dir/:
step_0_expression_extraction/annotations.jsonl— per-record output enriched withcleaned_captionandexpressions[](each withtext,expression_id,char_span,noun_chunk, emptyinstances[]).step_1_grounding/annotations.jsonl— same records withexpressions[].instances[]filled in (each instance hasbbox: [x1,y1,x2,y2]in pixel space,scorein[0.0, 1.0], andbbox_id).results_dir/annotations.jsonl— copy of the last step's output for convenience.step_<N>_*/.ckpt/<sample_id>.json— per-sample checkpoints used for resume.
Prerequisites
- Container:
nvcr.io/nvidia/tao/tao-toolkit:6.26.3-pyt - API access: At least one VLM endpoint (Gemini API key or OpenAI-compatible endpoint capable of image input)
04 同仓库更多 Skills
youtube-clipper
YouTube 视频智能剪辑工具。下载视频和字幕,AI 分析生成精细章节(几分钟级别), 用户选择片段后自动剪辑、翻译字幕为中英双语、烧录字幕到视频,并生成总结文案。 使用场景:当用户需要剪辑 YouTube 视频、生成短视频片段、制作双语字幕版本时。 关键词:视频剪辑、YouTube、字幕翻译、双语字幕、视频下载、clip video
viem-integration
Integrate EVM blockchains using viem. Use when user says "read blockchain data", "send transaction", "interact with smart contract", "connect to Ethereum", "use viem", "use wagmi", "wallet integration", "viem setup", or mentions blockchain/EVM development with TypeScript.
v4-security-foundations
Security-first Uniswap v4 hook development. Use when user mentions "v4 hooks", "hook security", "PoolManager", "beforeSwap", "afterSwap", or asks about V4 hook best practices, vulnerabilities, or audit requirements.
v4-sdk-integration
App-layer SDK guide for building swap and liquidity experiences directly with the Uniswap v4 SDK. Use when user asks about "v4 sdk", "uniswap v4", "v4 swap", "v4 liquidity", "PoolManager", "V4Planner", "StateView", "PositionManager", "pool state", "v4 position", "uniswap sdk", or when building swap/liquidity UX directly with SDKs rather than via the Trading API.
v4-hook-generator
Generate Uniswap v4 hook contracts via OpenZeppelin MCP. Use when building custom swap logic, async swaps, hook-owned liquidity, custom curves, dynamic fees, MEV protection, limit orders, or oracle hooks.
testing
For testing with Light Protocol programs and clients on localnet, devnet, and mainnet validation.