diff --git a/run-on-jobs/index.html b/run-on-jobs/index.html index 2c955ea..9783235 100644 --- a/run-on-jobs/index.html +++ b/run-on-jobs/index.html @@ -78,12 +78,15 @@ @media (prefers-reduced-motion:reduce){summary::before{transition:none}} +.entry{font-size:15px;margin-top:12px}

Use Hub models on your own data

Run a model over your whole dataset, or fine-tune it on your own examples. Each command starts a job on Hugging Face Jobs and saves the results to the Hub. There is nothing to clone.

+

Each recipe is a uv script: one Python file that lists its own dependencies, so it runs straight from its URL.

+

Find your model · New to Jobs? Try an example below, or read the quickstart.

Read seven scanned NASA pages with GLM-OCR and save the text to a new public dataset on your account. Add --private to keep it private.

hf jobs uv run \
@@ -212,182 +215,186 @@ 
Use Hub models on your own data

 
 Food System Constraints
 
-The primary objective in designing a food system for the Space Shuttle is to provide food that is safe and nutritious, light in weight and compact, and is packaged in a convenient form that allows easy manipulation in the weightless environment of an orbiting spacecraft. To achieve this objective requires a careful consideration of three important factors: biological, operational, and engineering. Under each factor are many constraints that affect the final choice of food and how it is

Our run: davanstrien/ocr-demo-results. 7 pages, 2.4 minutes of processing on a10g-small, after a few minutes of setup.

+The primary objective in designing a food system for the Space Shuttle is to provide food that is safe and nutritious, light in weight and compact, and is packaged in a convenient form that allows easy manipulation in the weightless environment of an orbiting spacecraft. To achieve this objective requires a careful consideration of three important factors: biological, operational, and engineering. Under each factor are many constraints that affect the final choice of food and how it is

Our run: davanstrien/ocr-demo-results. 7 pages, 2.4 minutes of processing on a10g-small, after a few minutes of setup.

More recipes and options in uv-scripts/ocr.

Transcribe four minutes of a NASA podcast, with speaker tags and word timestamps, into your own Bucket.

hf buckets create your-username/transcripts --private
 hf jobs uv run --timeout 45m \
     -v hf://buckets/uv-scripts/transcription-demo/podcast:/input:ro \
     -v hf://buckets/your-username/transcripts/podcast:/output \
     https://huggingface.co/datasets/uv-scripts/transcription/raw/main/nemotron-transcribe.py \
-    /input /output --timestamps
<spk:0> Jim, thank you so much for coming on the podcast today. ... <spk:3> It is, uh, the mission control room, of course, is now open for business ...

Our run: about 7.5 minutes on rtx-pro-6000, about $0.34. Most of that was model setup.

+ /input /output --timestamps
<spk:0> Jim, thank you so much for coming on the podcast today. ... <spk:3> It is, uh, the mission control room, of course, is now open for business ...

Our run: about 7.5 minutes on rtx-pro-6000, about $0.34. Most of that was model setup.

More recipes and options in uv-scripts/transcription.

Embed the first 100 IMDb reviews and save them as a private dataset.

hf jobs uv run --flavor l4x1 -s HF_TOKEN \
     https://huggingface.co/datasets/uv-scripts/embeddings/raw/main/generate-embeddings.py \
     stanfordnlp/imdb your-username/imdb-embeddings \
-    --column text --max-samples 100 --private

One L4 GPU embeds about 900 rows per second with all-MiniLM-L6-v2.

+ --column text --max-samples 100 --private

One L4 GPU embeds about 900 rows per second with all-MiniLM-L6-v2.

More recipes and options in uv-scripts/embeddings.

Train a news topic classifier from 16 examples, with no GPU, and save the model to your account.

hf jobs uv run --flavor cpu-basic --secrets HF_TOKEN \
     https://huggingface.co/datasets/uv-scripts/classification/raw/main/train-setfit.py \
     fancyzhx/ag_news your-username/ag-news-setfit \
     --num-samples 4 --max-train-pool 2000 --max-eval-samples 100 \
-    --max-seq-length 128 --sampling-strategy undersampling --max-minutes 5

Our run: 0.77 accuracy on 100 test articles, against 0.32 for always guessing the most common topic. Training took 20 seconds on cpu-basic.

+ --max-seq-length 128 --sampling-strategy undersampling --max-minutes 5

Our run: 0.77 accuracy on 100 test articles, against 0.32 for always guessing the most common topic. Training took 20 seconds on cpu-basic.

More recipes and options in uv-scripts/classification.

You need the hf CLI (1.32 or newer), a login and pay-as-you-go credit. Add --dry-run to see the GPU and settings without starting a job.

-

Find your model

+

Find your model

Replace the capitalised placeholders and your-username. Check the script if you're unsure what a value should be.

-

OCR

Transcription

Transcription

Segmentation

Segmentation

Object detection

Object detection

Video

Video

Classification

Classification

Embeddings

Embeddings

Text generation

Text generation

Fine-tuning

Fine-tuning

+ --num-samples 8

From uv-scripts/classification, with more options and examples.

+

Change a recipe

+

Download a script, edit it and run your copy: hf jobs uv run ./my-script.py. Flags before the script, such as --flavor, change the hardware.

+

Learn more

+

For agents

Read AGENTS.md and recipes.json.

diff --git a/run-on-jobs/recipes.json b/run-on-jobs/recipes.json index deaa39d..b171da0 100644 --- a/run-on-jobs/recipes.json +++ b/run-on-jobs/recipes.json @@ -4,6 +4,8 @@ "kind": "run", "task": "Classification", "script": "classification/classify-dataset-sglang.py", + "repo": "uv-scripts/classification", + "page": "https://huggingface.co/datasets/uv-scripts/classification/blob/main/classify-dataset-sglang.py", "url": "https://huggingface.co/datasets/uv-scripts/classification/raw/main/classify-dataset-sglang.py", "models": [ "HuggingFaceTB/SmolLM3-3B" @@ -20,6 +22,8 @@ "kind": "run", "task": "Classification", "script": "classification/classify-dataset.py", + "repo": "uv-scripts/classification", + "page": "https://huggingface.co/datasets/uv-scripts/classification/blob/main/classify-dataset.py", "url": "https://huggingface.co/datasets/uv-scripts/classification/raw/main/classify-dataset.py", "models": [ "HuggingFaceTB/SmolLM3-3B" @@ -36,6 +40,8 @@ "kind": "run", "task": "Classification", "script": "classification/classify-gliner2.py", + "repo": "uv-scripts/classification", + "page": "https://huggingface.co/datasets/uv-scripts/classification/blob/main/classify-gliner2.py", "url": "https://huggingface.co/datasets/uv-scripts/classification/raw/main/classify-gliner2.py", "models": [ "fastino/gliner2.5-multi-v1" @@ -52,6 +58,8 @@ "kind": "train", "task": "Fine-tuning", "script": "classification/train-classifier.py", + "repo": "uv-scripts/classification", + "page": "https://huggingface.co/datasets/uv-scripts/classification/blob/main/train-classifier.py", "url": "https://huggingface.co/datasets/uv-scripts/classification/raw/main/train-classifier.py", "models": [ "LiquidAI/LFM2.5-Encoder-350M" @@ -68,6 +76,8 @@ "kind": "train", "task": "Fine-tuning", "script": "classification/train-gliner2.py", + "repo": "uv-scripts/classification", + "page": "https://huggingface.co/datasets/uv-scripts/classification/blob/main/train-gliner2.py", "url": "https://huggingface.co/datasets/uv-scripts/classification/raw/main/train-gliner2.py", "models": [ "fastino/gliner2.5-multi-v1" @@ -84,6 +94,8 @@ "kind": "train", "task": "Fine-tuning", "script": "classification/train-setfit.py", + "repo": "uv-scripts/classification", + "page": "https://huggingface.co/datasets/uv-scripts/classification/blob/main/train-setfit.py", "url": "https://huggingface.co/datasets/uv-scripts/classification/raw/main/train-setfit.py", "models": [ "sentence-transformers/all-MiniLM-L6-v2" @@ -100,6 +112,8 @@ "kind": "run", "task": "Embeddings", "script": "embeddings/embed-to-lance.py", + "repo": "uv-scripts/embeddings", + "page": "https://huggingface.co/datasets/uv-scripts/embeddings/blob/main/embed-to-lance.py", "url": "https://huggingface.co/datasets/uv-scripts/embeddings/raw/main/embed-to-lance.py", "models": [ "BAAI/bge-base-en-v1.5" @@ -116,6 +130,8 @@ "kind": "run", "task": "Embeddings", "script": "embeddings/generate-embeddings-vllm.py", + "repo": "uv-scripts/embeddings", + "page": "https://huggingface.co/datasets/uv-scripts/embeddings/blob/main/generate-embeddings-vllm.py", "url": "https://huggingface.co/datasets/uv-scripts/embeddings/raw/main/generate-embeddings-vllm.py", "models": [ "Qwen/Qwen3-Embedding-0.6B" @@ -132,6 +148,8 @@ "kind": "run", "task": "Embeddings", "script": "embeddings/generate-embeddings.py", + "repo": "uv-scripts/embeddings", + "page": "https://huggingface.co/datasets/uv-scripts/embeddings/blob/main/generate-embeddings.py", "url": "https://huggingface.co/datasets/uv-scripts/embeddings/raw/main/generate-embeddings.py", "models": [ "sentence-transformers/all-MiniLM-L6-v2" @@ -148,6 +166,8 @@ "kind": "run", "task": "Embeddings", "script": "embeddings/launch-embedding-fleet.py", + "repo": "uv-scripts/embeddings", + "page": "https://huggingface.co/datasets/uv-scripts/embeddings/blob/main/launch-embedding-fleet.py", "url": "https://huggingface.co/datasets/uv-scripts/embeddings/raw/main/launch-embedding-fleet.py", "models": [ "sentence-transformers/all-MiniLM-L6-v2" @@ -164,6 +184,8 @@ "kind": "run", "task": "OCR", "script": "ocr/deepseek-ocr-vllm.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/deepseek-ocr-vllm.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/deepseek-ocr-vllm.py", "models": [ "deepseek-ai/DeepSeek-OCR" @@ -180,6 +202,8 @@ "kind": "run", "task": "OCR", "script": "ocr/deepseek-ocr2-vllm.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/deepseek-ocr2-vllm.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/deepseek-ocr2-vllm.py", "models": [ "deepseek-ai/DeepSeek-OCR-2" @@ -196,6 +220,8 @@ "kind": "run", "task": "OCR", "script": "ocr/dots-mocr.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/dots-mocr.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/dots-mocr.py", "models": [ "rednote-hilab/dots.mocr" @@ -212,6 +238,8 @@ "kind": "run", "task": "OCR", "script": "ocr/dots-ocr.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/dots-ocr.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/dots-ocr.py", "models": [ "rednote-hilab/dots.ocr" @@ -228,6 +256,8 @@ "kind": "run", "task": "OCR", "script": "ocr/glm-ocr.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/glm-ocr.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/glm-ocr.py", "models": [ "zai-org/GLM-OCR" @@ -244,6 +274,8 @@ "kind": "run", "task": "OCR", "script": "ocr/hunyuan-ocr-1.5.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/hunyuan-ocr-1.5.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/hunyuan-ocr-1.5.py", "models": [ "tencent/HunyuanOCR" @@ -260,6 +292,8 @@ "kind": "run", "task": "OCR", "script": "ocr/lift-extract.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/lift-extract.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/lift-extract.py", "models": [ "datalab-to/lift" @@ -276,6 +310,8 @@ "kind": "run", "task": "OCR", "script": "ocr/lighton-ocr2-saturate.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/lighton-ocr2-saturate.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/lighton-ocr2-saturate.py", "models": [ "lightonai/LightOnOCR-2-1B" @@ -283,7 +319,7 @@ "support": "core", "source": "header", "flavor": "a10g-small", - "timeout": null, + "timeout": "4h", "command": "hf jobs uv run \\\n https://huggingface.co/datasets/uv-scripts/ocr/raw/main/lighton-ocr2-saturate.py \\\n ", "summary": "Convert document images to markdown using LightOnOCR-2 via saturate.", "declared": true @@ -292,6 +328,8 @@ "kind": "run", "task": "OCR", "script": "ocr/lighton-ocr2.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/lighton-ocr2.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/lighton-ocr2.py", "models": [ "lightonai/LightOnOCR-2-1B" @@ -299,7 +337,7 @@ "support": "core", "source": "header", "flavor": "a10g-small", - "timeout": null, + "timeout": "1h", "command": "hf jobs uv run \\\n https://huggingface.co/datasets/uv-scripts/ocr/raw/main/lighton-ocr2.py \\\n ", "summary": "Convert document images to markdown using LightOnOCR-2 with vLLM.", "declared": true @@ -308,6 +346,8 @@ "kind": "run", "task": "OCR", "script": "ocr/lighton-ocr3-bucket.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/lighton-ocr3-bucket.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/lighton-ocr3-bucket.py", "models": [ "lightonai/LightOnOCR-3-4B" @@ -324,6 +364,8 @@ "kind": "run", "task": "OCR", "script": "ocr/lighton-ocr3.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/lighton-ocr3.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/lighton-ocr3.py", "models": [ "lightonai/LightOnOCR-3-4B" @@ -340,6 +382,8 @@ "kind": "run", "task": "OCR", "script": "ocr/minicpm-ocr.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/minicpm-ocr.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/minicpm-ocr.py", "models": [ "openbmb/MiniCPM-OCR-1.0" @@ -356,6 +400,8 @@ "kind": "run", "task": "OCR", "script": "ocr/nuextract3.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/nuextract3.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nuextract3.py", "models": [ "numind/NuExtract3" @@ -372,6 +418,8 @@ "kind": "run", "task": "OCR", "script": "ocr/olmocr2-vllm.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/olmocr2-vllm.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/olmocr2-vllm.py", "models": [ "allenai/olmOCR-2-7B-1025-FP8" @@ -379,7 +427,7 @@ "support": "core", "source": "header", "flavor": "a10g-small", - "timeout": null, + "timeout": "2h", "command": "hf jobs uv run \\\n https://huggingface.co/datasets/uv-scripts/ocr/raw/main/olmocr2-vllm.py \\\n INPUT_DATASET OUTPUT_DATASET", "summary": "Convert document images to markdown using olmOCR-2 with vLLM.", "declared": true @@ -388,6 +436,8 @@ "kind": "run", "task": "OCR", "script": "ocr/ovis-ocr2-saturate.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/ovis-ocr2-saturate.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/ovis-ocr2-saturate.py", "models": [ "ATH-MaaS/OvisOCR2" @@ -395,7 +445,7 @@ "support": "core", "source": "header", "flavor": "a10g-small", - "timeout": null, + "timeout": "4h", "command": "hf jobs uv run \\\n https://huggingface.co/datasets/uv-scripts/ocr/raw/main/ovis-ocr2-saturate.py \\\n ", "summary": "Convert document images to markdown using OvisOCR2 via saturate.", "declared": true @@ -404,6 +454,8 @@ "kind": "run", "task": "OCR", "script": "ocr/ovis-ocr2.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/ovis-ocr2.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/ovis-ocr2.py", "models": [ "ATH-MaaS/OvisOCR2" @@ -420,6 +472,8 @@ "kind": "run", "task": "OCR", "script": "ocr/paddleocr-vl-1.6.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/paddleocr-vl-1.6.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/paddleocr-vl-1.6.py", "models": [ "PaddlePaddle/PaddleOCR-VL-1.6" @@ -436,6 +490,8 @@ "kind": "run", "task": "OCR", "script": "ocr/qianfan-ocr.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/qianfan-ocr.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/qianfan-ocr.py", "models": [ "baidu/Qianfan-OCR" @@ -452,6 +508,8 @@ "kind": "run", "task": "OCR", "script": "ocr/surya-ocr.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/surya-ocr.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/surya-ocr.py", "models": [ "datalab-to/surya-ocr-2" @@ -468,6 +526,8 @@ "kind": "run", "task": "OCR", "script": "ocr/unlimited-ocr-vllm.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/unlimited-ocr-vllm.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/unlimited-ocr-vllm.py", "models": [ "baidu/Unlimited-OCR" @@ -484,6 +544,8 @@ "kind": "run", "task": "OCR", "script": "ocr/unlimited-ocr2-live.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/unlimited-ocr2-live.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/unlimited-ocr2-live.py", "models": [ "baidu/Unlimited-OCR-2" @@ -500,6 +562,8 @@ "kind": "run", "task": "OCR", "script": "ocr/unlimited-ocr2-saturate.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/unlimited-ocr2-saturate.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/unlimited-ocr2-saturate.py", "models": [ "baidu/Unlimited-OCR-2" @@ -516,6 +580,8 @@ "kind": "run", "task": "OCR", "script": "ocr/unlimited-ocr2.py", + "repo": "uv-scripts/ocr", + "page": "https://huggingface.co/datasets/uv-scripts/ocr/blob/main/unlimited-ocr2.py", "url": "https://huggingface.co/datasets/uv-scripts/ocr/raw/main/unlimited-ocr2.py", "models": [ "baidu/Unlimited-OCR-2" @@ -532,6 +598,8 @@ "kind": "run", "task": "Segmentation", "script": "sam3/detect-objects.py", + "repo": "uv-scripts/sam3", + "page": "https://huggingface.co/datasets/uv-scripts/sam3/blob/main/detect-objects.py", "url": "https://huggingface.co/datasets/uv-scripts/sam3/raw/main/detect-objects.py", "models": [ "facebook/sam3" @@ -548,6 +616,8 @@ "kind": "run", "task": "Segmentation", "script": "sam3/segment-objects.py", + "repo": "uv-scripts/sam3", + "page": "https://huggingface.co/datasets/uv-scripts/sam3/blob/main/segment-objects.py", "url": "https://huggingface.co/datasets/uv-scripts/sam3/raw/main/segment-objects.py", "models": [ "facebook/sam3" @@ -564,6 +634,8 @@ "kind": "run", "task": "Transcription", "script": "transcription/cohere-transcribe-vllm.py", + "repo": "uv-scripts/transcription", + "page": "https://huggingface.co/datasets/uv-scripts/transcription/blob/main/cohere-transcribe-vllm.py", "url": "https://huggingface.co/datasets/uv-scripts/transcription/raw/main/cohere-transcribe-vllm.py", "models": [ "CohereLabs/cohere-transcribe-03-2026" @@ -580,6 +652,8 @@ "kind": "run", "task": "Transcription", "script": "transcription/cohere-transcribe.py", + "repo": "uv-scripts/transcription", + "page": "https://huggingface.co/datasets/uv-scripts/transcription/blob/main/cohere-transcribe.py", "url": "https://huggingface.co/datasets/uv-scripts/transcription/raw/main/cohere-transcribe.py", "models": [ "CohereLabs/cohere-transcribe-03-2026" @@ -596,6 +670,8 @@ "kind": "run", "task": "Transcription", "script": "transcription/easytranscriber-transcribe.py", + "repo": "uv-scripts/transcription", + "page": "https://huggingface.co/datasets/uv-scripts/transcription/blob/main/easytranscriber-transcribe.py", "url": "https://huggingface.co/datasets/uv-scripts/transcription/raw/main/easytranscriber-transcribe.py", "models": [ "CohereLabs/cohere-transcribe-03-2026", @@ -614,6 +690,8 @@ "kind": "run", "task": "Transcription", "script": "transcription/granite-turboctc-transcribe.py", + "repo": "uv-scripts/transcription", + "page": "https://huggingface.co/datasets/uv-scripts/transcription/blob/main/granite-turboctc-transcribe.py", "url": "https://huggingface.co/datasets/uv-scripts/transcription/raw/main/granite-turboctc-transcribe.py", "models": [ "ibm-granite/granite-speech-5.0-470m-turboctc" @@ -630,6 +708,8 @@ "kind": "run", "task": "Transcription", "script": "transcription/moss-transcribe-diarize-server.py", + "repo": "uv-scripts/transcription", + "page": "https://huggingface.co/datasets/uv-scripts/transcription/blob/main/moss-transcribe-diarize-server.py", "url": "https://huggingface.co/datasets/uv-scripts/transcription/raw/main/moss-transcribe-diarize-server.py", "models": [ "OpenMOSS-Team/MOSS-Transcribe-Diarize" @@ -646,6 +726,8 @@ "kind": "run", "task": "Transcription", "script": "transcription/moss-transcribe-diarize.py", + "repo": "uv-scripts/transcription", + "page": "https://huggingface.co/datasets/uv-scripts/transcription/blob/main/moss-transcribe-diarize.py", "url": "https://huggingface.co/datasets/uv-scripts/transcription/raw/main/moss-transcribe-diarize.py", "models": [ "OpenMOSS-Team/MOSS-Transcribe-Diarize" @@ -662,6 +744,8 @@ "kind": "run", "task": "Transcription", "script": "transcription/nemotron-transcribe.py", + "repo": "uv-scripts/transcription", + "page": "https://huggingface.co/datasets/uv-scripts/transcription/blob/main/nemotron-transcribe.py", "url": "https://huggingface.co/datasets/uv-scripts/transcription/raw/main/nemotron-transcribe.py", "models": [ "nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B" @@ -669,7 +753,7 @@ "support": null, "source": "header", "flavor": "rtx-pro-6000", - "timeout": null, + "timeout": "4h", "command": "hf jobs uv run \\\n -v hf://buckets/your-username/audio-files:/input:ro \\\n -v hf://buckets/your-username/transcripts:/output \\\n https://huggingface.co/datasets/uv-scripts/transcription/raw/main/nemotron-transcribe.py \\\n /input /output", "summary": "Transcribe audio with NVIDIA Nemotron 3.5 Transcribe via an in-job vLLM server.", "declared": true @@ -678,6 +762,8 @@ "kind": "run", "task": "Video", "script": "video/marlin-caption.py", + "repo": "uv-scripts/video", + "page": "https://huggingface.co/datasets/uv-scripts/video/blob/main/marlin-caption.py", "url": "https://huggingface.co/datasets/uv-scripts/video/raw/main/marlin-caption.py", "models": [ "NemoStation/Marlin-2B" @@ -694,6 +780,8 @@ "kind": "run", "task": "Video", "script": "video/qwen38-caption.py", + "repo": "uv-scripts/video", + "page": "https://huggingface.co/datasets/uv-scripts/video/blob/main/qwen38-caption.py", "url": "https://huggingface.co/datasets/uv-scripts/video/raw/main/qwen38-caption.py", "models": [ "Qwen/Qwen3.8-27B" @@ -701,7 +789,7 @@ "support": null, "source": "header", "flavor": "a100-large", - "timeout": null, + "timeout": "2h", "command": "hf jobs uv run \\\n -v hf://buckets/your-username/my-videos:/input:ro \\\n https://huggingface.co/datasets/uv-scripts/video/raw/main/qwen38-caption.py \\\n /input hf://buckets/your-username/my-videos/captions", "summary": "Caption whole videos — no chunking — with Qwen3.8-27B, writing a resumable dataset.", "declared": true @@ -710,6 +798,8 @@ "kind": "run", "task": "Text generation", "script": "vllm/generate-responses.py", + "repo": "uv-scripts/vllm", + "page": "https://huggingface.co/datasets/uv-scripts/vllm/blob/main/generate-responses.py", "url": "https://huggingface.co/datasets/uv-scripts/vllm/raw/main/generate-responses.py", "models": [ "Qwen/Qwen3-30B-A3B-Instruct-2507" @@ -726,6 +816,8 @@ "kind": "run", "task": "Text generation", "script": "vllm/vlm-classify.py", + "repo": "uv-scripts/vllm", + "page": "https://huggingface.co/datasets/uv-scripts/vllm/blob/main/vlm-classify.py", "url": "https://huggingface.co/datasets/uv-scripts/vllm/raw/main/vlm-classify.py", "models": [ "Qwen/Qwen2-VL-7B-Instruct" @@ -742,6 +834,8 @@ "kind": "run", "task": "Object detection", "script": "vlm-object-detection/qwen3vl-detect-fewshot.py", + "repo": "uv-scripts/vlm-object-detection", + "page": "https://huggingface.co/datasets/uv-scripts/vlm-object-detection/blob/main/qwen3vl-detect-fewshot.py", "url": "https://huggingface.co/datasets/uv-scripts/vlm-object-detection/raw/main/qwen3vl-detect-fewshot.py", "models": [ "Qwen/Qwen3.6-35B-A3B" @@ -758,6 +852,8 @@ "kind": "run", "task": "Object detection", "script": "vlm-object-detection/qwen3vl-detect.py", + "repo": "uv-scripts/vlm-object-detection", + "page": "https://huggingface.co/datasets/uv-scripts/vlm-object-detection/blob/main/qwen3vl-detect.py", "url": "https://huggingface.co/datasets/uv-scripts/vlm-object-detection/raw/main/qwen3vl-detect.py", "models": [ "Qwen/Qwen3.6-35B-A3B" diff --git a/tools/build-catalog.py b/tools/build-catalog.py index a387360..1f01231 100755 --- a/tools/build-catalog.py +++ b/tools/build-catalog.py @@ -236,6 +236,8 @@ def load_recipes(): "kind": "train" if is_training else "run", "task": TRAINING_TASK if is_training else task, "script": script, + "repo": f"uv-scripts/{folder}", + "page": f"{HUB_BASE}/{folder}/blob/main/{path.name}", "url": url, "models": models, "support": support, @@ -276,13 +278,14 @@ def recipe_html(recipe): flavor = f" · {html.escape(recipe['flavor'])}" if recipe["flavor"] else "" summary = f'

{html.escape(recipe["summary"])}

' if recipe["summary"] else "" if recipe["source"] == "flags-needed": - body = f'

Read the script for its options. There is no ready-made command yet.

' + body = f'

Read the script for its options. There is no ready-made command yet.

' else: body = command_block(recipe["command"]) name = recipe["script"].split("/")[-1] return ( - f'

{html.escape(name)} ' - f'{status}{flavor}

{summary}{body}
' + f'

{html.escape(name)} ' + f'{status}{flavor}

{summary}{body}' + f'

From {html.escape(recipe["repo"])}, with more options and examples.

' ) diff --git a/tools/run-on-jobs-template.html b/tools/run-on-jobs-template.html index 5fee821..98812b6 100644 --- a/tools/run-on-jobs-template.html +++ b/tools/run-on-jobs-template.html @@ -78,12 +78,15 @@ @media (prefers-reduced-motion:reduce){summary::before{transition:none}} +.entry{font-size:15px;margin-top:12px}

Use Hub models on your own data

Run a model over your whole dataset, or fine-tune it on your own examples. Each command starts a job on Hugging Face Jobs and saves the results to the Hub. There is nothing to clone.

+

Each recipe is a uv script: one Python file that lists its own dependencies, so it runs straight from its URL.

+

Find your model · New to Jobs? Try an example below, or read the quickstart.

Read seven scanned NASA pages with GLM-OCR and save the text to a new public dataset on your account. Add --private to keep it private.

hf jobs uv run \
@@ -212,25 +215,25 @@ 
Use Hub models on your own data

 
 Food System Constraints
 
-The primary objective in designing a food system for the Space Shuttle is to provide food that is safe and nutritious, light in weight and compact, and is packaged in a convenient form that allows easy manipulation in the weightless environment of an orbiting spacecraft. To achieve this objective requires a careful consideration of three important factors: biological, operational, and engineering. Under each factor are many constraints that affect the final choice of food and how it is

Our run: davanstrien/ocr-demo-results. 7 pages, 2.4 minutes of processing on a10g-small, after a few minutes of setup.

+The primary objective in designing a food system for the Space Shuttle is to provide food that is safe and nutritious, light in weight and compact, and is packaged in a convenient form that allows easy manipulation in the weightless environment of an orbiting spacecraft. To achieve this objective requires a careful consideration of three important factors: biological, operational, and engineering. Under each factor are many constraints that affect the final choice of food and how it is

Our run: davanstrien/ocr-demo-results. 7 pages, 2.4 minutes of processing on a10g-small, after a few minutes of setup.

More recipes and options in uv-scripts/ocr.

Transcribe four minutes of a NASA podcast, with speaker tags and word timestamps, into your own Bucket.

hf buckets create your-username/transcripts --private
 hf jobs uv run --timeout 45m \
     -v hf://buckets/uv-scripts/transcription-demo/podcast:/input:ro \
     -v hf://buckets/your-username/transcripts/podcast:/output \
     https://huggingface.co/datasets/uv-scripts/transcription/raw/main/nemotron-transcribe.py \
-    /input /output --timestamps
<spk:0> Jim, thank you so much for coming on the podcast today. ... <spk:3> It is, uh, the mission control room, of course, is now open for business ...

Our run: about 7.5 minutes on rtx-pro-6000, about $0.34. Most of that was model setup.

+ /input /output --timestamps
<spk:0> Jim, thank you so much for coming on the podcast today. ... <spk:3> It is, uh, the mission control room, of course, is now open for business ...

Our run: about 7.5 minutes on rtx-pro-6000, about $0.34. Most of that was model setup.

More recipes and options in uv-scripts/transcription.

Embed the first 100 IMDb reviews and save them as a private dataset.

hf jobs uv run --flavor l4x1 -s HF_TOKEN \
     https://huggingface.co/datasets/uv-scripts/embeddings/raw/main/generate-embeddings.py \
     stanfordnlp/imdb your-username/imdb-embeddings \
-    --column text --max-samples 100 --private

One L4 GPU embeds about 900 rows per second with all-MiniLM-L6-v2.

+ --column text --max-samples 100 --private

One L4 GPU embeds about 900 rows per second with all-MiniLM-L6-v2.

More recipes and options in uv-scripts/embeddings.

Train a news topic classifier from 16 examples, with no GPU, and save the model to your account.

hf jobs uv run --flavor cpu-basic --secrets HF_TOKEN \
     https://huggingface.co/datasets/uv-scripts/classification/raw/main/train-setfit.py \
     fancyzhx/ag_news your-username/ag-news-setfit \
     --num-samples 4 --max-train-pool 2000 --max-eval-samples 100 \
-    --max-seq-length 128 --sampling-strategy undersampling --max-minutes 5

Our run: 0.77 accuracy on 100 test articles, against 0.32 for always guessing the most common topic. Training took 20 seconds on cpu-basic.

+ --max-seq-length 128 --sampling-strategy undersampling --max-minutes 5

Our run: 0.77 accuracy on 100 test articles, against 0.32 for always guessing the most common topic. Training took 20 seconds on cpu-basic.

More recipes and options in uv-scripts/classification.

You need the hf CLI (1.32 or newer), a login and pay-as-you-go credit. Add --dry-run to see the GPU and settings without starting a job.

-

Find your model

+

Find your model

Replace the capitalised placeholders and your-username. Check the script if you're unsure what a value should be.

@@ -238,6 +241,10 @@

Find your model

{{GROUPS}}

Data tools with no model

    {{TOOLS}}
+

Change a recipe

+

Download a script, edit it and run your copy: hf jobs uv run ./my-script.py. Flags before the script, such as --flavor, change the hardware.

+

Learn more

+

For agents

Read AGENTS.md and recipes.json.