Instructions to use WootzappLab/fara-9b-baseline-eval with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use WootzappLab/fara-9b-baseline-eval with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="WootzappLab/fara-9b-baseline-eval")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("WootzappLab/fara-9b-baseline-eval")
model = AutoModelForMultimodalLM.from_pretrained("WootzappLab/fara-9b-baseline-eval", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use WootzappLab/fara-9b-baseline-eval with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "WootzappLab/fara-9b-baseline-eval"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "WootzappLab/fara-9b-baseline-eval",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/WootzappLab/fara-9b-baseline-eval

SGLang

How to use WootzappLab/fara-9b-baseline-eval with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "WootzappLab/fara-9b-baseline-eval" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "WootzappLab/fara-9b-baseline-eval",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "WootzappLab/fara-9b-baseline-eval" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "WootzappLab/fara-9b-baseline-eval",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use WootzappLab/fara-9b-baseline-eval with Docker Model Runner:
```
docker model run hf.co/WootzappLab/fara-9b-baseline-eval
```

Fara 1.5-9B W8 baseline evaluation

This repository preserves the exact Fara 1.5-9B checkpoint used for the W8 DOM browser-agent baseline, together with commit-level provenance for the task dataset and evaluation environment.

The checkpoint was not trained or fine-tuned for this run. Its files are an exact mirror of the pinned Microsoft checkpoint. The W8 integration translated the recorder's DOM-based OpenAI Responses contract to the model-serving endpoint; that infrastructure code does not modify the weights.

Baseline result

Measure	Result
Selected tasks	90
Passed	6
Verified failures	60
Execution errors	24
Infrastructure errors	0
Raw Pass@1	6.67%
Pass rate among verified tasks	9.09%

The 24 execution errors require a separate audit. The raw Pass@1 should not be treated as a clean estimate of model capability until those errors are classified and any invalid trials are rerun.

Provenance

Component	Repository	Exact revision
Base checkpoint	`microsoft/Fara1.5-9B`	`1a93677cd89d5601bc2ed759791e981f3a520032`
Fara reference source	`microsoft/fara`	`a675d6d61c41c47ae87bacefeab22caad18e3e84`
Task dataset	`WootzappLab/cua-bench`	`745eebb8f07abb92fdaae356bb6bee3a1f2975af`
Task recorder, browser harness, environments, and verifier	`tokenbender/w8-cua-bench`	`66037db6bae4b00a2b5860aa08ebb69b559e6871`
DOM serving infrastructure	Pending infrastructure PR	Pending

The Hugging Face revision pins the task dataset. The Git revision separately pins the recorder, browser harness, browser environments, and verifier used to execute and score those tasks.

Serving configuration

Model dtype: BF16
KV-cache dtype: BF16 through vLLM auto
Maximum model length: 32,768 tokens
Tensor parallel size: 1
Runtime: vLLM 0.19.1, PyTorch 2.10.0, Transformers 5.6.2
Hardware: one NVIDIA H100 80 GB HBM3
GCP instance: w8-fara-9b-8115f370-head-3ibtw2ah-compute
Run ID: fara15-dom-baseline
Run date: 2026-09-25

Repository contents

Model configuration, tokenizer, processor, chat template, and Safetensors weights at the repository root, copied from the pinned upstream revision.
CHECKPOINT_FILES.json: upstream Git blob identities and LFS SHA256 hashes, including all four Safetensors shards.
MANIFEST.json: machine-readable checkpoint, code, and run provenance.
PROVENANCE.md: concise revision and experiment-boundary receipt.
CHECKSUMS.sha256: hashes for the repository's reproducibility metadata.

Evaluation artifacts, the detailed results report, and the infrastructure PR will be linked after they are published. Their absence here does not affect the identity of the preserved checkpoint.

Load the checkpoint

import torch
from transformers import AutoModelForMultimodalLM, AutoProcessor

model_id = "WootzappLab/fara-9b-baseline-eval"
processor = AutoProcessor.from_pretrained(model_id)
model = AutoModelForMultimodalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)

License and attribution

The mirrored base checkpoint is released by Microsoft under the MIT License. See the upstream Fara 1.5-9B model card for its original documentation, intended use, and limitations.

Downloads last month: 21

Safetensors

Model size

9B params

Tensor type

BF16

Model tree for WootzappLab/fara-9b-baseline-eval

Base model

Qwen/Qwen3.5-9B-Base

Finetuned

Qwen/Qwen3.5-9B

Finetuned

microsoft/Fara1.5-9B

Finetuned

(1)

this model