Instructions to use harshatheg/GPC-1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use harshatheg/GPC-1 with Transformers:

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

pipe = pipeline("image-text-to-text", model="harshatheg/GPC-1")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("harshatheg/GPC-1")
model = AutoModelForMultimodalLM.from_pretrained("harshatheg/GPC-1", device_map="auto")

PEFT
How to use harshatheg/GPC-1 with PEFT:
```
Task type is invalid.
```
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use harshatheg/GPC-1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "harshatheg/GPC-1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "harshatheg/GPC-1",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/harshatheg/GPC-1

SGLang

How to use harshatheg/GPC-1 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 "harshatheg/GPC-1" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "harshatheg/GPC-1",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "harshatheg/GPC-1" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "harshatheg/GPC-1",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use harshatheg/GPC-1 with Docker Model Runner:
```
docker model run hf.co/harshatheg/GPC-1
```

GPC-1 / MODEL_CARD.md

harshatheg

Release v1

c87a0aa 14 days ago

preview code

Raw History Blame Contribute Delete

1.73 kB

	# GPC-1 model card

	GPC-1 is a structured prediction model based on Qwen3.5-35B-A3B. It supports categorical choices, bounded numeric estimates, image-conditioned coordinates, and complete records selected from a caller-defined set.

	## Model details

	\| Property \| Value \|
	\| --- \| --- \|
	\| API model ID \| `gpc-1` \|
	\| Backbone family \| `Qwen/Qwen3.5-35B-A3B` \|
	\| Architecture \| Multimodal mixture of experts \|
	\| Distribution \| GPC-1 backbone and matching adapter \|
	\| Runtime \| NVIDIA, BF16, Transformers / PyTorch \|
	\| Categorical support \| 2–255 caller-defined choices \|
	\| Numeric support \| 101 positions per declared range \|
	\| Joint output \| Caller-enumerated complete records \|
	\| Server input ceiling \| 256K tokens (262,144), including compiled request overhead \|
	\| License \| Apache-2.0, with applicable third-party notices \|

	The release package contains GPC-1's backbone weights and matching adapter. The runtime loads and verifies both. See [setup](README.md#get-started), [download options](docs/DOWNLOADS.md), and [context configuration](API.md#context-window).

	## Intended use

	Document and message classification, bounded numerical estimates, image-coordinate annotation, and structured workflow decisions. Define labels, units, and reference frames explicitly in each request.

	Validate application-specific accuracy before using predictions in medical, legal, financial, or safety-critical decisions. The [API guide](API.md) defines score interpretation and supported output structures.

	## License

	GPC-1 inference code and the Qwen base use Apache-2.0. Preserve applicable upstream notices. Example media retain their own licenses. See [LICENSE](LICENSE), [NOTICE](NOTICE), and [attribution](docs/DATA_AND_LICENSES.md).

	# GPC-1 model card

	GPC-1 is a structured prediction model based on Qwen3.5-35B-A3B. It supports categorical choices, bounded numeric estimates, image-conditioned coordinates, and complete records selected from a caller-defined set.

	## Model details

	\| Property \| Value \|
	\| --- \| --- \|
	\| API model ID \| `gpc-1` \|
	\| Backbone family \| `Qwen/Qwen3.5-35B-A3B` \|
	\| Architecture \| Multimodal mixture of experts \|
	\| Distribution \| GPC-1 backbone and matching adapter \|
	\| Runtime \| NVIDIA, BF16, Transformers / PyTorch \|
	\| Categorical support \| 2–255 caller-defined choices \|
	\| Numeric support \| 101 positions per declared range \|
	\| Joint output \| Caller-enumerated complete records \|
	\| Server input ceiling \| 256K tokens (262,144), including compiled request overhead \|
	\| License \| Apache-2.0, with applicable third-party notices \|

	The release package contains GPC-1's backbone weights and matching adapter. The runtime loads and verifies both. See [setup](README.md#get-started), [download options](docs/DOWNLOADS.md), and [context configuration](API.md#context-window).

	## Intended use

	Document and message classification, bounded numerical estimates, image-coordinate annotation, and structured workflow decisions. Define labels, units, and reference frames explicitly in each request.

	Validate application-specific accuracy before using predictions in medical, legal, financial, or safety-critical decisions. The [API guide](API.md) defines score interpretation and supported output structures.

	## License

	GPC-1 inference code and the Qwen base use Apache-2.0. Preserve applicable upstream notices. Example media retain their own licenses. See [LICENSE](LICENSE), [NOTICE](NOTICE), and [attribution](docs/DATA_AND_LICENSES.md).