Instructions to use Huggggooo/ProtoCycle-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Huggggooo/ProtoCycle-7B with Transformers:

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

pipe = pipeline("text-generation", model="Huggggooo/ProtoCycle-7B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Huggggooo/ProtoCycle-7B")
model = AutoModelForCausalLM.from_pretrained("Huggggooo/ProtoCycle-7B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Huggggooo/ProtoCycle-7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Huggggooo/ProtoCycle-7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Huggggooo/ProtoCycle-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Huggggooo/ProtoCycle-7B

SGLang

How to use Huggggooo/ProtoCycle-7B 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 "Huggggooo/ProtoCycle-7B" \
    --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": "Huggggooo/ProtoCycle-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "Huggggooo/ProtoCycle-7B" \
        --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": "Huggggooo/ProtoCycle-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Huggggooo/ProtoCycle-7B with Docker Model Runner:
```
docker model run hf.co/Huggggooo/ProtoCycle-7B
```

Huggggooo commited on Apr 18

Commit

d331fe7

verified ·

1 Parent(s): 23dedf4

Add model card

Browse files

Files changed (1) hide show

README.md +88 -0

README.md ADDED Viewed

	@@ -0,0 +1,88 @@

+---
+license: apache-2.0
+library_name: transformers
+pipeline_tag: text-generation
+base_model: Huggggooo/ProtoCycle-7B-SFT
+tags:
+  - protein-design
+  - agentic
+  - tool-use
+  - qwen2.5
+  - reinforcement-learning
+  - grpo
+language:
+  - en
+---
+# ProtoCycle-7B
+RL checkpoint for **ProtoCycle** — an agentic protein design model that
+performs multi-step, tool-augmented sequence design.
+This is the **GRPO-TCR (Group Relative Policy Optimization with Tool-Call
+Reward) stage**, initialised from the SFT checkpoint
+[`Huggggooo/ProtoCycle-7B-SFT`](https://huggingface.co/Huggggooo/ProtoCycle-7B-SFT).
+It corresponds to `global_step_20` of the `save10` run.
+- Base model: `Huggggooo/ProtoCycle-7B-SFT`
+  (itself fine-tuned from `Qwen/Qwen2.5-7B-Instruct`)
+- Training framework: [VeRL](https://github.com/volcengine/verl) /
+  [Open-AgentRL](https://github.com/Gen-Verse/Open-AgentRL)
+- Stage: agentic RL with GRPO-TCR
+- Rollouts per prompt: 8, max turns: 16
+- Max prompt / response: 8k / 20k tokens
+- Reward manager: `protein` (see
+  [ProtoCycle/verl/workers/reward_manager/protein.py](https://github.com/huggggoooooo/ProtoCycle/blob/main/verl/workers/reward_manager/protein.py))
+## Reward Signal
+The GRPO-TCR reward combines:
+1. **Protocol compliance** — `<think>` / `<plan>` / `<tool_call>` / `<answer>`
+   format.
+2. **Tool-call quality** — per-step ProTrek score of the best sequence after
+   each Stage-1 / Stage-3 tool call.
+3. **Outcome signal** — final (global-best) ProTrek score vs the requirement.
+4. **Efficiency** — penalty for excessively long rollouts.
+See
+[`recipe/protein/reward.py`](https://github.com/huggggoooooo/ProtoCycle/blob/main/recipe/protein/reward.py)
+for the exact formulation.
+## Agent Protocol
+```
+<think>  ... reasoning ...  </think>
+<plan>   ... stage plan ...  </plan>
+<tool_call>{"name": "...", "arguments": {...}}</tool_call>
+...
+<answer>MAEGEITPLKTF...</answer>
+```
+Registered tools (10):
+- **Stage 1:** `function2seq`, `pathway2seq`, `domain2seq`, `go2seq`,
+  `dna_binding2seq`
+- **Stage 2:** `cofactor2constraints`, `motif2constraints`,
+  `signal2constraints`
+- **Stage 3:** `esm_inpaint`, `get_score`
+## How to Use
+Inference requires the biology tools shipped in the
+[ProtoCycle](https://github.com/huggggoooooo/ProtoCycle) repo:
+```bash
+export MODEL_DIR=/path/to/ProtoCycle-7B       # this checkpoint
+export MODEL_NAME=ProtoCycle-7B
+export CONDA_ROOT=/path/to/miniconda3
+# Also export PROTREK_*_DIR, ESM_MODEL_PATH, etc. (see ProtoCycle README)
+bash infer_tools.sh
+```
+## License
+Apache-2.0.
+## Citation
+If you find this work useful, please cite ProtoCycle (forthcoming) and the
+upstream frameworks: VeRL, Open-AgentRL, ProTrek, ESM.