Instructions to use OmAhire369/safe-genai-ppo-full with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OmAhire369/safe-genai-ppo-full with Transformers:

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

pipe = pipeline("text-generation", model="OmAhire369/safe-genai-ppo-full")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("OmAhire369/safe-genai-ppo-full")
model = AutoModelForCausalLM.from_pretrained("OmAhire369/safe-genai-ppo-full", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use OmAhire369/safe-genai-ppo-full with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OmAhire369/safe-genai-ppo-full"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OmAhire369/safe-genai-ppo-full",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/OmAhire369/safe-genai-ppo-full

SGLang

How to use OmAhire369/safe-genai-ppo-full 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 "OmAhire369/safe-genai-ppo-full" \
    --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": "OmAhire369/safe-genai-ppo-full",
		"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 "OmAhire369/safe-genai-ppo-full" \
        --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": "OmAhire369/safe-genai-ppo-full",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use OmAhire369/safe-genai-ppo-full with Docker Model Runner:
```
docker model run hf.co/OmAhire369/safe-genai-ppo-full
```

safe-genai-ppo-full

File size: 828 Bytes

ecd48aa

{
  "task": "ppo",
  "strategy": "full",
  "base_model": "gpt2-medium",
  "reward_model": "OmAhire369/safe-genai-reward-full",
  "n_train_prompts": 4000,
  "ppo_steps": 120,
  "batch_size": 32,
  "learning_rate": 1.41e-05,
  "param_stats": {
    "total_params": 354823168,
    "trainable_params": 354823168,
    "trainable_pct": 100.0,
    "total_M": 354.82,
    "trainable_M": 354.823
  },
  "build_notes": [],
  "reward_start": -4.85760498046875,
  "reward_final": -1.817138671875,
  "reward_best": 0.2688255310058594,
  "best_probe_tag": "step-80",
  "used_checkpoint": "best",
  "reward_delta": 3.04046630859375,
  "stopped_early": false,
  "length_collapsed": true,
  "length_start": 46.96875,
  "length_final": 48.0,
  "train_seconds": 1751.58,
  "peak_gpu_mb": 9386.4,
  "data_source": "hf:nrizwan/safe_ai_assignment_1"
}