Instructions to use zoeythanayot/gemma3-it-grpo-thai with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use zoeythanayot/gemma3-it-grpo-thai with Transformers:

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

pipe = pipeline("text-generation", model="zoeythanayot/gemma3-it-grpo-thai")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("zoeythanayot/gemma3-it-grpo-thai", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use zoeythanayot/gemma3-it-grpo-thai with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "zoeythanayot/gemma3-it-grpo-thai"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "zoeythanayot/gemma3-it-grpo-thai",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/zoeythanayot/gemma3-it-grpo-thai

SGLang

How to use zoeythanayot/gemma3-it-grpo-thai 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 "zoeythanayot/gemma3-it-grpo-thai" \
    --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": "zoeythanayot/gemma3-it-grpo-thai",
		"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 "zoeythanayot/gemma3-it-grpo-thai" \
        --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": "zoeythanayot/gemma3-it-grpo-thai",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use zoeythanayot/gemma3-it-grpo-thai with Docker Model Runner:
```
docker model run hf.co/zoeythanayot/gemma3-it-grpo-thai
```

zoeythanayot commited on Sep 16, 2025

Commit

35c2ef4

verified ·

1 Parent(s): b7378bf

Update README.md

Browse files

Files changed (1) hide show

README.md +11 -17

README.md CHANGED Viewed

@@ -78,22 +78,13 @@ Below are the reward values observed during training:
 | Step | Reward |
 |------|--------|
-| 10   | 0.0005 |
-| 20   | 0.0016 |
-| 30   | 0.0028 |
-| 40   | 0.0043 |
-| 50   | 0.0048 |
-| 60   | 0.0046 |
-| 70   | 0.0046 |
-| 80   | 0.0048 |
-| 90   | 0.0048 |
-| 100  | 0.0048 |
-| 110  | 0.0049 |
-| 120  | 0.0050 |
-- Reward ค่อย ๆ เพิ่มขึ้นช่วงแรก (10 → 50)
-- Stabilize ที่ ~0.0048–0.0050 หลัง step 60
-- แสดงถึง convergence ของโมเดลต่อ reward function
 ---
@@ -132,7 +123,10 @@ with torch.inference_mode():
         top_p=0.9
     )
-print(tok.decode(output_ids[0], skip_special_tokens=True))
 ```
 ---

 | Step | Reward |
 |------|--------|
+| 100   | 0.0030 |
+| 200   | 0.0040 |
+| 280   | 0.0042 |
+- Reward มีแนวโน้ม ค่อย ๆ เพิ่มขึ้นต่อเนื่อง ในช่วงต้นการเทรน (Step 100 → 200 → 280)
+- ค่า Reward อยู่ราว ๆ 0.0030 → 0.0040 → 0.0042 แสดงถึงการปรับตัวของโมเดลตาม reward function
+- แนวโน้มชี้ว่าโมเดลกำลัง เข้าใกล้ภาวะเสถียร (convergence) แต่ยังไม่ถึงจุด plateau เหมือนกรณีที่ stabilize แถว ~0.0048–0.0050 หากเทรนต่อไปอีก มีโอกาสที่ค่า Reward จะคงที่ในระดับสูงขึ้น (plateau)
 ---
         top_p=0.9
     )
+input_length = inputs.shape[1]
+new_tokens = output_ids[0, input_length:]
+resp = tok.decode(new_tokens, skip_special_tokens=True)
+print(resp.strip())
 ```
 ---