Instructions to use Chokun00032/gemma-1b-pruned-th with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Chokun00032/gemma-1b-pruned-th with Transformers:

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

pipe = pipeline("text-generation", model="Chokun00032/gemma-1b-pruned-th")

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("Chokun00032/gemma-1b-pruned-th")
model = AutoModelForMultimodalLM.from_pretrained("Chokun00032/gemma-1b-pruned-th")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Chokun00032/gemma-1b-pruned-th with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Chokun00032/gemma-1b-pruned-th"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Chokun00032/gemma-1b-pruned-th",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Chokun00032/gemma-1b-pruned-th

SGLang

How to use Chokun00032/gemma-1b-pruned-th 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 "Chokun00032/gemma-1b-pruned-th" \
    --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": "Chokun00032/gemma-1b-pruned-th",
		"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 "Chokun00032/gemma-1b-pruned-th" \
        --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": "Chokun00032/gemma-1b-pruned-th",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Chokun00032/gemma-1b-pruned-th with Docker Model Runner:
```
docker model run hf.co/Chokun00032/gemma-1b-pruned-th
```

gemma-1b-pruned-th

Depth-pruned (layer dropping) + healing SFT of google/gemma-3-4b (unsloth mirror) for Thai.

Spec

Base: google/gemma-3-4b (unsloth mirror)
Params: 2.70B (kept 17/34 decoder layers; drop middle, keep head+tail)
Healing: SFT on SEA-PILE v2 Thai (~8k docs), bf16
Requires: transformers>=4.50, accelerate

Usage

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
m = "Chokun00032/gemma-1b-pruned-th"
tok = AutoTokenizer.from_pretrained(m)
model = AutoModelForCausalLM.from_pretrained(m, torch_dtype=torch.bfloat16, device_map="cuda")
ids = tok("ปัญญาประดิษฐ์ คือ", return_tensors="pt").to(model.device)
out = model.generate(**ids, max_new_tokens=120, do_sample=True,
                     temperature=0.7, top_p=0.9, repetition_penalty=1.3)
print(tok.decode(out[0], skip_special_tokens=True))

Notes

Pruned base healed on raw corpus: Thai grammar is fluent, but factual/arithmetic ability is weak.
Use repetition_penalty>=1.2 to avoid loops.
Best used as a base for further instruction fine-tuning.

Downloads last month: 43