GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
Chang-chih
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
Chang-chih
Model Tree
deepseek-ai/deepseek-llm-7b-chat
Input Modalities
Output Modalities
Supported Functionality
leviathan-16B-final (旗艦參謀)deepseek-ai/deepseek-llm-7b-chatdistillation_dataset.jsonl (336 筆，由導師模型深度解析產出)python
from transformers import AutoModelForCausalLM, AutoTokenizerimport torchmodel_id = "Chang-chih/leviathan-9B-MoE"tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)model = AutoModelForCausalLM.from_pretrained(model_id,device_map="auto",torch_dtype=torch.bfloat16,trust_remote_code=True)prompt = "你是誰？"inputs = tokenizer(prompt, return_tensors="pt").to("cuda")outputs = model.generate(**inputs, max_new_tokens=50)response = tokenizer.decode(outputs[0], skip_special_tokens=True)print(response)使用 Ollamabashollama run hf.co/Chang-chih/leviathan-9B-MoE "你的指令"硬體需求 (Hardware Requirements)GPU VRAM：建議 16GB 以上 (可使用 4-bit 量化)系統：Linux / Windows (透過 WSL)授權 (License)本模型採用 Apache-2.0 授權條款。引用 (Citation)如果您使用此模型，請引用：bibtex@misc{leviathan-9b-moe,author = {Chang-chih},title = {Leviathan-9B-MoE: A Mixture-of-Experts Model for Local Vision-Language Integration},year = {2026},publisher = {Hugging Face},url = {https://huggingface.co/Chang-chih/leviathan-9B-MoE}}