GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
0PValencia
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
0PValencia
Model Tree
Qwen/Qwen3-8B
Input Modalities
Output Modalities
Supported Functionality
| Skill | Descripción |
|---|---|
| extraction | Texto libre → JSON de oportunidad |
| classification | Tipo: beca, pasantía, curso, evento, empleo junior, voluntariado |
| matching | Compatibilidad usuario ↔ oportunidad (alto/medio/bajo/nulo + score + reason) |
| decision | Acción del agente: send_alert / skip / watch |
python
import torchfrom peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfigbase_id = "Qwen/Qwen3-8B"adapter_id = "0PValencia/oppy-qwen3-lora"bnb = BitsAndBytesConfig(load_in_4bit=True,bnb_4bit_quant_type="nf4",bnb_4bit_use_double_quant=True,bnb_4bit_compute_dtype=torch.float16,)tokenizer = AutoTokenizer.from_pretrained(base_id, trust_remote_code=True)model = AutoModelForCausalLM.from_pretrained(base_id,quantization_config=bnb,device_map="auto",trust_remote_code=True,)model = PeftModel.from_pretrained(model, adapter_id)model.eval()messages = [{"role": "system","content": ("Eres Oppy, un agente que encuentra y recomienda oportunidades ""en Bolivia. Cuando pidan estructura, responde solo JSON válido."),},{"role": "user","content": ("Clasifica esta oportunidad. Responde solo con JSON válido.\n\n""Convocatoria de beca para estudiantes universitarios en Bolivia."),},]prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True, enable_thinking=False)inputs = tokenizer(prompt, return_tensors="pt").to(model.device)out = model.generate(**inputs, max_new_tokens=256, do_sample=False)print(tokenizer.decode(out[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))
0PValencia/oppy-bolivia-sftQwen/Qwen3-8BApache-2.0 (respetar también la licencia del modelo base Qwen3).