GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
omerfaksoy
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
omerfaksoy
Model Tree
google/paligemma-3b-pt-224
Input Modalities
Output Modalities
Supported Functionality
Evaluation was run on the TR-DocVQA-Synth test split with 2,000 examples.
| Model | Setting | Test Samples | Normalized EM | ANLS | Token F1 | Empty Prediction Rate | Invalid Prediction Rate |
|---|---|---|---|---|---|---|---|
| PaliGemma-3B LoRA | Fine-tuned LoRA | 2000 | 0.7205 | 0.8745 | 0.7294 | 0.0000 | 0.0000 |
Additional paper-ready metrics, per-field breakdowns, error analysis, and LaTeX tables are included under evaluation/.
python
import torchfrom PIL import Imagefrom transformers import AutoProcessor, PaliGemmaForConditionalGenerationfrom peft import PeftModelbase_model = "google/paligemma-3b-pt-224"adapter_id = "omerfaksoy/trdocvqa-paligemma-3b-lora"processor = AutoProcessor.from_pretrained(adapter_id)model = PaliGemmaForConditionalGeneration.from_pretrained(base_model,torch_dtype=torch.bfloat16,device_map="auto",)model = PeftModel.from_pretrained(model, adapter_id)model.eval()image = Image.open("document.png").convert("RGB")question = "Toplam tutar nedir?"prompt = f"answer tr {question}\n"inputs = processor(text=prompt, images=image, return_tensors="pt").to(model.device)with torch.inference_mode():generated = model.generate(**inputs, max_new_tokens=64, do_sample=False, num_beams=1)prompt_len = inputs["input_ids"].shape[-1]answer = processor.batch_decode(generated[:, prompt_len:], skip_special_tokens=True)[0].strip()print(answer)
google/paligemma-3b-pt-224Ethosoft/TR-DocVQA-SynthThis repository contains a LoRA adapter, not a full merged copy of the base PaliGemma model. Users must comply with the terms of the base model and accept any gated access requirements for google/paligemma-3b-pt-224.
The model was developed for research use on synthetic Turkish document VQA data. Before production use, evaluate on real documents from the target domain and review privacy, licensing, and bias considerations.