GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
AGViveiros
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
AGViveiros
Model Tree
Qwen/Qwen2.5-VL-3B-Instruct
Input Modalities
Output Modalities
Supported Functionality
Table 3 in the paper was trained and evaluated with grayscale intermediate rotation strips, as described in the paper/code. Afterwards, we ran additional experiments with colorized intermediate images and found little impact on task accuracy, but more variation in the oracle latent representations. We standardized on the colorized setup afterwards, and this checkpoint comes from that run — not the grayscale one Table 3 reports.
Data: AGViveiros/LanteRn-Tetris-Data
LantErn extends Qwen2.5-VL-3B-Instruct with
Latent Visual Reasoning (LVR) tokens. Instead of always verbalising what it sees, the model can emit
compressed visual embeddings (<|lvr_start|>…<|lvr_end|>) during its chain-of-thought, enabling
non-verbalized visual reasoning interleaved with text.
Special tokens:
| Token | Role |
|---|---|
<lvr_start> | Begin a latent visual reasoning block |
<lvr_sep> | Placeholder replaced by compressed visual embeddings (8 tokens) |
<lvr_end> | End a latent visual reasoning block |
Codebase: github.com/GuilhermeViveiros/LantErn
bash
git clone https://github.com/GuilhermeViveiros/LantErn.gitcd LantErnpip install -r requirements.txtpip install -e .python
import torchfrom PIL import Imagefrom qwen_vl_utils import process_vision_info from src.lantern_generate.generate import generate as lantern_generatefrom src.models import load_model # ── 1. Load model + processor ─────────────────────────────────────────────────device = "cuda" if torch.cuda.is_available() else "cpu"model, processor = load_model("AGViveiros/LanteRn-3B-Tetris", compute_dtype=torch.bfloat16, use_cache=True)model.eval().to(device)processor.tokenizer.padding_side = "left" # ── 2. Build inputs ───────────────────────────────────────────────────────────image = Image.open("path/to/image.jpg").convert("RGB")question = "Your question here"messages = [{ "role": "user", "content": [ {"type": "image", "image": image}, {"type": "text", "text": question}, ],}]text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)image_inputs, _ = process_vision_info(messages)inputs = processor(text=[text], images=image_inputs, return_tensors="pt").to(device)prompt_len = inputs["input_ids"].shape[1] # ── 3. Generate with latent visual reasoning ──────────────────────────────────output = model.generate( **inputs, max_new_tokens=512, do_sample=False, custom_generate=lantern_generate, use_cache=True, return_dict_in_generate=True,) generated = output.sequences[0][prompt_len:]print(processor.decode(generated, skip_special_tokens=False))bibtex
@article{viveiros2026holding, title={What's Holding Back Latent Visual Reasoning?}, author={Viveiros, Andr{\'e} G and Gon{\c{c}}alves, Nuno and Martins, Andr{\'e} FT and Lindemann, Matthias}, journal={arXiv preprint arXiv:2605.18445}, year={2026}}git clone https://github.com/GuilhermeViveiros/LantErn.gitcd LantErnpip install -r requirements.txtpip install -e .import torchfrom PIL import Imagefrom qwen_vl_utils import process_vision_info from src.lantern_generate.generate import generate as lantern_generatefrom src.models import load_model # ── 1. Load model + processor ─────────────────────────────────────────────────device = "cuda" if torch.cuda.is_available() else "cpu"model, processor = load_model("AGViveiros/LanteRn-3B-Tetris", compute_dtype=torch.bfloat16, use_cache=True)model.eval().to(device)processor.tokenizer.padding_side = "left" # ── 2. Build inputs ───────────────────────────────────────────────────────────image = Image.open("path/to/image.jpg").convert("RGB")question = "Your question here"messages = [{ "role": "user", "content": [ {"type": "image", "image": image}, {"type": "text", "text": question}, ],}]text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)image_inputs, _ = process_vision_info(messages)inputs = processor(text=[text], images=image_inputs, return_tensors="pt").to(device)prompt_len = inputs["input_ids"].shape[1] # ── 3. Generate with latent visual reasoning ──────────────────────────────────output = model.generate( **inputs, max_new_tokens=512, do_sample=False, custom_generate=lantern_generate, use_cache=True, return_dict_in_generate=True,) generated = output.sequences[0][prompt_len:]print(processor.decode(generated, skip_special_tokens=False))@article{viveiros2026holding, title={What's Holding Back Latent Visual Reasoning?}, author={Viveiros, Andr{\'e} G and Gon{\c{c}}alves, Nuno and Martins, Andr{\'e} FT and Lindemann, Matthias}, journal={arXiv preprint arXiv:2605.18445}, year={2026}}