GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
hiyasvyas
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
hiyasvyas
Model Tree
unsloth/Qwen2.5-VL-3B-Instruct-bnb-4bit
Input Modalities
Output Modalities
Supported Functionality
Given an image of a 3D unit-cube solid, output exactly three ASCII grids labeled `top:`, `front:`, `right:` — using only `.` and `#`, with no prose before, between, or after the grids.
A well-prompted base VLM cannot do this reliably; it never even produces the three-grid format. Fine-tuning fixes the format completely and improves the spatial content.
| metric | BASE | TUNED | delta |
|---|---|---|---|
| `format_ok` | 0.0% | 100.0% | +100.0 |
| `top_ok` | 0.0% | 7.0% | +7.0 |
| `front_ok` | 0.0% | 25.0% | +25.0 |
| `right_ok` | 0.0% | 20.0% | +20.0 |
| `exact` | 0.0% | 3.0% | +3.0 |
Read: the base model scores zero everywhere — it cannot produce the format at all. The tuned model produces the correct 3-grid format 100% of the time and gets individual views right a meaningful fraction of the time. Format adherence (the target behavior) is fully instilled; per-view spatial accuracy is the next data-iteration target.
```python from transformers import AutoProcessor, Qwen2_5_VLForConditionalGeneration from peft import PeftModel from PIL import Image import torch
base = "unsloth/Qwen2.5-VL-3B-Instruct-bnb-4bit" model = Qwen2_5_VLForConditionalGeneration.from_pretrained(base, torch_dtype=torch.float16, device_map="auto") model = PeftModel.from_pretrained(model, "hiyasvyas/ortho-vision-lora") processor = AutoProcessor.from_pretrained("hiyasvyas/ortho-vision-lora")
prompt = ("Look at this 3D object built from unit cubes. Output its third-angle " "orthographic views as three ASCII grids labeled 'top:', 'front:', and " "'right:', using only '.' and '#'. Output nothing except the three grids.") img = Image.open("your_block_image.png").convert("RGB") msgs = [{"role": "user", "content": [{"type": "image", "image": img}, {"type": "text", "text": prompt}]}] inputs = processor.apply_chat_template(msgs, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt").to(model.device) out = model.generate(**inputs, max_new_tokens=256) print(processor.decode(out[0], skip_special_tokens=True)) ```
Per-view spatial accuracy is still low (`exact` = 3%). `top_ok` (7%) lags `front`/`right`, suggesting the top-down footprint is hardest to read from the rendered perspective — a data/rendering signal to fix in the next iteration, not a hyperparameter problem.