Name: transformers
Version: 5.5.0
Summary: Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
Home-page: https://github.com/huggingface/transformers
Author: The Hugging Face team (past and future) with the help of all our contributors (https://github.com/huggingface/transformers/graphs/contributors)
Author-email: transformers@huggingface.co
License: Apache 2.0 License
Location: /usr/local/lib/python3.13/dist-packages
Requires: huggingface-hub, numpy, packaging, pyyaml, regex, safetensors, tokenizers, tqdm, typer
Required-by: peft, sentence-transformers, trl, unsloth, unsloth_zoo
---
Name: torch
Version: 2.11.0+cu128
Summary: Tensors and Dynamic neural networks in Python with strong GPU acceleration
Home-page: https://pytorch.org
Author: 
Author-email: PyTorch Team <packages@pytorch.org>
License: BSD-3-Clause
Location: /usr/local/lib/python3.13/dist-packages
Requires: cuda-bindings, cuda-toolkit, filelock, fsspec, jinja2, networkx, nvidia-cudnn-cu12, nvidia-cusparselt-cu12, nvidia-nccl-cu12, nvidia-nvshmem-cu12, setuptools, sympy, triton, typing-extensions
Required-by: accelerate, bitsandbytes, cut-cross-entropy, fastai, peft, sentence-transformers, timm, torchdata, torchvision, unsloth, unsloth_zoo, xformers
---
Name: unsloth
Version: 2026.8.19
Summary: 2-5X faster training, reinforcement learning & finetuning
Home-page: https://unsloth.ai
Author: Unsloth AI team
Author-email: info@unsloth.ai
License: 
Location: /usr/local/lib/python3.13/dist-packages
Requires: accelerate, bitsandbytes, click, datasets, diffusers, hf_transfer, huggingface_hub, nest-asyncio, numpy, packaging, peft, protobuf, psutil, pydantic, pyyaml, rich, sentencepiece, structlog, torch, torchvision, tqdm, transformers, triton, trl, typer, tyro, unsloth_zoo, wheel, xformers
Required-by: 
---
Name: bitsandbytes
Version: 0.50.1
Summary: k-bit optimizers and matrix multiplication routines.
Home-page: https://github.com/bitsandbytes-foundation/bitsandbytes
Author: 
Author-email: Tim Dettmers <dettmers@cs.washington.edu>
License: 
Location: /usr/local/lib/python3.13/dist-packages
Requires: numpy, packaging, torch
Required-by: unsloth

INFERENCE

python

import sys
import os
import contextlib
from PIL import Image

# 1. Download image using wget and load it
image_url = "https://picsum.photos/300/300"
image_filename = "test_image.jpg"
os.system(f"wget -q -O {image_filename} {image_url}")

image = Image.open(image_filename).convert("RGB")

python

import sys
import os
import contextlib

# Suppress all C/C++/Python low-level file descriptor prints during imports
@contextlib.contextmanager
def suppress_all_output():
    with open(os.devnull, "w") as devnull:
        old_stdout = sys.stdout
        old_stderr = sys.stderr
        sys.stdout = devnull
        sys.stderr = devnull
        try:
            yield
        finally:
            sys.stdout = old_stdout
            sys.stderr = old_stderr

# Completely silence unsloth/transformers startup output and progress bars
os.environ["UNSLOTH_DISABLE_LOGGING"] = "1"
os.environ["TRANSVERSE_NO_PROGRESS_BARS"] = "1"
os.environ["TQDM_DISABLE"] = "1"

with suppress_all_output():
    import torch
    import numpy as np
    
    # Globally enforce weights_only=False for PyTorch 2.6+ checkpoint loading
    original_torch_load = torch.load
    def patched_torch_load(*args, **kwargs):
        kwargs["weights_only"] = False
        return original_torch_load(*args, **kwargs)
    torch.load = patched_torch_load

    from huggingface_hub import hf_hub_download
    from PIL import Image
    from unsloth import FastVisionModel

MODEL_ID = "frankmorales2020/topo-gemma-4-e4b-vision-13tasks"

with suppress_all_output():
    ckpt_path = hf_hub_download(repo_id=MODEL_ID, filename="pytorch_model.bin")
    checkpoint = torch.load(ckpt_path, map_location="cpu")
    BASE_MODEL = checkpoint.get("base_model", "frankmorales2020/gemma-4-e4b-unesco-optimized")

    model, tokenizer = FastVisionModel.from_pretrained(
        model_name=BASE_MODEL,
        load_in_4bit=True,
        dtype=torch.bfloat16,
    )
    FastVisionModel.for_inference(model)

# 1. Load test image

# 1.1. Download image using wget and load it
from PIL import Image
image_url = "https://picsum.photos/300/300"
image_filename = "test_image.jpg"
os.system(f"wget -q -O {image_filename} {image_url}")

image = Image.open("test_image.jpg").convert("RGB")

# 2. Define all 13 tasks
tasks = [
    ("Task A", "Animal vs Vehicle", "Does this image depict an animal or a vehicle?"),
    ("Task B", "Natural vs Man-Made", "Is this subject natural or man-made?"),
    ("Task C", "Living vs Non-Living", "Is the primary subject living or non-living?"),
    ("Task D", "Large vs Small", "Is the subject large or small in scale?"),
    ("Task E", "Ground vs Air/Water", "Does this subject belong to ground or air/water?"),
    ("Task F", "Domestic vs Wild", "Is this subject domestic or wild?"),
    ("Task G", "Mammal vs Non-Mammal", "Is this subject a mammal or non-mammal?"),
    ("Task H", "Flying vs Non-Flying", "Is this subject flying or non-flying?"),
    ("Task I", "Fast vs Slow", "Is this subject characterized as fast or slow?"),
    ("Task J", "Urban vs Rural", "Does this setting represent an urban or rural environment?"),
    ("Task K", "Predator vs Prey", "Is this subject a predator or prey?"),
    ("Task L", "Nocturnal vs Diurnal", "Is this subject nocturnal or diurnal?"),
    ("Task M", "Domesticated vs Wild Animals", "Is this animal domesticated or wild?")
]

print("\n" + "="*80)
print("🚀 EVALUATING ALL 13 TOPO-2026 TASKS")
print("="*80)

for task_id, task_name, prompt in tasks:
    messages = [
        {
            "role": "user",
            "content": [
                {"type": "image"},
                {"type": "text", "text": f"{task_id} ({task_name}): {prompt}"}
            ]
        }
    ]
    input_text = tokenizer.apply_chat_template(messages, add_generation_prompt=True)
    
    inputs = tokenizer(
        image,
        input_text,
        add_special_tokens=False,
        return_tensors="pt",
    ).to("cuda")

    with torch.inference_mode():
        output_tokens = model.generate(
            **inputs,
            max_new_tokens=24,
            do_sample=False,
            use_cache=True,
        )

    response = tokenizer.decode(output_tokens[0], skip_special_tokens=True)
    answer = response.split("model")[-1].strip() if "model" in response else response
    print(f"[{task_id}] {task_name:<30} ➔ {answer}")

print("="*80)
print("🎉 EVALUATION COMPLETE!")
print("="*80)

EXPECTED - INFERENCE - OUTPUT

text

 Loading weights: 100% 2130/2130 [00:03<00:00, 1004.30it/s]
================================================================================
🚀 EVALUATING ALL 13 TOPO-2026 TASKS
================================================================================
[Task A] Animal vs Vehicle              ➔ This image depicts **neither** an animal nor a vehicle. It is a landscape photograph of the **ocean/sea**
[Task B] Natural vs Man-Made            ➔ This subject is **natural**.

It depicts a seascape with waves, ocean, and a distant landmass under a dramatic
[Task C] Living vs Non-Living           ➔ The primary subject in the image is the **ocean/sea** and the **sky/weather**.

Both the ocean
[Task D] Large vs Small                 ➔ Based on the image, the **subject** (the ocean, waves, and coastline) is **large in scale**.
[Task E] Ground vs Air/Water            ➔ This subject belongs to **both ground and air/water**.

Here's why:

* **Water:** The
[Task F] Domestic vs Wild               ➔ This subject is **wild**.

The image depicts a natural scene: the ocean, waves, and the sky. These
[Task G] Mammal vs Non-Mammal           ➔ Based on the image provided, there is **no subject** visible that is an animal. The image is a landscape photograph
[Task H] Flying vs Non-Flying           ➔ Based on the image provided, there is **no subject** that is clearly flying or non-flying.

The image
[Task I] Fast vs Slow                   ➔ Based on the image, the subject matter is a **seascape** (ocean waves, sky, and coastline).
[Task J] Urban vs Rural                 ➔ This setting represents a **rural** environment.

Here's why:

* **Natural Landscape:** The image is
[Task K] Predator vs Prey               ➔ Based on the image provided, which is a **landscape photograph of the ocean at sunset/sunrise**, there are **no
[Task L] Nocturnal vs Diurnal           ➔ Based on the image, the subject is a **seascape** (ocean, waves, sky).

The concept of
[Task M] Domesticated vs Wild Animals   ➔ I'm sorry, but you have provided an image of a **seascape (ocean waves and sky)**, not
================================================================================
🎉 EVALUATION COMPLETE!
================================================================================