با استفاده از کتابخانه Gemma، به صورت تصویری و متنی اعلان ارسال کنید

pip install -q gemma

# This will prompt you to enter your Kaggle API token
import kagglehub
kagglehub.login()

VBox(children=(HTML(value='<center> <img\nsrc=https://www.kaggle.com/static/images/site-logo.png\nalt=\'Kaggle…
Kaggle credentials set.
Warning: Looks like you're using an outdated `kagglehub` version (installed: 0.3.13), please consider upgrading to the latest version (0.4.1).
Kaggle credentials successfully validated.

import os
os.environ["XLA_PYTHON_CLIENT_MEM_FRACTION"]="1.00"

# Common imports
import os
import jax
import jax.numpy as jnp
import tensorflow_datasets as tfds

# Gemma imports
from gemma import gm

tokenizer = gm.text.Gemma3Tokenizer()

model = gm.nn.Gemma3_4B()

params = gm.ckpts.load_params(gm.ckpts.CheckpointPath.GEMMA3_4B_IT)

WARNING:absl:Provided metadata contains unknown key custom. Adding it to custom_metadata.

sampler = gm.text.Sampler(
    model=model,
    params=params,
    tokenizer=tokenizer,
)

sampler.sample('Roses are red.', max_new_tokens=30)

'\nViolets are blue.\nI love you,\nAnd I want you too.\n\n---\n\nThis is a classic, simple, and'

ds = tfds.data_source('oxford_flowers102', split='train')
image = ds[0]['image']

# display the image
image

WARNING:absl:Variant folder /root/tensorflow_datasets/oxford_flowers102/2.1.1 has no dataset_info.json
Downloading and preparing dataset Unknown size (download: Unknown size, generated: Unknown size, total: Unknown size) to /root/tensorflow_datasets/oxford_flowers102/2.1.1...
Dl Completed...: 0 url [00:00, ? url/s]
Dl Size...: 0 MiB [00:00, ? MiB/s]
Extraction completed...: 0 file [00:00, ? file/s]
Generating splits...:   0%|          | 0/3 [00:00<?, ? splits/s]
Generating train examples...: 0 examples [00:00, ? examples/s]
Shuffling /root/tensorflow_datasets/oxford_flowers102/incomplete.2TXIHE_2.1.1/oxford_flowers102-train.array_re…
Generating test examples...: 0 examples [00:00, ? examples/s]
Shuffling /root/tensorflow_datasets/oxford_flowers102/incomplete.2TXIHE_2.1.1/oxford_flowers102-test.array_rec…
Generating validation examples...: 0 examples [00:00, ? examples/s]
Shuffling /root/tensorflow_datasets/oxford_flowers102/incomplete.2TXIHE_2.1.1/oxford_flowers102-validation.arr…
Dataset oxford_flowers102 downloaded and prepared to /root/tensorflow_datasets/oxford_flowers102/2.1.1. Subsequent calls will reuse this data.

prompt = """<start_of_turn>user
Describe the contents of this image.

<start_of_image>

<end_of_turn>
<start_of_turn>model
"""

sampler = gm.text.Sampler(
    model=model,
    params=params,
    tokenizer=tokenizer,
)

out = sampler.sample(prompt, images=image, max_new_tokens=500)
print(out)

/usr/local/lib/python3.12/dist-packages/jax/_src/ops/scatter.py:108: FutureWarning: scatter inputs have incompatible types: cannot safely cast value from dtype=float32 to dtype=bfloat16 with jax_numpy_dtype_promotion='standard'. In future JAX releases this will result in an error.
  warnings.warn(
Here's a description of the image's contents:

**Subject:** The image features a close-up shot of a water lily in full bloom.

**Details:**

*   **Flower:** The lily is predominantly white with a delicate pink hue at the base of the petals. It has a striking, star-like shape with pointed petals that curve upwards.
*   **Center:** The flower's center is a vibrant yellow, with prominent stamens extending outwards. There are a few water droplets clinging to the petals.
*   **Stem and Pads:** The lily is resting on a dark, textured stem and broad, paddle-shaped pads that are characteristic of water lilies.
*   **Background:** The background is completely black, creating a dramatic contrast and isolating the flower, making it the clear focal point.

**Overall Impression:** The image has a serene and elegant feel, emphasizing the beauty and detail of the water lily. The dark background and lighting create a sense of depth and highlight the flower's form.<end_of_turn>