Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,537 @@
+---
+license: gemma
+library_name: transformers
+pipeline_tag: image-text-to-text
+extra_gated_heading: Access Gemma on Hugging Face
+extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and
+  agree to Google’s usage license. To do this, please ensure you’re logged in to Hugging
+  Face and click below. Requests are processed immediately.
+extra_gated_button_content: Acknowledge license
+base_model: google/gemma-3-12b-it
+---
+
+# NVFP4 quantization of gemma-3-12b-it model using llm-compressor
+
+# Gemma 3 model card
+
+**Model Page**: [Gemma](https://ai.google.dev/gemma/docs/core)
+
+**Resources and Technical Documentation**:
+
+* [Gemma 3 Technical Report][g3-tech-report]
+* [Responsible Generative AI Toolkit][rai-toolkit]
+* [Gemma on Kaggle][kaggle-gemma]
+* [Gemma on Vertex Model Garden][vertex-mg-gemma3]
+
+**Terms of Use**: [Terms][terms]
+
+**Authors**: Google DeepMind
+
+## Model Information
+
+Summary description and brief definition of inputs and outputs.
+
+### Description
+
+Gemma is a family of lightweight, state-of-the-art open models from Google,
+built from the same research and technology used to create the Gemini models.
+Gemma 3 models are multimodal, handling text and image input and generating text
+output, with open weights for both pre-trained variants and instruction-tuned
+variants. Gemma 3 has a large, 128K context window, multilingual support in over
+140 languages, and is available in more sizes than previous versions. Gemma 3
+models are well-suited for a variety of text generation and image understanding
+tasks, including question answering, summarization, and reasoning. Their
+relatively small size makes it possible to deploy them in environments with
+limited resources such as laptops, desktops or your own cloud infrastructure,
+democratizing access to state of the art AI models and helping foster innovation
+for everyone.
+
+### Inputs and outputs
+
+-   **Input:**
+    -  Text string, such as a question, a prompt, or a document to be summarized
+    -  Images, normalized to 896 x 896 resolution and encoded to 256 tokens
+       each
+    -  Total input context of 128K tokens for the 4B, 12B, and 27B sizes, and
+       32K tokens for the 1B size
+
+-   **Output:**
+    -   Generated text in response to the input, such as an answer to a
+        question, analysis of image content, or a summary of a document
+    -   Total output context of 8192 tokens
+
+### Usage
+
+Below, there are some code snippets on how to get quickly started with running the model. First, install the Transformers library. Gemma 3 is supported starting from transformers 4.50.0. 
+
+```sh
+$ pip install -U transformers
+```
+
+Then, copy the snippet from the section that is relevant for your use case.
+
+#### Running with the `pipeline` API
+
+You can initialize the model and processor for inference with `pipeline` as follows.
+
+```python
+from transformers import pipeline
+import torch
+
+pipe = pipeline(
+    "image-text-to-text",
+    model="google/gemma-3-12b-it",
+    device="cuda",
+    torch_dtype=torch.bfloat16
+)
+```
+
+With instruction-tuned models, you need to use chat templates to process our inputs first. Then, you can pass it to the pipeline.
+
+```python
+messages = [
+    {
+        "role": "system",
+        "content": [{"type": "text", "text": "You are a helpful assistant."}]
+    },
+    {
+        "role": "user",
+        "content": [
+            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
+            {"type": "text", "text": "What animal is on the candy?"}
+        ]
+    }
+]
+
+output = pipe(text=messages, max_new_tokens=200)
+print(output[0]["generated_text"][-1]["content"])
+# Okay, let's take a look! 
+# Based on the image, the animal on the candy is a **turtle**. 
+# You can see the shell shape and the head and legs.
+```
+
+#### Running the model on a single / multi GPU
+
+```python
+# pip install accelerate
+
+from transformers import AutoProcessor, Gemma3ForConditionalGeneration
+from PIL import Image
+import requests
+import torch
+
+model_id = "google/gemma-3-12b-it"
+
+model = Gemma3ForConditionalGeneration.from_pretrained(
+    model_id, device_map="auto"
+).eval()
+
+processor = AutoProcessor.from_pretrained(model_id)
+
+messages = [
+    {
+        "role": "system",
+        "content": [{"type": "text", "text": "You are a helpful assistant."}]
+    },
+    {
+        "role": "user",
+        "content": [
+            {"type": "image", "image": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
+            {"type": "text", "text": "Describe this image in detail."}
+        ]
+    }
+]
+
+inputs = processor.apply_chat_template(
+    messages, add_generation_prompt=True, tokenize=True,
+    return_dict=True, return_tensors="pt"
+).to(model.device, dtype=torch.bfloat16)
+
+input_len = inputs["input_ids"].shape[-1]
+
+with torch.inference_mode():
+    generation = model.generate(**inputs, max_new_tokens=100, do_sample=False)
+    generation = generation[0][input_len:]
+
+decoded = processor.decode(generation, skip_special_tokens=True)
+print(decoded)
+
+# **Overall Impression:** The image is a close-up shot of a vibrant garden scene, 
+# focusing on a cluster of pink cosmos flowers and a busy bumblebee. 
+# It has a slightly soft, natural feel, likely captured in daylight.
+```
+
+### Citation
+
+```none
+@article{gemma_2025,
+    title={Gemma 3},
+    url={https://goo.gle/Gemma3Report},
+    publisher={Kaggle},
+    author={Gemma Team},
+    year={2025}
+}
+```
+
+## Model Data
+
+Data used for model training and how the data was processed.
+
+### Training Dataset
+
+These models were trained on a dataset of text data that includes a wide variety
+of sources. The 27B model was trained with 14 trillion tokens, the 12B model was
+trained with 12 trillion tokens, 4B model was trained with 4 trillion tokens and
+1B with 2 trillion tokens. Here are the key components:
+
+-   Web Documents: A diverse collection of web text ensures the model is
+    exposed to a broad range of linguistic styles, topics, and vocabulary. The
+    training dataset includes content in over 140 languages.
+-   Code: Exposing the model to code helps it to learn the syntax and
+    patterns of programming languages, which improves its ability to generate
+    code and understand code-related questions.
+-   Mathematics: Training on mathematical text helps the model learn logical
+    reasoning, symbolic representation, and to address mathematical queries.
+-   Images: A wide range of images enables the model to perform image
+    analysis and visual data extraction tasks.
+
+The combination of these diverse data sources is crucial for training a powerful
+multimodal model that can handle a wide variety of different tasks and data
+formats.
+
+### Data Preprocessing
+
+Here are the key data cleaning and filtering methods applied to the training
+data:
+
+-   CSAM Filtering: Rigorous CSAM (Child Sexual Abuse Material) filtering
+    was applied at multiple stages in the data preparation process to ensure
+    the exclusion of harmful and illegal content.
+-   Sensitive Data Filtering: As part of making Gemma pre-trained models
+    safe and reliable, automated techniques were used to filter out certain
+    personal information and other sensitive data from training sets.
+-   Additional methods: Filtering based on content quality and safety in
+    line with [our policies][safety-policies].
+
+## Implementation Information
+
+Details about the model internals.
+
+### Hardware
+
+Gemma was trained using [Tensor Processing Unit (TPU)][tpu] hardware (TPUv4p,
+TPUv5p and TPUv5e). Training vision-language models (VLMS) requires significant
+computational power. TPUs, designed specifically for matrix operations common in
+machine learning, offer several advantages in this domain:
+
+-   Performance: TPUs are specifically designed to handle the massive
+    computations involved in training VLMs. They can speed up training
+    considerably compared to CPUs.
+-   Memory: TPUs often come with large amounts of high-bandwidth memory,
+    allowing for the handling of large models and batch sizes during training.
+    This can lead to better model quality.
+-   Scalability: TPU Pods (large clusters of TPUs) provide a scalable
+    solution for handling the growing complexity of large foundation models.
+    You can distribute training across multiple TPU devices for faster and more
+    efficient processing.
+-   Cost-effectiveness: In many scenarios, TPUs can provide a more
+    cost-effective solution for training large models compared to CPU-based
+    infrastructure, especially when considering the time and resources saved
+    due to faster training.
+-   These advantages are aligned with
+    [Google's commitments to operate sustainably][sustainability].
+
+### Software
+
+Training was done using [JAX][jax] and [ML Pathways][ml-pathways].
+
+JAX allows researchers to take advantage of the latest generation of hardware,
+including TPUs, for faster and more efficient training of large models. ML
+Pathways is Google's latest effort to build artificially intelligent systems
+capable of generalizing across multiple tasks. This is specially suitable for
+foundation models, including large language models like these ones.
+
+Together, JAX and ML Pathways are used as described in the
+[paper about the Gemini family of models][gemini-2-paper]; *"the 'single
+controller' programming model of Jax and Pathways allows a single Python
+process to orchestrate the entire training run, dramatically simplifying the
+development workflow."*
+
+## Evaluation
+
+Model evaluation metrics and results.
+
+### Benchmark Results
+
+These models were evaluated against a large collection of different datasets and
+metrics to cover different aspects of text generation:
+
+#### Reasoning and factuality
+
+| Benchmark                      | Metric         | Gemma 3 PT 1B  | Gemma 3 PT 4B | Gemma 3 PT 12B | Gemma 3 PT 27B |
+| ------------------------------ |----------------|:--------------:|:-------------:|:--------------:|:--------------:|
+| [HellaSwag][hellaswag]         | 10-shot        |      62.3      |      77.2     |      84.2      |      85.6      |
+| [BoolQ][boolq]                 | 0-shot         |      63.2      |      72.3     |      78.8      |      82.4      |
+| [PIQA][piqa]                   | 0-shot         |      73.8      |      79.6     |      81.8      |      83.3      |
+| [SocialIQA][socialiqa]         | 0-shot         |      48.9      |      51.9     |      53.4      |      54.9      |
+| [TriviaQA][triviaqa]           | 5-shot         |      39.8      |      65.8     |      78.2      |      85.5      |
+| [Natural Questions][naturalq]  | 5-shot         |      9.48      |      20.0     |      31.4      |      36.1      |
+| [ARC-c][arc]                   | 25-shot        |      38.4      |      56.2     |      68.9      |      70.6      |
+| [ARC-e][arc]                   | 0-shot         |      73.0      |      82.4     |      88.3      |      89.0      |
+| [WinoGrande][winogrande]       | 5-shot         |      58.2      |      64.7     |      74.3      |      78.8      |
+| [BIG-Bench Hard][bbh]          | few-shot       |      28.4      |      50.9     |      72.6      |      77.7      |
+| [DROP][drop]                   | 1-shot         |      42.4      |      60.1     |      72.2      |      77.2      |
+
+[hellaswag]: https://arxiv.org/abs/1905.07830
+[boolq]: https://arxiv.org/abs/1905.10044
+[piqa]: https://arxiv.org/abs/1911.11641
+[socialiqa]: https://arxiv.org/abs/1904.09728
+[triviaqa]: https://arxiv.org/abs/1705.03551
+[naturalq]: https://github.com/google-research-datasets/natural-questions
+[arc]: https://arxiv.org/abs/1911.01547
+[winogrande]: https://arxiv.org/abs/1907.10641
+[bbh]: https://paperswithcode.com/dataset/bbh
+[drop]: https://arxiv.org/abs/1903.00161
+
+#### STEM and code
+
+| Benchmark                      | Metric         | Gemma 3 PT 4B | Gemma 3 PT 12B | Gemma 3 PT 27B |
+| ------------------------------ |----------------|:-------------:|:--------------:|:--------------:|
+| [MMLU][mmlu]                   | 5-shot         |      59.6     |      74.5      |      78.6      |
+| [MMLU][mmlu] (Pro COT)         | 5-shot         |      29.2     |      45.3      |      52.2      |
+| [AGIEval][agieval]             | 3-5-shot       |      42.1     |      57.4      |      66.2      |
+| [MATH][math]                   | 4-shot         |      24.2     |      43.3      |      50.0      |
+| [GSM8K][gsm8k]                 | 8-shot         |      38.4     |      71.0      |      82.6      |
+| [GPQA][gpqa]                   | 5-shot         |      15.0     |      25.4      |      24.3      |
+| [MBPP][mbpp]                   | 3-shot         |      46.0     |      60.4      |      65.6      |
+| [HumanEval][humaneval]         | 0-shot         |      36.0     |      45.7      |      48.8      |
+
+[mmlu]: https://arxiv.org/abs/2009.03300
+[agieval]: https://arxiv.org/abs/2304.06364
+[math]: https://arxiv.org/abs/2103.03874
+[gsm8k]: https://arxiv.org/abs/2110.14168
+[gpqa]: https://arxiv.org/abs/2311.12022
+[mbpp]: https://arxiv.org/abs/2108.07732
+[humaneval]: https://arxiv.org/abs/2107.03374
+
+#### Multilingual
+
+| Benchmark                            | Gemma 3 PT 1B | Gemma 3 PT 4B | Gemma 3 PT 12B | Gemma 3 PT 27B |
+| ------------------------------------ |:-------------:|:-------------:|:--------------:|:--------------:|
+| [MGSM][mgsm]                         |      2.04     |      34.7     |      64.3     |      74.3     |
+| [Global-MMLU-Lite][global-mmlu-lite] |      24.9     |      57.0     |      69.4     |      75.7     |
+| [WMT24++][wmt24pp] (ChrF)            |      36.7     |      48.4     |      53.9     |      55.7     |
+| [FloRes][flores]                     |      29.5     |      39.2     |      46.0     |      48.8     |
+| [XQuAD][xquad] (all)                 |      43.9     |      68.0     |      74.5     |      76.8     |
+| [ECLeKTic][eclektic]                 |      4.69     |      11.0     |      17.2     |      24.4     |
+| [IndicGenBench][indicgenbench]       |      41.4     |      57.2     |      61.7     |      63.4     |
+
+[mgsm]: https://arxiv.org/abs/2210.03057
+[flores]: https://arxiv.org/abs/2106.03193
+[xquad]: https://arxiv.org/abs/1910.11856v3
+[global-mmlu-lite]: https://huggingface.co/datasets/CohereForAI/Global-MMLU-Lite
+[wmt24pp]: https://arxiv.org/abs/2502.12404v1
+[eclektic]: https://arxiv.org/abs/2502.21228
+[indicgenbench]: https://arxiv.org/abs/2404.16816
+
+#### Multimodal
+
+| Benchmark                      | Gemma 3 PT 4B | Gemma 3 PT 12B | Gemma 3 PT 27B |
+| ------------------------------ |:-------------:|:--------------:|:--------------:|
+| [COCOcap][coco-cap]            |      102      |      111       |      116       |
+| [DocVQA][docvqa] (val)         |      72.8     |      82.3      |      85.6      |
+| [InfoVQA][info-vqa] (val)      |      44.1     |      54.8      |      59.4      |
+| [MMMU][mmmu] (pt)              |      39.2     |      50.3      |      56.1      |
+| [TextVQA][textvqa] (val)       |      58.9     |      66.5      |      68.6      |
+| [RealWorldQA][realworldqa]     |      45.5     |      52.2      |      53.9      |
+| [ReMI][remi]                   |      27.3     |      38.5      |      44.8      |
+| [AI2D][ai2d]                   |      63.2     |      75.2      |      79.0      |
+| [ChartQA][chartqa]             |      63.6     |      74.7      |      76.3      |
+| [VQAv2][vqav2]                 |      63.9     |      71.2      |      72.9      |
+| [BLINK][blinkvqa]              |      38.0     |      35.9      |      39.6      |
+| [OKVQA][okvqa]                 |      51.0     |      58.7      |      60.2      |
+| [TallyQA][tallyqa]             |      42.5     |      51.8      |      54.3      |
+| [SpatialSense VQA][ss-vqa]     |      50.9     |      60.0      |      59.4      |
+| [CountBenchQA][countbenchqa]   |      26.1     |      17.8      |      68.0      |
+
+[coco-cap]: https://cocodataset.org/#home
+[docvqa]: https://www.docvqa.org/
+[info-vqa]: https://arxiv.org/abs/2104.12756
+[mmmu]: https://arxiv.org/abs/2311.16502
+[textvqa]: https://textvqa.org/
+[realworldqa]: https://paperswithcode.com/dataset/realworldqa
+[remi]: https://arxiv.org/html/2406.09175v1
+[ai2d]: https://allenai.org/data/diagrams
+[chartqa]: https://arxiv.org/abs/2203.10244
+[vqav2]: https://visualqa.org/index.html
+[blinkvqa]: https://arxiv.org/abs/2404.12390
+[okvqa]: https://okvqa.allenai.org/
+[tallyqa]: https://arxiv.org/abs/1810.12440
+[ss-vqa]: https://arxiv.org/abs/1908.02660
+[countbenchqa]: https://github.com/google-research/big_vision/blob/main/big_vision/datasets/countbenchqa/
+
+## Ethics and Safety
+
+Ethics and safety evaluation approach and results.
+
+### Evaluation Approach
+
+Our evaluation methods include structured evaluations and internal red-teaming
+testing of relevant content policies. Red-teaming was conducted by a number of
+different teams, each with different goals and human evaluation metrics. These
+models were evaluated against a number of different categories relevant to
+ethics and safety, including:
+
+-   **Child Safety**: Evaluation of text-to-text and image to text prompts
+    covering child safety policies, including child sexual abuse and
+    exploitation.
+-   **Content Safety:** Evaluation of text-to-text and image to text prompts
+    covering safety policies including, harassment, violence and gore, and hate
+    speech.
+-   **Representational Harms**: Evaluation of text-to-text and image to text
+    prompts covering safety policies including bias, stereotyping, and harmful
+    associations or inaccuracies.
+
+In addition to development level evaluations, we conduct "assurance
+evaluations" which are our 'arms-length' internal evaluations for responsibility
+governance decision making. They are conducted separately from the model
+development team, to inform decision making about release. High level findings
+are fed back to the model team, but prompt sets are held-out to prevent
+overfitting and preserve the results' ability to inform decision making.
+Assurance evaluation results are reported to our Responsibility & Safety Council
+as part of release review.
+
+### Evaluation Results
+
+For all areas of safety testing, we saw major improvements in the categories of
+child safety, content safety, and representational harms relative to previous
+Gemma models. All testing was conducted without safety filters to evaluate the
+model capabilities and behaviors. For both text-to-text and image-to-text, and
+across all model sizes, the model produced minimal policy violations, and showed
+significant improvements over previous Gemma models' performance with respect
+to ungrounded inferences. A limitation of our evaluations was they included only
+English language prompts.
+
+## Usage and Limitations
+
+These models have certain limitations that users should be aware of.
+
+### Intended Usage
+
+Open vision-language models (VLMs) models have a wide range of applications
+across various industries and domains. The following list of potential uses is
+not comprehensive. The purpose of this list is to provide contextual information
+about the possible use-cases that the model creators considered as part of model
+training and development.
+
+-   Content Creation and Communication
+    -   Text Generation: These models can be used to generate creative text
+        formats such as poems, scripts, code, marketing copy, and email drafts.
+    -   Chatbots and Conversational AI: Power conversational interfaces
+        for customer service, virtual assistants, or interactive applications.
+    -   Text Summarization: Generate concise summaries of a text corpus,
+        research papers, or reports.
+    -   Image Data Extraction: These models can be used to extract,
+        interpret, and summarize visual data for text communications.
+-   Research and Education
+    -   Natural Language Processing (NLP) and VLM Research: These
+        models can serve as a foundation for researchers to experiment with VLM
+        and NLP techniques, develop algorithms, and contribute to the
+        advancement of the field.
+    -   Language Learning Tools: Support interactive language learning
+        experiences, aiding in grammar correction or providing writing practice.
+    -   Knowledge Exploration: Assist researchers in exploring large
+        bodies of text by generating summaries or answering questions about
+        specific topics.
+
+### Limitations
+
+-   Training Data
+    -   The quality and diversity of the training data significantly
+        influence the model's capabilities. Biases or gaps in the training data
+        can lead to limitations in the model's responses.
+    -   The scope of the training dataset determines the subject areas
+        the model can handle effectively.
+-   Context and Task Complexity
+    -   Models are better at tasks that can be framed with clear
+        prompts and instructions. Open-ended or highly complex tasks might be
+        challenging.
+    -   A model's performance can be influenced by the amount of context
+        provided (longer context generally leads to better outputs, up to a
+        certain point).
+-   Language Ambiguity and Nuance
+    -   Natural language is inherently complex. Models might struggle
+        to grasp subtle nuances, sarcasm, or figurative language.
+-   Factual Accuracy
+    -   Models generate responses based on information they learned
+        from their training datasets, but they are not knowledge bases. They
+        may generate incorrect or outdated factual statements.
+-   Common Sense
+    -   Models rely on statistical patterns in language. They might
+        lack the ability to apply common sense reasoning in certain situations.
+
+### Ethical Considerations and Risks
+
+The development of vision-language models (VLMs) raises several ethical
+concerns. In creating an open model, we have carefully considered the following:
+
+-   Bias and Fairness
+    -   VLMs trained on large-scale, real-world text and image data can
+        reflect socio-cultural biases embedded in the training material. These
+        models underwent careful scrutiny, input data pre-processing described
+        and posterior evaluations reported in this card.
+-   Misinformation and Misuse
+    -   VLMs can be misused to generate text that is false, misleading,
+        or harmful.
+    -   Guidelines are provided for responsible use with the model, see the
+        [Responsible Generative AI Toolkit][rai-toolkit].
+-   Transparency and Accountability:
+    -   This model card summarizes details on the models' architecture,
+        capabilities, limitations, and evaluation processes.
+    -   A responsibly developed open model offers the opportunity to
+        share innovation by making VLM technology accessible to developers and
+        researchers across the AI ecosystem.
+
+Risks identified and mitigations:
+
+-   **Perpetuation of biases**: It's encouraged to perform continuous
+    monitoring (using evaluation metrics, human review) and the exploration of
+    de-biasing techniques during model training, fine-tuning, and other use
+    cases.
+-   **Generation of harmful content**: Mechanisms and guidelines for content
+    safety are essential. Developers are encouraged to exercise caution and
+    implement appropriate content safety safeguards based on their specific
+    product policies and application use cases.
+-   **Misuse for malicious purposes**: Technical limitations and developer
+    and end-user education can help mitigate against malicious applications of
+    VLMs. Educational resources and reporting mechanisms for users to flag
+    misuse are provided. Prohibited uses of Gemma models are outlined in the
+    [Gemma Prohibited Use Policy][prohibited-use].
+-   **Privacy violations**: Models were trained on data filtered for removal
+    of certain personal information and other sensitive data. Developers are
+    encouraged to adhere to privacy regulations with privacy-preserving
+    techniques.
+
+### Benefits
+
+At the time of release, this family of models provides high-performance open
+vision-language model implementations designed from the ground up for
+responsible AI development compared to similarly sized models.
+
+Using the benchmark evaluation metrics described in this document, these models
+have shown to provide superior performance to other, comparably-sized open model
+alternatives.
+
+[g3-tech-report]: https://goo.gle/Gemma3Report
+[rai-toolkit]: https://ai.google.dev/responsible
+[kaggle-gemma]: https://www.kaggle.com/models/google/gemma-3
+[vertex-mg-gemma3]: https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma3
+[terms]: https://ai.google.dev/gemma/terms
+[safety-policies]: https://ai.google/static/documents/ai-responsibility-update-published-february-2025.pdf
+[prohibited-use]: https://ai.google.dev/gemma/prohibited_use_policy
+[tpu]: https://cloud.google.com/tpu/docs/intro-to-tpu
+[sustainability]: https://sustainability.google/operating-sustainably/
+[jax]: https://github.com/jax-ml/jax
+[ml-pathways]: https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture/
+[sustainability]: https://sustainability.google/operating-sustainably/
+[gemini-2-paper]: https://arxiv.org/abs/2312.11805

chat_template.jinja

@@ -0,0 +1,47 @@
+{{ bos_token }}
+{%- if messages[0]['role'] == 'system' -%}
+    {%- if messages[0]['content'] is string -%}
+        {%- set first_user_prefix = messages[0]['content'] + '
+
+' -%}
+    {%- else -%}
+        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '
+
+' -%}
+    {%- endif -%}
+    {%- set loop_messages = messages[1:] -%}
+{%- else -%}
+    {%- set first_user_prefix = "" -%}
+    {%- set loop_messages = messages -%}
+{%- endif -%}
+{%- for message in loop_messages -%}
+    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
+        {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
+    {%- endif -%}
+    {%- if (message['role'] == 'assistant') -%}
+        {%- set role = "model" -%}
+    {%- else -%}
+        {%- set role = message['role'] -%}
+    {%- endif -%}
+    {{ '<start_of_turn>' + role + '
+' + (first_user_prefix if loop.first else "") }}
+    {%- if message['content'] is string -%}
+        {{ message['content'] | trim }}
+    {%- elif message['content'] is iterable -%}
+        {%- for item in message['content'] -%}
+            {%- if item['type'] == 'image' -%}
+                {{ '<start_of_image>' }}
+            {%- elif item['type'] == 'text' -%}
+                {{ item['text'] | trim }}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- else -%}
+        {{ raise_exception("Invalid content type") }}
+    {%- endif -%}
+    {{ '<end_of_turn>
+' }}
+{%- endfor -%}
+{%- if add_generation_prompt -%}
+    {{'<start_of_turn>model
+'}}
+{%- endif -%}

config.json

@@ -0,0 +1,318 @@
+{
+  "architectures": [
+    "Gemma3ForConditionalGeneration"
+  ],
+  "boi_token_index": 255999,
+  "dtype": "bfloat16",
+  "eoi_token_index": 256000,
+  "eos_token_id": [
+    1,
+    106
+  ],
+  "image_token_index": 262144,
+  "initializer_range": 0.02,
+  "mm_tokens_per_image": 256,
+  "model_type": "gemma3",
+  "quantization_config": {
+    "config_groups": {
+      "group_0": {
+        "format": "nvfp4-pack-quantized",
+        "input_activations": {
+          "actorder": null,
+          "block_structure": null,
+          "dynamic": "local",
+          "group_size": 16,
+          "num_bits": 4,
+          "observer": "static_minmax",
+          "observer_kwargs": {},
+          "strategy": "tensor_group",
+          "symmetric": true,
+          "type": "float"
+        },
+        "output_activations": null,
+        "targets": [
+          "Linear"
+        ],
+        "weights": {
+          "actorder": null,
+          "block_structure": null,
+          "dynamic": false,
+          "group_size": 16,
+          "num_bits": 4,
+          "observer": "static_minmax",
+          "observer_kwargs": {},
+          "strategy": "tensor_group",
+          "symmetric": true,
+          "type": "float"
+        }
+      }
+    },
+    "format": "nvfp4-pack-quantized",
+    "global_compression_ratio": null,
+    "ignore": [
+      "model.vision_tower.vision_model.encoder.layers.0.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.0.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.0.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.0.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.0.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.0.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.1.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.1.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.1.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.1.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.1.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.1.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.2.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.2.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.2.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.2.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.2.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.2.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.3.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.3.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.3.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.3.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.3.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.3.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.4.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.4.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.4.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.4.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.4.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.4.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.5.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.5.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.5.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.5.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.5.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.5.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.6.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.6.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.6.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.6.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.6.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.6.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.7.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.7.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.7.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.7.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.7.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.7.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.8.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.8.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.8.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.8.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.8.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.8.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.9.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.9.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.9.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.9.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.9.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.9.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.10.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.10.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.10.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.10.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.10.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.10.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.11.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.11.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.11.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.11.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.11.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.11.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.12.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.12.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.12.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.12.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.12.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.12.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.13.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.13.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.13.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.13.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.13.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.13.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.14.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.14.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.14.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.14.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.14.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.14.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.15.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.15.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.15.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.15.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.15.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.15.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.16.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.16.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.16.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.16.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.16.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.16.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.17.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.17.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.17.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.17.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.17.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.17.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.18.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.18.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.18.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.18.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.18.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.18.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.19.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.19.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.19.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.19.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.19.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.19.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.20.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.20.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.20.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.20.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.20.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.20.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.21.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.21.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.21.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.21.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.21.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.21.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.22.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.22.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.22.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.22.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.22.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.22.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.23.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.23.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.23.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.23.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.23.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.23.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.24.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.24.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.24.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.24.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.24.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.24.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.25.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.25.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.25.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.25.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.25.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.25.mlp.fc2",
+      "model.vision_tower.vision_model.encoder.layers.26.self_attn.k_proj",
+      "model.vision_tower.vision_model.encoder.layers.26.self_attn.v_proj",
+      "model.vision_tower.vision_model.encoder.layers.26.self_attn.q_proj",
+      "model.vision_tower.vision_model.encoder.layers.26.self_attn.out_proj",
+      "model.vision_tower.vision_model.encoder.layers.26.mlp.fc1",
+      "model.vision_tower.vision_model.encoder.layers.26.mlp.fc2",
+      "lm_head"
+    ],
+    "kv_cache_scheme": null,
+    "quant_method": "compressed-tensors",
+    "quantization_status": "compressed",
+    "sparsity_config": {},
+    "transform_config": {},
+    "version": "0.12.3.a20251028"
+  },
+  "text_config": {
+    "_sliding_window_pattern": 6,
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "attn_logit_softcapping": null,
+    "final_logit_softcapping": null,
+    "head_dim": 256,
+    "hidden_activation": "gelu_pytorch_tanh",
+    "hidden_size": 3840,
+    "initializer_range": 0.02,
+    "intermediate_size": 15360,
+    "layer_types": [
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "full_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "full_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "full_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "full_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "full_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "full_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "full_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "sliding_attention",
+      "full_attention"
+    ],
+    "max_position_embeddings": 131072,
+    "model_type": "gemma3_text",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 48,
+    "num_key_value_heads": 8,
+    "query_pre_attn_scalar": 256,
+    "rms_norm_eps": 1e-06,
+    "rope_local_base_freq": 10000.0,
+    "rope_scaling": {
+      "factor": 8.0,
+      "rope_type": "linear"
+    },
+    "rope_theta": 1000000.0,
+    "sliding_window": 1024,
+    "use_bidirectional_attention": false,
+    "use_cache": true,
+    "vocab_size": 262208
+  },
+  "transformers_version": "4.57.1",
+  "vision_config": {
+    "attention_dropout": 0.0,
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 896,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_channels": 3,
+    "num_hidden_layers": 27,
+    "patch_size": 14,
+    "vision_use_head": false
+  }
+}

generation_config.json

@@ -0,0 +1,13 @@
+{
+  "bos_token_id": 2,
+  "cache_implementation": "hybrid",
+  "do_sample": true,
+  "eos_token_id": [
+    1,
+    106
+  ],
+  "pad_token_id": 0,
+  "top_k": 64,
+  "top_p": 0.95,
+  "transformers_version": "4.57.1"
+}

model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b1a5404aa5f0212ef5154ab07630ca412a3b53e2cfc73dc5d383a83051f196a
+size 4996570464

model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8dd269d16dbc07e45d6013d9dfed5288564d5bbba538c433602c7a3398da247c
+size 3913167880

model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ead2180d70715d2eca76bf634d4defc4d0f41a8690e74c3cc8735177ffcf853d
+size 2013757584

preprocessor_config.json

@@ -0,0 +1,29 @@
+{
+  "do_convert_rgb": null,
+  "do_normalize": true,
+  "do_pan_and_scan": null,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "Gemma3ImageProcessor",
+  "image_seq_length": 256,
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "pan_and_scan_max_num_crops": null,
+  "pan_and_scan_min_crop_size": null,
+  "pan_and_scan_min_ratio_to_activate": null,
+  "processor_class": "Gemma3Processor",
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 896,
+    "width": 896
+  }
+}

processor_config.json

@@ -0,0 +1,4 @@
+{
+  "image_seq_length": 256,
+  "processor_class": "Gemma3Processor"
+}

recipe.yaml

@@ -0,0 +1,6 @@
+default_stage:
+  default_modifiers:
+    QuantizationModifier:
+      targets: [Linear]
+      ignore: [lm_head, 're:model\.vision_tower.*', 're:model\.multi_modal_projector.*']
+      scheme: NVFP4

special_tokens_map.json

@@ -0,0 +1,33 @@
+{
+  "boi_token": "<start_of_image>",
+  "bos_token": {
+    "content": "<bos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eoi_token": "<end_of_image>",
+  "eos_token": {
+    "content": "<eos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "image_token": "<image_soft_token>",
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4667f2089529e8e7657cfb6d1c19910ae71ff5f28aa7ab2ff2763330affad795
+size 33384568