Upload folder using huggingface_hub

README.md

@@ -34,7 +34,7 @@ tags:
     <a href="https://discord.gg/unsloth">
       <img src="https://github.com/unslothai/unsloth/raw/main/images/Discord%20button.png" width="173">
     </a>
-    <a href="https://docs.unsloth.ai/basics/qwen3-how-to-run-and-fine-tune">
+    <a href="https://docs.unsloth.ai/">
       <img src="https://raw.githubusercontent.com/unslothai/unsloth/refs/heads/main/images/documentation%20green%20button.png" width="143">
     </a>
   </div>
@@ -53,7 +53,7 @@ tags:
     issues): [MedGemma](https://github.com/google-health/medgemma)
 *   Quick start notebook: [GitHub](https://github.com/google-health/medgemma/blob/main/notebooks/quick_start_with_hugging_face.ipynb)
 *   Fine-tuning notebook: [GitHub](https://github.com/google-health/medgemma/blob/main/notebooks/fine_tune_with_hugging_face.ipynb)
-*   [Patient Education Demo built using MedGemma](https://huggingface.co/spaces/google/rad_explain)
+*   Concept applications built using MedGemma: [Collection](https://huggingface.co/collections/google/medgemma-concept-apps-686ea036adb6d51416b0928a)
 *   Support: See [Contact](https://developers.google.com/health-ai-developer-foundations/medgemma/get-started.md#contact)
 *   License: The use of MedGemma is governed by the [Health AI Developer
     Foundations terms of
@@ -70,32 +70,50 @@ This section describes the MedGemma model and how to use it.
 MedGemma is a collection of [Gemma 3](https://ai.google.dev/gemma/docs/core)
 variants that are trained for performance on medical text and image
 comprehension. Developers can use MedGemma to accelerate building
-healthcare-based AI applications. MedGemma currently comes in two variants: a 4B
-multimodal version and a 27B text-only version.
-
-MedGemma 4B utilizes a [SigLIP](https://arxiv.org/abs/2303.15343) image encoder
-that has been specifically pre-trained on a variety of de-identified medical
-data, including chest X-rays, dermatology images, ophthalmology images, and
-histopathology slides. Its LLM component is trained on a diverse set of medical
-data, including radiology images, histopathology patches, ophthalmology images,
-and dermatology images.
+healthcare-based AI applications. MedGemma currently comes in three variants: a
+4B multimodal version and 27B text-only and multimodal versions.
+
+Both MedGemma multimodal versions utilize a
+[SigLIP](https://arxiv.org/abs/2303.15343) image encoder that has been
+specifically pre-trained on a variety of de-identified medical data, including
+chest X-rays, dermatology images, ophthalmology images, and histopathology
+slides. Their LLM components are trained on a diverse set of medical data,
+including medical text, medical question-answer pairs, FHIR-based electronic
+health record data (27B multimodal only), radiology images, histopathology
+patches, ophthalmology images, and dermatology images.
 
 MedGemma 4B is available in both pre-trained (suffix: `-pt`) and
 instruction-tuned (suffix `-it`) versions. The instruction-tuned version is a
-better starting point for most applications. The pre-trained version notably
-achieves better performance on MIMIC-style chest X-ray reporting.
-
-MedGemma 27B has been trained exclusively on medical text and optimized for
-inference-time computation. MedGemma 27B is only available as an
-instruction-tuned model.
+better starting point for most applications. The pre-trained version is
+available for those who want to experiment more deeply with the models.
+
+MedGemma 27B multimodal has pre-training on medical image, medical record and
+medical record comprehension tasks. MedGemma 27B text-only has been trained
+exclusively on medical text. Both models have been optimized for inference-time
+computation on medical reasoning. This means it has slightly higher performance
+on some text benchmarks than MedGemma 27B multimodal. Users who want to work
+with a single model for both medical text, medical record and medical image
+tasks are better suited for MedGemma 27B multimodal. Those that only need text
+use-cases may be better served with the text-only variant. Both MedGemma 27B
+variants are only available in instruction-tuned versions.
 
 MedGemma variants have been evaluated on a range of clinically relevant
-benchmarks to illustrate their baseline performance. These include both open
-benchmark datasets and curated datasets. Developers can fine-tune MedGemma
-variants for improved performance. Consult the Intended Use section below for
-more details.
-
-A full technical report will be available soon.
+benchmarks to illustrate their baseline performance. These evaluations are based
+on both open benchmark datasets and curated datasets. Developers can fine-tune
+MedGemma variants for improved performance. Consult the [Intended
+Use](https://developers.google.com/health-ai-developer-foundations/medgemma/model-card#intended_use)
+section below for more details.
+
+MedGemma is optimized for medical applications that involve a text generation
+component. For medical image-based applications that do not involve text
+generation, such as data-efficient classification, zero-shot classification, or
+content-based or semantic image retrieval, the [MedSigLIP image
+encoder](https://developers.google.com/health-ai-developer-foundations/medsiglip/model-card)
+is recommended. MedSigLIP is based on the same image encoder that powers
+MedGemma.
+
+Please consult the [MedGemma Technical Report](https://arxiv.org/abs/2507.05201)
+for more details.
 
 ### How to use
 
@@ -104,14 +122,6 @@ model locally on GPU. If you want to use the model at scale, we recommend that
 you create a production version using [Model
 Garden](https://cloud.google.com/model-garden).
 
-
-First, install the Transformers library. Gemma 3 is supported starting from
-transformers 4.50.0.
-
-```sh
-$ pip install -U transformers
-```
-
 First, install the Transformers library. Gemma 3 is supported starting from
 transformers 4.50.0.
 
@@ -146,8 +156,8 @@ messages = [
     {
         "role": "user",
         "content": [
-            {"type": "text", "text": "Describe this X-ray"}
-            {"type": "image", "image": image},
+            {"type": "text", "text": "Describe this X-ray"},
+            {"type": "image", "image": image}
         ]
     }
 ]
@@ -213,43 +223,47 @@ See the following Colab notebooks for examples of how to use MedGemma:
 
 *   To give the model a quick try, running it locally with weights from Hugging
     Face, see [Quick start notebook in
-    Colab](https://colab.research.google.com/github/google-health/medgemma/blob/main/notebooks/quick_start_with_hugging_face.ipynb). Note that you will need to use Colab
-    Enterprise to run the 27B model without quantization.
+    Colab](https://colab.research.google.com/github/google-health/medgemma/blob/main/notebooks/quick_start_with_hugging_face.ipynb).
+    Note that you will need to use Colab Enterprise to obtain adequate GPU
+    resources to run either 27B model without quantization.
 
-*   For an example of fine-tuning the model, see the [Fine-tuning notebook in
+*   For an example of fine-tuning the 4B model, see the [Fine-tuning notebook in
     Colab](https://colab.research.google.com/github/google-health/medgemma/blob/main/notebooks/fine_tune_with_hugging_face.ipynb).
+    The 27B models can be fine tuned in a similar manner but will require more
+    time and compute resources than the 4B model.
 
 ### Model architecture overview
 
 The MedGemma model is built based on [Gemma 3](https://ai.google.dev/gemma/) and
-uses the same decoder-only transformer architecture as Gemma 3. To read more
+uses the same decoder-only transformer architecture as Gemma 3\. To read more
 about the architecture, consult the Gemma 3 [model
 card](https://ai.google.dev/gemma/docs/core/model_card_3).
 
 ### Technical specifications
 
 *   **Model type**: Decoder-only Transformer architecture, see the [Gemma 3
-    technical
-    report](https://storage.googleapis.com/deepmind-media/gemma/Gemma3Report.pdf)
-*   **Modalities**: **4B**: Text, vision; **27B**: Text only
-*   **Attention mechanism**: Utilizes grouped-query attention (GQA)
+    Technical
+    Report](https://storage.googleapis.com/deepmind-media/gemma/Gemma3Report.pdf)
+*   **Input Modalities**: Text, vision
+*   **Output Modality:** Text only
+*   **Attention mechanism**: Grouped-query attention (GQA)
 *   **Context length**: Supports long context, at least 128K tokens
-*   **Key publication**: Coming soon
-*   **Model created**: May 20, 2025
-*   **Model version**: 1.0.0
+*   **Key publication**: https://arxiv.org/abs/2507.05201
+*   **Model created**: July 9, 2025
+
+*   **Model version**: 1.0.1
 
 ### Citation
 
-A technical report is coming soon. In the meantime, if you publish using this
-model, please cite the Hugging Face model page:
+When using this model, please cite: Sellergren et al. "MedGemma Technical
+Report." *arXiv preprint arXiv:2507.05201* (2025).
 
 ```none
-@misc{medgemma-hf,
-    author = {Google},
-    title = {MedGemma Hugging Face}
-    howpublished = {\url{https://huggingface.co/collections/google/medgemma-release-680aade845f90bec6a3f60c4}},
-    year = {2025},
-    note = {Accessed: [Insert Date Accessed, e.g., 2025-05-20]}
+@article{sellergren2025medgemma,
+  title={MedGemma Technical Report},
+  author={Sellergren, Andrew and Kazemzadeh, Sahar and Jaroensri, Tiam and Kiraly, Atilla and Traverse, Madeleine and Kohlberger, Timo and Xu, Shawn and Jamil, Fayaz and Hughes, Cían and Lau, Charles and others},
+  journal={arXiv preprint arXiv:2507.05201},
+  year={2025}
 }
 ```
 
@@ -276,33 +290,42 @@ report generation, visual question answering, and text-based tasks.
 
 #### Imaging evaluations
 
-The multimodal performance of MedGemma 4B was evaluated across a range of
-benchmarks, focusing on radiology, dermatology, histopathology, ophthalmology,
-and multimodal clinical reasoning.
+The multimodal performance of MedGemma 4B and 27B multimodal was evaluated
+across a range of benchmarks, focusing on radiology, dermatology,
+histopathology, ophthalmology, and multimodal clinical reasoning.
 
 MedGemma 4B outperforms the base Gemma 3 4B model across all tested multimodal
 health benchmarks.
 
-| Task and metric | MedGemma 4B | Gemma 3 4B |
+| Task and metric | Gemma 3 4B | MedGemma 4B |
 | :---- | :---- | :---- |
 | **Medical image classification** |  |  |
-| MIMIC CXR \- Average F1 for top 5 conditions | 88.9 | 81.1 |
-| CheXpert CXR \- Average F1 for top 5 conditions | 48.1 | 31.2 |
-| DermMCQA\* \- Accuracy | 71.8 | 42.6 |
+| MIMIC CXR\*\* \- macro F1 for top 5 conditions | 81.2 | 88.9 |
+| CheXpert CXR \- macro F1 for top 5 conditions | 32.6 | 48.1 |
+| CXR14 \- macro F1 for 3 conditions | 32.0 | 50.1 |
+| PathMCQA\* (histopathology, internal\*\*)  \- Accuracy | 37.1 | 69.8 |
+| US-DermMCQA\* \- Accuracy | 52.5 | 71.8 |
+| EyePACS\* (fundus, internal) \- Accuracy | 14.4 | 64.9 |
 | **Visual question answering** |  |  |
-| SlakeVQA (radiology) \- Tokenized F1 | 62.3 | 38.6 |
-| VQA-Rad\*\* (radiology) \- Tokenized F1 | 49.9 | 38.6 |
-| PathMCQA (histopathology, internal\*\*\*)  \- Accuracy | 69.8 | 37.1 |
-| **Knowledge and reasoning** |  |  |
-| MedXpertQA (text \+ multimodal questions) \- Accuracy | 18.8 | 16.4 |
-
-*Based on [ref](https://www.nature.com/articles/s41591-020-0842-3), presented as
-a 4-way MCQ per example for skin condition classification.
-
-**On balanced split, see [ref](https://arxiv.org/pdf/2405.03162).
-
-***Based on multiple datasets, presented as 3-9 way MCQ per example for
-identification, grading, and subtype for breast, cervical, and prostate cancer.
+| SLAKE (radiology) \- Tokenized F1 | 40.2 | 72.3 |
+| VQA-RAD\*\*\* (radiology) \- Tokenized F1 | 33.6 | 49.9 |
+| **Knowledge and reasoning** |  |  |  |  |
+| MedXpertQA (text \+ multimodal questions) \- Accuracy | 16.4 | 18.8 |
+
+*Internal datasets. US-DermMCQA is described in [Liu (2020, Nature
+medicine)](https://www.nature.com/articles/s41591-020-0842-3), presented as a
+4-way MCQ per example for skin condition classification. PathMCQA is based on
+multiple datasets, presented as 3-9 way MCQ per example for identification,
+grading, and subtype for breast, cervical, and prostate cancer. EyePACS is a
+dataset of fundus images with classification labels based on 5-level diabetic
+retinopathy severity (None, Mild, Moderate, Severe, Proliferative). More details
+in the [MedGemma Technical Report](https://arxiv.org/abs/2507.05201).
+
+**Based on radiologist adjudicated labels, described in [Yang (2024,
+arXiv)](https://arxiv.org/pdf/2405.03162) Section A.1.1.
+
+***Based on "balanced split," described in [Yang (2024,
+arXiv)](https://arxiv.org/pdf/2405.03162).
 
 #### Chest X-ray report generation
 
@@ -312,15 +335,17 @@ F1 metric](https://arxiv.org/abs/2106.14463). We compare the MedGemma
 pre-trained checkpoint with our previous best model for CXR report generation,
 [PaliGemma 2](https://arxiv.org/abs/2412.03555).
 
-| Metric | MedGemma 4B (pre-trained) | PaliGemma 2 3B (tuned for CXR) | PaliGemma 2 10B (tuned for CXR) |
-| :---- | :---- | :---- | :---- |
-| **Chest X-ray report generation** |  |  |  |
-| MIMIC CXR \- RadGraph F1 | 29.5 | 28.8 | 29.5 |
+| Metric | MedGemma 4B (pre-trained) | MedGemma 4B (tuned for CXR)| PaliGemma 2 3B (tuned for CXR) | PaliGemma 2 10B (tuned for CXR) |
+| :---- | :---- | :---- | :---- | :---- |
+| MIMIC CXR \- RadGraph F1 | 29.5 | 30.3 |28.8 | 29.5 |
+
+
 
-The instruction-tuned versions of MedGemma 4B and Gemma 3 4B achieve lower
-scores (0.22 and 0.12, respectively) due to the differences in reporting style
+The instruction-tuned versions of MedGemma 4B and MedGemma 27B achieve lower
+scores (21.9 and 21.3, respectively) due to the differences in reporting style
 compared to the MIMIC ground truth reports. Further fine-tuning on MIMIC reports
-will enable users to achieve improved performance.
+enables users to achieve improved performance, as shown by the improved
+performance of the MedGemma 4B model that was tuned for CXR.
 
 #### Text evaluations
 
@@ -330,18 +355,30 @@ text-only benchmarks for medical knowledge and reasoning.
 The MedGemma models outperform their respective base Gemma models across all
 tested text-only health benchmarks.
 
-| Metric | MedGemma 27B | Gemma 3 27B | MedGemma 4B | Gemma 3 4B |
-| :---- | :---- | :---- | :---- | :---- |
-| MedQA (4-op) | 89.8 (best-of-5) 87.7 (0-shot) | 74.9 | 64.4 | 50.7 |
-| MedMCQA | 74.2 | 62.6 | 55.7 | 45.4 |
-| PubMedQA | 76.8 | 73.4 | 73.4 | 68.4 |
-| MMLU Med (text only) | 87.0 | 83.3 | 70.0 | 67.2 |
-| MedXpertQA (text only) | 26.7 | 15.7 | 14.2 | 11.6 |
-| AfriMed-QA | 84.0 | 72.0 | 52.0 | 48.0 |
+| Metric | Gemma 3 4B | MedGemma 4B |
+| :---- | :---- | :---- |
+| MedQA (4-op) | 50.7 | 64.4 |
+| MedMCQA | 45.4 | 55.7 |
+| PubMedQA | 68.4 | 73.4 |
+| MMLU Med | 67.2 | 70.0 |
+| MedXpertQA (text only) | 11.6 | 14.2 |
+| AfriMed-QA (25 question test set) | 48.0 | 52.0 |
 
 For all MedGemma 27B results, [test-time
 scaling](https://arxiv.org/abs/2501.19393) is used to improve performance.
 
+#### Medical record evaluations
+
+All models were evaluated on a question answer dataset from synthetic FHIR data
+to answer questions about patient records. MedGemma 27B multimodal's
+FHIR-specific training gives it significant improvement over other MedGemma and
+Gemma models.
+
+| Metric | Gemma 3 4B | MedGemma 4B |
+| :---- | :---- | :---- |
+| EHRQA | 70.9 | 67.6 |
+
+
 ### Ethics and safety evaluation
 
 #### Evaluation approach
@@ -405,23 +442,88 @@ medical image modalities. These include both open benchmark datasets and curated
 datasets, with a focus on expert human evaluations for tasks like CXR report
 generation and radiology VQA.
 
+### Ethics and safety evaluation
+
+#### 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.
+*   **General medical harms:** Evaluation of text-to-text and image-to-text
+    prompts covering safety policies, including information quality 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. Notable 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 safe levels of performance across the
+categories of child safety, content safety, and representational harms. All
+testing was conducted without safety filters to evaluate the model capabilities
+and behaviors. For text-to-text, image-to-text, and audio-to-text, and across
+both MedGemma model sizes, the model produced minimal policy violations. A
+limitation of our evaluations was that they included primarily English language
+prompts.
+
+## Data card
+
+### Dataset overview
+
+#### Training
+
+The base Gemma models are pre-trained on a large corpus of text and code data.
+MedGemma multimodal variants utilize a
+[SigLIP](https://arxiv.org/abs/2303.15343) image encoder that has been
+specifically pre-trained on a variety of de-identified medical data, including
+radiology images, histopathology images, ophthalmology images, and dermatology
+images. Their LLM component is trained on a diverse set of medical data,
+including medical text, medical question-answer pairs, FHIR-based electronic
+health record data (27B multimodal only), radiology images, histopathology
+patches, ophthalmology images, and dermatology images.
+
+#### Evaluation
+
+MedGemma models have been evaluated on a comprehensive set of clinically
+relevant benchmarks, including over 22 datasets across 6 different tasks and 4
+medical image modalities. These benchmarks include both open and internal
+datasets.
+
 #### Source
 
 MedGemma utilizes a combination of public and private datasets.
 
 This model was trained on diverse public datasets including MIMIC-CXR (chest
-X-rays and reports), Slake-VQA (multimodal medical images and questions),
-PAD-UFES-20 (skin lesion images and data), SCIN (dermatology images), TCGA
-(cancer genomics data), CAMELYON (lymph node histopathology images), PMC-OA
-(biomedical literature with images), and Mendeley Digital Knee X-Ray (knee
-X-rays).
+X-rays and reports), ChestImaGenome: Set of bounding boxes linking image
+findings with anatomical regions for MIMIC-CXR (MedGemma 27B multimodal only),
+SLAKE (multimodal medical images and questions), PAD-UFES-20 (skin lesion images
+and data), SCIN (dermatology images), TCGA (cancer genomics data), CAMELYON
+(lymph node histopathology images), PMC-OA (biomedical literature with images),
+and Mendeley Digital Knee X-Ray (knee X-rays).
 
 Additionally, multiple diverse proprietary datasets were licensed and
 incorporated (described next).
 
 ### Data Ownership and Documentation
 
-*   [Mimic-CXR](https://physionet.org/content/mimic-cxr/2.1.0/): MIT Laboratory
+*   [MIMIC-CXR](https://physionet.org/content/mimic-cxr/2.1.0/): MIT Laboratory
     for Computational Physiology and Beth Israel Deaconess Medical Center
     (BIDMC).
 *   [Slake-VQA](https://www.med-vqa.com/slake/): The Hong Kong Polytechnic
@@ -457,81 +559,125 @@ incorporated (described next).
     created by a research team led by Jason J. Lau, Soumya Gayen, Asma Ben
     Abacha, and Dina Demner-Fushman and their affiliated institutions (the US
     National Library of Medicine and National Institutes of Health)
+*   [Chest ImaGenome](https://physionet.org/content/chest-imagenome/1.0.0/): IBM
+    Research.
 *   [MedExpQA](https://www.sciencedirect.com/science/article/pii/S0933365724001805):
     This dataset was created by researchers at the HiTZ Center (Basque Center
     for Language Technology and Artificial Intelligence).
 *   [MedXpertQA](https://huggingface.co/datasets/TsinghuaC3I/MedXpertQA): This
     dataset was developed by researchers at Tsinghua University (Beijing, China)
     and Shanghai Artificial Intelligence Laboratory (Shanghai, China).
+*   [HealthSearchQA](https://huggingface.co/datasets/katielink/healthsearchqa):
+    This dataset consists of consisting of 3,173 commonly searched consumer
+    questions
 
 In addition to the public datasets listed above, MedGemma was also trained on
-de-identified datasets licensed for research or collected internally at Google
-from consented participants.
-
-*   Radiology dataset 1: De-identified dataset of different CT studies across
-    body parts from a US-based radiology outpatient diagnostic center network.
-*   Ophthalmology dataset 1: De-identified dataset of fundus images from
-    diabetic retinopathy screening.
-*   Dermatology dataset 1: De-identified dataset of teledermatology skin
+de-identified, licensed datasets or datasets collected internally at Google from
+consented participants.
+
+*   **Radiology dataset 1:** De-identified dataset of different CT studies
+    across body parts from a US-based radiology outpatient diagnostic center
+    network.
+*   **Ophthalmology dataset 1 (EyePACS):** De-identified dataset of fundus
+    images from diabetic retinopathy screening.
+*   **Dermatology dataset 1:** De-identified dataset of teledermatology skin
     condition images (both clinical and dermatoscopic) from Colombia.
-*   Dermatology dataset 2: De-identified dataset of skin cancer images (both
+*   **Dermatology dataset 2:** De-identified dataset of skin cancer images (both
     clinical and dermatoscopic) from Australia.
-*   Dermatology dataset 3: De-identified dataset of non-diseased skin images
+*   **Dermatology dataset 3:** De-identified dataset of non-diseased skin images
     from an internal data collection effort.
-*   Pathology dataset 1: De-identified dataset of histopathology H&E whole slide
-    images created in collaboration with an academic research hospital and
+*   **Pathology dataset 1:** De-identified dataset of histopathology H\&E whole
+    slide images created in collaboration with an academic research hospital and
     biobank in Europe. Comprises de-identified colon, prostate, and lymph nodes.
-*   Pathology dataset 2: De-identified dataset of lung histopathology H&E and
-    IHC whole slide images created by a commercial biobank in the United States.
-*   Pathology dataset 3: De-identified dataset of prostate and lymph node H&E
-    and IHC histopathology whole slide images created by a contract research
-    organization in the United States.
-*   Pathology dataset 4: De-identified dataset of histopathology, predominantly
-    H\&E whole slide images created in collaboration with a large, tertiary
-    teaching hospital in the United States. Comprises a diverse set of tissue
-    and stain types, predominantly H&E.
+*   **Pathology dataset 2:** De-identified dataset of lung histopathology H\&E
+    and IHC whole slide images created by a commercial biobank in the United
+    States.
+*   **Pathology dataset 3:** De-identified dataset of prostate and lymph node
+    H\&E and IHC histopathology whole slide images created by a contract
+    research organization in the United States.
+*   **Pathology dataset 4:** De-identified dataset of histopathology whole slide
+    images created in collaboration with a large, tertiary teaching hospital in
+    the United States. Comprises a diverse set of tissue and stain types,
+    predominantly H\&E.
+*   **EHR dataset 1:** Question/answer dataset drawn from synthetic FHIR records
+    created by [Synthea.](https://synthetichealth.github.io/synthea/) The test
+    set includes 19 unique patients with 200 questions per patient divided into
+    10 different categories.
 
 ### Data citation
 
-*   MIMIC-CXR Johnson, A., Pollard, T., Mark, R., Berkowitz, S., & Horng, S.
-    (2024). MIMIC-CXR Database (version 2.1.0). PhysioNet.
-*   Johnson, A.E.W., Pollard, T.J., Berkowitz, S.J. et al. [MIMIC-CXR, a
-    de-identified publicly available database of chest radiographs with
-    free-text reports. Sci Data 6, 317
-    (2019).](https://doi.org/10.1038/s41597-019-0322-0)
-*   Available on Physionet Goldberger, A., Amaral, L., Glass, L., Hausdorff, J.,
-    Ivanov, P. C., Mark, R., ... & Stanley, H. E. (2000). [PhysioBank,
-    PhysioToolkit, and PhysioNet: Components of a new research resource for
-    complex physiologic signals. Circulation \[Online\]. 101 (23), pp.
-    E215–e220.](https://pubmed.ncbi.nlm.nih.gov/10851218/)
-*   Bo Liu, Li-Ming Zhan, etc. [SLAKE: A Semantically-Labeled Knowledge-Enhanced
-    Dataset for Medical Visual Question
-    Answering](https://arxiv.org/abs/2102.09542).
-*   [PAD-UFES-20: A skin lesion dataset composed of patient data and clinical
-    images collected from
-    smartphones](https://pmc.ncbi.nlm.nih.gov/articles/PMC7479321/)
-*   [The Cancer Genome Atlas Program (TCGA)](https://www.cancer.gov/ccg/research/genome-sequencing/tcga)
-*   Babak Ehteshami Bejnordi, etc.: [Diagnostic Assessment of Deep Learning
+*   **MIMIC-CXR:** Johnson, A., Pollard, T., Mark, R., Berkowitz, S., & Horng,
+    S. (2024). MIMIC-CXR Database (version 2.1.0). PhysioNet.
+    [https://physionet.org/content/mimic-cxr/2.1.0/](https://physionet.org/content/mimic-cxr/2.1.0/)
+    *and* Johnson, Alistair E. W., Tom J. Pollard, Seth J. Berkowitz, Nathaniel
+    R. Greenbaum, Matthew P. Lungren, Chih-Ying Deng, Roger G. Mark, and Steven
+    Horng. 2019\. "MIMIC-CXR, a de-Identified Publicly Available Database of
+    Chest Radiographs with Free-Text Reports." *Scientific Data 6* (1): 1–8.
+
+*   **SLAKE:** Liu, Bo, Li-Ming Zhan, Li Xu, Lin Ma, Yan Yang, and Xiao-Ming Wu.
+    2021.SLAKE: A Semantically-Labeled Knowledge-Enhanced Dataset for Medical
+    Visual Question Answering."
+    [http://arxiv.org/abs/2102.09542](http://arxiv.org/abs/2102.09542).
+
+*   **PAD-UEFS-20:** Pacheco, Andre GC, et al. "PAD-UFES-20: A skin lesion
+    dataset composed of patient data and clinical images collected from
+    smartphones." *Data in brief* 32 (2020): 106221\.
+
+*   **SCIN:** Ward, Abbi, Jimmy Li, Julie Wang, Sriram Lakshminarasimhan, Ashley
+    Carrick, Bilson Campana, Jay Hartford, et al. 2024\. "Creating an Empirical
+    Dermatology Dataset Through Crowdsourcing With Web Search Advertisements."
+    *JAMA Network Open 7* (11): e2446615–e2446615.
+
+*   **TCGA:** The results shown here are in whole or part based upon data
+    generated by the TCGA Research Network:
+    [https://www.cancer.gov/tcga](https://www.cancer.gov/tcga).
+
+*   **CAMELYON16:** Ehteshami Bejnordi, Babak, Mitko Veta, Paul Johannes van
+    Diest, Bram van Ginneken, Nico Karssemeijer, Geert Litjens, Jeroen A. W. M.
+    van der Laak, et al. 2017\. "Diagnostic Assessment of Deep Learning
     Algorithms for Detection of Lymph Node Metastases in Women With Breast
-    Cancer](https://jamanetwork.com/journals/jama/fullarticle/2665774)
-*   MedQA: [https://arxiv.org/abs/2009.13081](https://arxiv.org/abs/2009.13081)
-*   Mendeley Digital Knee X-Ray: Gornale, Shivanand; Patravali, Pooja (2020),
-    "Digital Knee X-ray Images", Mendeley Data, V1, doi: 10.17632/t9ndx37v5h.1
-*   AfriMed-QA: [https://arxiv.org/abs/2411.15640](https://arxiv.org/abs/2411.15640)
-*   VQA-RAD: [Lau, J., Gayen, S., Ben Abacha, A. et al. A dataset of clinically
-    generated visual questions and answers about radiology images. Sci Data 5,
-    180251 (2018).
-    https://doi.org/10.1038/sdata.2018.251](https://doi.org/10.1038/sdata.2018.251)
-*   [MedExpQA: Multilingual benchmarking of Large Language Models for
-    Medical Question
-    Answering](https://www.sciencedirect.com/science/article/pii/S0933365724001805)
-*   MedXpertQA: [arXiv:2501.18362v2](https://arxiv.org/abs/2501.18362)
+    Cancer." *JAMA 318* (22): 2199–2210.
+
+*   **Mendeley Digital Knee X-Ray:** Gornale, Shivanand; Patravali, Pooja
+    (2020), "Digital Knee X-ray Images", Mendeley Data, V1, doi:
+    10.17632/t9ndx37v5h.1
+
+*   **VQA-RAD:** Lau, Jason J., Soumya Gayen, Asma Ben Abacha, and Dina
+    Demner-Fushman. 2018\. "A Dataset of Clinically Generated Visual Questions
+    and Answers about Radiology Images." *Scientific Data 5* (1): 1–10.
+
+*   **Chest ImaGenome:** Wu, J., Agu, N., Lourentzou, I., Sharma, A., Paguio,
+    J., Yao, J. S., Dee, E. C., Mitchell, W., Kashyap, S., Giovannini, A., Celi,
+    L. A., Syeda-Mahmood, T., & Moradi, M. (2021). Chest ImaGenome Dataset
+    (version 1.0.0). PhysioNet. RRID:SCR\_007345.
+    [https://doi.org/10.13026/wv01-y230](https://doi.org/10.13026/wv01-y230)
+
+*   **MedQA:** Jin, Di, Eileen Pan, Nassim Oufattole, Wei-Hung Weng, Hanyi Fang,
+    and Peter Szolovits. 2020\. "What Disease Does This Patient Have? A
+    Large-Scale Open Domain Question Answering Dataset from Medical Exams."
+    [http://arxiv.org/abs/2009.13081](http://arxiv.org/abs/2009.13081).
+
+*   **AfrimedQA:** Olatunji, Tobi, Charles Nimo, Abraham Owodunni, Tassallah
+    Abdullahi, Emmanuel Ayodele, Mardhiyah Sanni, Chinemelu Aka, et al. 2024\.
+    "AfriMed-QA: A Pan-African, Multi-Specialty, Medical Question-Answering
+    Benchmark Dataset."
+    [http://arxiv.org/abs/2411.15640](http://arxiv.org/abs/2411.15640).
+
+*   **MedExpQA:** Alonso, I., Oronoz, M., & Agerri, R. (2024). MedExpQA:
+    Multilingual Benchmarking of Large Language Models for Medical Question
+    Answering. *arXiv preprint arXiv:2404.05590*. Retrieved from
+    [https://arxiv.org/abs/2404.05590](https://arxiv.org/abs/2404.05590)
+
+*   **MedXpertQA:** Zuo, Yuxin, Shang Qu, Yifei Li, Zhangren Chen, Xuekai Zhu,
+    Ermo Hua, Kaiyan Zhang, Ning Ding, and Bowen Zhou. 2025\. "MedXpertQA:
+    Benchmarking Expert-Level Medical Reasoning and Understanding."
+    [http://arxiv.org/abs/2501.18362](http://arxiv.org/abs/2501.18362).
 
 ### De-identification/anonymization:
 
-Google and partnerships utilize datasets that have been rigorously anonymized or
+Google and its partners utilize datasets that have been rigorously anonymized or
 de-identified to ensure the protection of individual research participants and
-patient privacy
+patient privacy.
 
 ## Implementation information
 
@@ -562,8 +708,8 @@ textual), however the model was pre-trained using chest X-ray, pathology,
 dermatology, and fundus images. Examples of tasks within MedGemma's training
 include visual question answering pertaining to medical images, such as
 radiographs, or providing answers to textual medical questions. Full details of
-all the tasks MedGemma has been evaluated can be found in an upcoming technical
-report.
+all the tasks MedGemma has been evaluated can be found in the [MedGemma
+Technical Report](https://arxiv.org/abs/2507.05201).
 
 ### Benefits
 
@@ -596,7 +742,7 @@ of multiple images.
 MedGemma has not been evaluated or optimized for multi-turn applications.
 
 MedGemma's training may make it more sensitive to the specific prompt used than
-Gemma 3.
+Gemma 3\.
 
 When adapting MedGemma developer should consider the following:
 
@@ -612,4 +758,14 @@ When adapting MedGemma developer should consider the following:
     overestimating its true ability to generalize to novel medical concepts.
     Developers should validate MedGemma on datasets not publicly available or
     otherwise made available to non-institutional researchers to mitigate this
-    risk.
+    risk.
+
+
+### Release notes
+
+*   May 20, 2025: Initial Release
+*   July 9, 2025 Bug Fix: Fixed the subtle degradation in the multimodal
+    performance. The issue was due to a missing end-of-image token in the model
+    vocabulary, impacting combined text-and-image tasks. This fix reinstates and
+    correctly maps that token, ensuring text-only tasks remain unaffected while
+    restoring multimodal performance.

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

@@ -15,13 +15,48 @@
     "attention_bias": false,
     "attention_dropout": 0.0,
     "attn_logit_softcapping": null,
-    "cache_implementation": "hybrid",
     "final_logit_softcapping": null,
     "head_dim": 256,
     "hidden_activation": "gelu_pytorch_tanh",
     "hidden_size": 2560,
     "initializer_range": 0.02,
     "intermediate_size": 10240,
+    "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"
+    ],
     "max_position_embeddings": 131072,
     "model_type": "gemma3_text",
     "num_attention_heads": 8,
@@ -36,13 +71,12 @@
     },
     "rope_theta": 1000000,
     "sliding_window": 1024,
-    "sliding_window_pattern": 6,
     "torch_dtype": "bfloat16",
     "use_cache": true,
     "vocab_size": 262208
   },
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.51.3",
+  "transformers_version": "4.53.1",
   "unsloth_fixed": true,
   "vision_config": {
     "attention_dropout": 0.0,

generation_config.json

@@ -1,13 +1,10 @@
 {
+  "_from_model_config": true,
   "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": "4.51.3"
+  "transformers_version": "4.53.1"
 }

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:359b85680f8fe50feedb13cb3d7044f6cd690a967b4c07cb0c24201b0ae7fd87
+oid sha256:acd1c5fbeda43edbd93f164edc23844da1cb6136d0af5120c7ece561be8fbd01
 size 4961251752

model.safetensors.index.json

@@ -1,5 +1,6 @@
 {
   "metadata": {
+    "total_parameters": 4971331952,
     "total_size": 8600158944
   },
   "weight_map": {

tokenizer_config.json

@@ -1107,7 +1107,7 @@
       "special": false
     },
     "138": {
-      "content": "\u2581\u2581",
+      "content": "▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1115,7 +1115,7 @@
       "special": false
     },
     "139": {
-      "content": "\u2581\u2581\u2581",
+      "content": "▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1123,7 +1123,7 @@
       "special": false
     },
     "140": {
-      "content": "\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1131,7 +1131,7 @@
       "special": false
     },
     "141": {
-      "content": "\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1139,7 +1139,7 @@
       "special": false
     },
     "142": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1147,7 +1147,7 @@
       "special": false
     },
     "143": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1155,7 +1155,7 @@
       "special": false
     },
     "144": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1163,7 +1163,7 @@
       "special": false
     },
     "145": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1171,7 +1171,7 @@
       "special": false
     },
     "146": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1179,7 +1179,7 @@
       "special": false
     },
     "147": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1187,7 +1187,7 @@
       "special": false
     },
     "148": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1195,7 +1195,7 @@
       "special": false
     },
     "149": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1203,7 +1203,7 @@
       "special": false
     },
     "150": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1211,7 +1211,7 @@
       "special": false
     },
     "151": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1219,7 +1219,7 @@
       "special": false
     },
     "152": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1227,7 +1227,7 @@
       "special": false
     },
     "153": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1235,7 +1235,7 @@
       "special": false
     },
     "154": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1243,7 +1243,7 @@
       "special": false
     },
     "155": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1251,7 +1251,7 @@
       "special": false
     },
     "156": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1259,7 +1259,7 @@
       "special": false
     },
     "157": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1267,7 +1267,7 @@
       "special": false
     },
     "158": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1275,7 +1275,7 @@
       "special": false
     },
     "159": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1283,7 +1283,7 @@
       "special": false
     },
     "160": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1291,7 +1291,7 @@
       "special": false
     },
     "161": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1299,7 +1299,7 @@
       "special": false
     },
     "162": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1307,7 +1307,7 @@
       "special": false
     },
     "163": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1315,7 +1315,7 @@
       "special": false
     },
     "164": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1323,7 +1323,7 @@
       "special": false
     },
     "165": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1331,7 +1331,7 @@
       "special": false
     },
     "166": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -1339,7 +1339,7 @@
       "special": false
     },
     "167": {
-      "content": "\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581\u2581",
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -51325,7 +51325,6 @@
   },
   "boi_token": "<start_of_image>",
   "bos_token": "<bos>",
-  "chat_template": "{{ bos_token }}\n{%- if messages[0]['role'] == 'system' -%}\n    {%- if messages[0]['content'] is string -%}\n        {%- set first_user_prefix = messages[0]['content'] + '\n\n' -%}\n    {%- else -%}\n        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '\n\n' -%}\n    {%- endif -%}\n    {%- set loop_messages = messages[1:] -%}\n{%- else -%}\n    {%- set first_user_prefix = \"\" -%}\n    {%- set loop_messages = messages -%}\n{%- endif -%}\n{%- for message in loop_messages -%}\n    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}\n        {{ raise_exception(\"Conversation roles must alternate user/assistant/user/assistant/...\") }}\n    {%- endif -%}\n    {%- if (message['role'] == 'assistant') -%}\n        {%- set role = \"model\" -%}\n    {%- else -%}\n        {%- set role = message['role'] -%}\n    {%- endif -%}\n    {{ '<start_of_turn>' + role + '\n' + (first_user_prefix if loop.first else \"\") }}\n    {%- if message['content'] is string -%}\n        {{ message['content'] | trim }}\n    {%- elif message['content'] is iterable -%}\n        {%- for item in message['content'] -%}\n            {%- if item['type'] == 'image' -%}\n                {{ '<start_of_image>' }}\n            {%- elif item['type'] == 'text' -%}\n                {{ item['text'] | trim }}\n            {%- endif -%}\n        {%- endfor -%}\n    {%- else -%}\n        {{ raise_exception(\"Invalid content type\") }}\n    {%- endif -%}\n    {{ '<end_of_turn>\n' }}\n{%- endfor -%}\n{%- if add_generation_prompt -%}\n    {{'<start_of_turn>model\n'}}\n{%- endif -%}\n",
   "clean_up_tokenization_spaces": false,
   "eoi_token": "<end_of_image>",
   "eos_token": "<end_of_turn>",
@@ -51343,5 +51342,6 @@
   "spaces_between_special_tokens": false,
   "tokenizer_class": "GemmaTokenizer",
   "unk_token": "<unk>",
-  "use_default_system_prompt": false
+  "use_default_system_prompt": false,
+  "chat_template": "{{ bos_token }}\n{%- if messages[0]['role'] == 'system' -%}\n    {%- if messages[0]['content'] is string -%}\n        {%- set first_user_prefix = messages[0]['content'] + '\n\n' -%}\n    {%- else -%}\n        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '\n\n' -%}\n    {%- endif -%}\n    {%- set loop_messages = messages[1:] -%}\n{%- else -%}\n    {%- set first_user_prefix = \"\" -%}\n    {%- set loop_messages = messages -%}\n{%- endif -%}\n{%- for message in loop_messages -%}\n    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}\n        {{ raise_exception(\"Conversation roles must alternate user/assistant/user/assistant/...\") }}\n    {%- endif -%}\n    {%- if (message['role'] == 'assistant') -%}\n        {%- set role = \"model\" -%}\n    {%- else -%}\n        {%- set role = message['role'] -%}\n    {%- endif -%}\n    {{ '<start_of_turn>' + role + '\n' + (first_user_prefix if loop.first else \"\") }}\n    {%- if message['content'] is string -%}\n        {{ message['content'] | trim }}\n    {%- elif message['content'] is iterable -%}\n        {%- for item in message['content'] -%}\n            {%- if item['type'] == 'image' -%}\n                {{ '<start_of_image>' }}\n            {%- elif item['type'] == 'text' -%}\n                {{ item['text'] | trim }}\n            {%- endif -%}\n        {%- endfor -%}\n    {%- else -%}\n        {{ raise_exception(\"Invalid content type\") }}\n    {%- endif -%}\n    {{ '<end_of_turn>\n' }}\n{%- endfor -%}\n{%- if add_generation_prompt -%}\n    {{'<start_of_turn>model\n'}}\n{%- endif -%}\n"
 }