.gitattributes

@@ -0,0 +1,38 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz 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
+model.safetensors filter=lfs diff=lfs merge=lfs -text
+tokenizer.model filter=lfs diff=lfs merge=lfs -text

1/model.py

@@ -0,0 +1,243 @@
+import triton_python_backend_utils as pb_utils
+from transformers import AutoModelForCausalLM, AutoTokenizer, GenerationConfig
+import numpy as np
+import json
+import os
+
+class TritonPythonModel:
+    def initialize(self, args):
+        """
+        모델이 로드될 때 딱 한 번만 호출됩니다.
+        `initialize` 함수를 구현하는 것은 선택 사항입니다. 이 함수를 통해 모델은
+        이 모델과 관련된 모든 상태를 초기화할 수 있습니다.
+        """
+        self.logger = pb_utils.Logger
+
+        current_file_path = os.path.abspath(__file__)
+        self.logger.log_info(f"current_file_path: {current_file_path}")
+
+
+        self.model_name = args["model_name"]
+        model_repository = args["model_repository"]
+        model_path = f"{model_repository}/{self.model_name}"
+        #model_path = "/cheetah/input/model/gemma-3-1b-it/gemma-3-1b-it"
+
+        input_model_path = model_path
+
+        if os.path.exists(input_model_path):
+            file_list = os.listdir(input_model_path)
+            self.logger.log_info(f"'{input_model_path}' 디렉토리의 파일 목록:")
+        for file_name in file_list:
+            self.logger.log_info(file_name)
+        else:
+            self.logger.log_info(f"'{input_model_path}' 디렉토리가 존재하지 않습니다.")
+
+        self.logger.log_info(f"model_repository: {model_repository}")
+        self.logger.log_info(f"model_path: {model_path}")
+
+        self.model_config = json.loads(args["model_config"])
+
+        # Hugging Face Transformers 라이브러리에서 사전 학습된 토크나이저를 로드합니다.
+        self.tokenizer = AutoTokenizer.from_pretrained(model_path)
+        self.tokenizer.pad_token_id = self.tokenizer.eos_token_id
+        self.supports_chat_template = self._check_chat_template_support()
+
+        # Hugging Face Transformers 라이브러리에서 사전 학습된 언어 모델을 로드합니다.
+        self.model = AutoModelForCausalLM.from_pretrained(
+            pretrained_model_name_or_path=model_path,
+            local_files_only=True,
+            device_map="auto"
+        )
+
+        self.enable_inference_trace = self._get_inference_trace_setting()
+
+        self.logger.log_info(f"'{self.model_name}' 모델 초기화 완료")
+
+
+    def execute(self, requests):
+        """
+        Triton이 각 추론 요청에 대해 호출하는 실행 함수입니다.
+        """
+        responses = []
+
+        # 각 추론 요청을 순회하며 처리합니다.
+        for request in requests:
+            # Triton 입력 파싱            
+            input_text = self._get_input_value(request, "text_input")
+            
+            text = ""
+            conversation = ""
+            input_token_length = 0  # 입력 토큰 길이를 저장할 변수
+
+            # 입력 텍스트가 JSON 형식의 대화 기록인지 확인합니다.
+            try:
+                conversation = json.loads(input_text)
+                is_chat = True
+                self.logger.log_info(f"입력 conversation 출력:\n{conversation}")
+            except:
+                # JSON 파싱에 실패하면 일반 텍스트로 처리합니다.
+                text = input_text
+                is_chat = False
+                self.logger.log_info(f"입력 text 출력:\n{text}")
+            
+            # 입력 텍스트를 토큰화합니다.
+            if self.supports_chat_template and is_chat:
+                self.logger.log_info(f"Chat 템플릿을 적용하여 토큰화합니다.")
+                inputs = self.tokenizer.apply_chat_template(
+                                            conversation,
+                                            tokenize=True,
+                                            add_generation_prompt=True,
+                                            return_tensors="pt",
+                                            return_dict=True
+                                            ).to(device=self.model.device)
+            else:
+                self.logger.log_info(f"입력 텍스트를 토큰화합니다.")
+                inputs = self.tokenizer(
+                            text,
+                            return_tensors="pt").to(device=self.model.device)
+
+            input_ids = inputs["input_ids"]
+            attention_mask = inputs["attention_mask"]
+            input_token_length = inputs["input_ids"].shape[-1]
+
+
+            # 언어 모델을 사용하여 텍스트를 생성합니다.
+            gened = self.model.generate(
+                input_ids=input_ids,
+                attention_mask=attention_mask,
+                generation_config=self._process_generation_config(request),
+                pad_token_id=self.tokenizer.pad_token_id,
+            )
+
+            # 생성된 토큰 시퀀스를 텍스트로 디코딩하고 입력 텍스트는 제외합니다.
+            generated_tokens = gened[0][input_token_length:] # 입력 토큰 이후부터 슬라이싱
+            gened_text = self.tokenizer.decode(generated_tokens, skip_special_tokens=True)
+            self.logger.log_info(f"모델이 생성한 토큰 시퀀스 (입력 텍스트 제외):\n{gened_text}")
+
+            output = gened_text.strip()
+
+            # 생성된 텍스트를 Triton 출력 텐서로 변환합니다.
+            output_tensor = pb_utils.Tensor("text_output", np.array(output.encode('utf-8'), dtype=np.bytes_))
+
+            # 응답 객체를 생성하고 출력 텐서를 추가합니다.
+            responses.append(pb_utils.InferenceResponse(output_tensors=[output_tensor]))
+
+        return responses
+    
+    def _process_generation_config(self, request):
+        """
+        추론 요청에서 생성 설정 관련 파라미터들을 추출하여 GenerationConfig 객체를 생성합니다.
+
+        Args:
+            request (pb_utils.InferenceRequest): Triton 추론 요청 객체.
+
+        Returns:
+            transformers.GenerationConfig: GenerationConfig 객체.
+        """
+        max_length = self._get_input_value(request, "max_length", default=20)
+        max_new_tokens = self._get_input_value(request, "max_new_tokens")
+        temperature = self._get_input_value(request, "temperature")
+        do_sample = self._get_input_value(request, "do_sample")
+        top_k = self._get_input_value(request, "top_k")
+        top_p = self._get_input_value(request, "top_p")
+        repetition_penalty = self._get_input_value(request, "repetition_penalty")
+        stream = self._get_input_value(request, "stream")
+
+        generation_config = GenerationConfig(
+            max_length=max_length,
+            max_new_tokens=max_new_tokens,
+            temperature=temperature,
+            do_sample=do_sample,
+            top_k=top_k,
+            top_p=top_p,
+            repetition_penalty=repetition_penalty,
+            stream=stream,
+        )
+
+        self.logger.log_info(f"추론 요청 GenerationConfig:\n{generation_config}")
+
+        return generation_config
+
+    def _get_inference_trace_setting(self):
+        """
+        모델 설정(config.pbxt)에서 'enable_inference_trace' 값을 추출하여 반환합니다.
+
+        'enable_inference_trace' 설정이 없거나, 올바른 형식이 아닌 경우 기본적으로 False를 반환합니다.
+
+        Returns:
+            bool: 추론 추적 활성화 여부 (True 또는 False).
+        """
+        parameters = self.model_config.get('parameters', {})
+        trace_config = parameters.get('enable_inference_trace')
+        if isinstance(trace_config, dict) and 'string_value' in trace_config:
+            return trace_config['string_value'].lower() == 'true'  # 문자열 값을 bool로 변환하여 반환
+        return False
+    
+
+    def _check_chat_template_support(self):
+      """
+      주어진 허깅페이스 Transformer 모델이 Chat 템플릿을 지원하는지 확인하고 결과를 출력합니다.
+      
+      Returns:
+            bool: Chat 템플릿 지원 여부 (True 또는 False).
+      """
+      try:
+          if hasattr(self.tokenizer, "chat_template") and self.tokenizer.chat_template is not None:
+              self.logger.log_info(f"'{self.model_name}' 모델의 토크나이저는 Chat 템플릿을 지원합니다.")
+              self.logger.log_info("Chat 템플릿 내용:")
+              self.logger.log_info(self.tokenizer.chat_template)
+              return True
+          else:
+              self.logger.log_info(f"'{self.model_name}' 모델의 토크나이저는 Chat 템플릿을 직접적으로 지원하지 않거나, Chat 템플릿 정보가 없습니다.")
+              return False
+      except Exception as e:
+          self.logger.log_info(f"'{self.model_name}' 모델의 토크나이저를 로드하는 동안 오류가 발생했습니다: {e}")
+          return False
+    
+
+    def _get_input_value(self, request, input_name: str, default=None):
+        """
+        Triton 추론 요청에서 특정 이름의 입력 텐서 값을 가져옵니다.
+
+        Args:
+            request (pb_utils.InferenceRequest): Triton 추론 요청 객체.
+            input_name (str): 가져올 입력 텐서의 이름.
+            default (any, optional): 입력 텐서가 없을 경우 반환할 기본값. Defaults to None.
+
+        Returns:
+            any: 디코딩된 입력 텐서의 값. 텐서가 없으면 기본값을 반환합니다.
+        """
+        tensor_value = pb_utils.get_input_tensor_by_name(request, input_name)
+
+        if tensor_value is None:
+            return default
+        
+        return self._np_decoder(tensor_value.as_numpy()[0])
+        
+    def _np_decoder(self, obj):
+        """
+        NumPy 객체의 데이터 타입을 확인하고 Python 기본 타입으로 변환합니다.
+
+        Args:
+            obj (numpy.ndarray element): 변환할 NumPy 배열의 요소.
+
+        Returns:
+            any: 해당 NumPy 요소에 대응하는 Python 기본 타입 (str, int, float, bool).
+                 bytes 타입인 경우 UTF-8로 디코딩합니다.
+        """
+        if isinstance(obj, bytes):
+           return obj.decode('utf-8')
+        if np.issubdtype(obj, np.integer):
+           return int(obj)
+        if np.issubdtype(obj, np.floating):
+           return round(float(obj), 3)
+        if isinstance(obj, np.bool_):
+           return bool(obj)
+
+    def finalize(self):
+        """
+        모델 실행이 완료된 후 Triton 서버가 종료될 때 한 번 호출되는 함수입니다.
+        `finalize` 함수를 구현하는 것은 선택 사항입니다. 이 함수를 통해 모델은
+        종료 전에 필요한 모든 정리 작업을 수행할 수 있습니다.
+        """
+        pass

config.pbtxt

@@ -0,0 +1,89 @@
+# Triton backend to use
+name: "gemma-3-1b-it"
+backend: "python"
+max_batch_size: 0
+
+# Triton should expect as input a single string
+# input of variable length named 'text_input'
+input [
+  {
+    name: "text_input"
+    data_type: TYPE_STRING
+    dims: [ -1 ]
+  },
+  {
+    name: "max_length"
+    data_type: TYPE_INT32
+    dims: [ 1 ]
+    optional: true
+  },
+  {
+    name: "max_new_tokens"
+    data_type: TYPE_INT32
+    dims: [ 1 ]
+    optional: true
+  },
+  {
+    name: "do_sample"
+    data_type: TYPE_BOOL
+    dims: [ 1 ]
+    optional: true
+  },
+  {
+    name: "top_k"
+    data_type: TYPE_INT32
+    dims: [ 1 ]
+    optional: true
+  },
+  {
+    name: "top_p"
+    data_type: TYPE_FP32
+    dims: [ 1 ]
+    optional: true
+  },
+  {
+    name: "temperature"
+    data_type: TYPE_FP32
+    dims: [ 1 ]
+    optional: true
+  },
+  {
+    name: "repetition_penalty"
+    data_type: TYPE_FP32
+    dims: [ 1 ]
+    optional: true
+  },
+  {
+    name: "stream"
+    data_type: TYPE_BOOL
+    dims: [ 1 ]
+    optional: true
+  }
+]
+
+
+# Triton should expect to respond with a single string
+# output of variable length named 'text_output'
+output [ 
+  {
+    name: "text_output"
+    data_type: TYPE_STRING
+    dims: [ -1 ]
+  }
+]
+
+
+parameters: [
+  {
+    key: "enable_inference_trace",
+    value: {string_value: "False"}
+  }
+]
+
+instance_group [
+  {
+    kind: KIND_AUTO,
+    count: 1
+  }
+]
+

gemma-3-1b-it/README.md

@@ -0,0 +1,514 @@
+---
+license: gemma
+library_name: transformers
+pipeline_tag: text-generation
+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-3-1b-pt
+---
+
+# 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
+
+With instruction-tuned models, you need to use chat templates to process our inputs first. Then, you can pass it to the pipeline.
+
+```python
+from transformers import pipeline
+import torch
+
+pipe = pipeline("text-generation", model="google/gemma-3-1b-it", device="cuda", torch_dtype=torch.bfloat16)
+
+messages = [
+    [
+        {
+            "role": "system",
+            "content": [{"type": "text", "text": "You are a helpful assistant."},]
+        },
+        {
+            "role": "user",
+            "content": [{"type": "text", "text": "Write a poem on Hugging Face, the company"},]
+        },
+    ],
+]
+
+output = pipe(messages, max_new_tokens=50)
+```
+
+#### Running the model on a single / multi GPU
+
+```python
+from transformers import AutoTokenizer, BitsAndBytesConfig, Gemma3ForCausalLM
+import torch
+
+model_id = "google/gemma-3-1b-it"
+
+quantization_config = BitsAndBytesConfig(load_in_8bit=True)
+
+model = Gemma3ForCausalLM.from_pretrained(
+    model_id, quantization_config=quantization_config
+).eval()
+
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+
+messages = [
+    [
+        {
+            "role": "system",
+            "content": [{"type": "text", "text": "You are a helpful assistant."},]
+        },
+        {
+            "role": "user",
+            "content": [{"type": "text", "text": "Write a poem on Hugging Face, the company"},]
+        },
+    ],
+]
+inputs = tokenizer.apply_chat_template(
+    messages,
+    add_generation_prompt=True,
+    tokenize=True,
+    return_dict=True,
+    return_tensors="pt",
+).to(model.device).to(torch.bfloat16)
+
+
+with torch.inference_mode():
+    outputs = model.generate(**inputs, max_new_tokens=64)
+
+outputs = tokenizer.batch_decode(outputs)
+```
+
+
+### 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

gemma-3-1b-it/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<image_soft_token>": 262144
+}

gemma-3-1b-it/config.json

@@ -0,0 +1,37 @@
+{
+  "architectures": [
+    "Gemma3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "attn_logit_softcapping": null,
+  "bos_token_id": 2,
+  "cache_implementation": "hybrid",
+  "eos_token_id": [
+    1,
+    106
+  ],
+  "final_logit_softcapping": null,
+  "head_dim": 256,
+  "hidden_activation": "gelu_pytorch_tanh",
+  "hidden_size": 1152,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "model_type": "gemma3_text",
+  "num_attention_heads": 4,
+  "num_hidden_layers": 26,
+  "num_key_value_heads": 1,
+  "pad_token_id": 0,
+  "query_pre_attn_scalar": 256,
+  "rms_norm_eps": 1e-06,
+  "rope_local_base_freq": 10000,
+  "rope_scaling": null,
+  "rope_theta": 1000000,
+  "sliding_window": 512,
+  "sliding_window_pattern": 6,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.50.0.dev0",
+  "use_cache": true,
+  "vocab_size": 262144
+}

gemma-3-1b-it/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": "4.50.0.dev0"
+}

gemma-3-1b-it/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
+  }
+}

gemma-3-1b-it/sub/test.txt

@@ -0,0 +1,6 @@
+test.txt
+test.txt
+test.txt
+test.txt
+test.txt
+