GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
Video-R1
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
Video-R1
Model Tree
Qwen/Qwen2.5-VL-7B-Instruct
Input Modalities
Output Modalities
Supported Functionality
This repository contains the Video-R1-7B model as presented in Video-R1: Reinforcing Video Reasoning in MLLMs.
For training and evaluation, please refer to the code: https://github.com/tulerfeng/Video-R1
For inference on a single example, you may refer to: https://github.com/tulerfeng/Video-R1/blob/main/src/inference_example.py
python
import osimport torchfrom vllm import LLM, SamplingParamsfrom transformers import AutoProcessor, AutoTokenizerfrom qwen_vl_utils import process_vision_info # Set model pathmodel_path = "Video-R1/Video-R1-7B" # Set video path and questionvideo_path = "./src/example_video/video1.mp4"question = "Which move motion in the video lose the system energy?" # Choose the question type from 'multiple choice', 'numerical', 'OCR', 'free-form', 'regression'problem_type = 'free-form' # Initialize the LLMllm = LLM( model=model_path, tensor_parallel_size=1, max_model_len=81920, gpu_memory_utilization=0.8, limit_mm_per_prompt={"video": 1, "image": 1},) sampling_params = SamplingParams( temperature=0.1, top_p=0.001, max_tokens=1024,) # Load processor and tokenizerprocessor = AutoProcessor.from_pretrained(model_path)tokenizer = AutoTokenizer.from_pretrained(model_path)tokenizer.padding_side = "left"processor.tokenizer = tokenizer # Prompt templateQUESTION_TEMPLATE = ( "{Question}\n" "Please think about this question as if you were a human pondering deeply. " "Engage in an internal dialogue using expressions such as 'let me think', 'wait', 'Hmm', 'oh, I see', 'let's break it down', etc, or other natural language thought expressions " "It's encouraged to include self-reflection or verification in the reasoning process. " "Provide your detailed reasoning between the <think> and </think> tags, and then give your final answer between the <answer> and </answer> tags.") # Question type TYPE_TEMPLATE = { "multiple choice": " Please provide only the single option letter (e.g., A, B, C, D, etc.) within the <answer> </answer> tags.", "numerical": " Please provide the numerical value (e.g., 42 or 3.14) within the <answer> </answer> tags.", "OCR": " Please transcribe text from the image/video clearly and provide your text answer within the <answer> </answer> tags.", "free-form": " Please provide your text answer within the <answer> </answer> tags.", "regression": " Please provide the numerical value (e.g., 42 or 3.14) within the <answer> </answer> tags."} # Construct multimodal messagemessages = [ { "role": "user", "content": [ { "type": "video", "video": video_path, "max_pixels": 200704, # max pixels for each frame "nframes": 32 # max frame number }, { "type": "text", "text": QUESTION_TEMPLATE.format(Question=question) + TYPE_TEMPLATE[problem_type] }, ], }] # Convert to prompt stringprompt = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) # Process video inputimage_inputs, video_inputs, video_kwargs = process_vision_info(messages, return_video_kwargs=True) # Prepare vllm inputllm_inputs = [{ "prompt": prompt, "multi_modal_data": {"video": video_inputs[0]}, "mm_processor_kwargs": {key: val[0] for key, val in video_kwargs.items()},}] # Run inferenceoutputs = llm.generate(llm_inputs, sampling_params=sampling_params)output_text = outputs[0].outputs[0].text print(output_text)import osimport torchfrom vllm import LLM, SamplingParamsfrom transformers import AutoProcessor, AutoTokenizerfrom qwen_vl_utils import process_vision_info # Set model pathmodel_path = "Video-R1/Video-R1-7B" # Set video path and questionvideo_path = "./src/example_video/video1.mp4"question = "Which move motion in the video lose the system energy?" # Choose the question type from 'multiple choice', 'numerical', 'OCR', 'free-form', 'regression'problem_type = 'free-form' # Initialize the LLMllm = LLM( model=model_path, tensor_parallel_size=1, max_model_len=81920, gpu_memory_utilization=0.8, limit_mm_per_prompt={"video": 1, "image": 1},) sampling_params = SamplingParams( temperature=0.1, top_p=0.001, max_tokens=1024,) # Load processor and tokenizerprocessor = AutoProcessor.from_pretrained(model_path)tokenizer = AutoTokenizer.from_pretrained(model_path)tokenizer.padding_side = "left"processor.tokenizer = tokenizer # Prompt templateQUESTION_TEMPLATE = ( "{Question}\n" "Please think about this question as if you were a human pondering deeply. " "Engage in an internal dialogue using expressions such as 'let me think', 'wait', 'Hmm', 'oh, I see', 'let's break it down', etc, or other natural language thought expressions " "It's encouraged to include self-reflection or verification in the reasoning process. " "Provide your detailed reasoning between the <think> and </think> tags, and then give your final answer between the <answer> and </answer> tags.") # Question type TYPE_TEMPLATE = { "multiple choice": " Please provide only the single option letter (e.g., A, B, C, D, etc.) within the <answer> </answer> tags.", "numerical": " Please provide the numerical value (e.g., 42 or 3.14) within the <answer> </answer> tags.", "OCR": " Please transcribe text from the image/video clearly and provide your text answer within the <answer> </answer> tags.", "free-form": " Please provide your text answer within the <answer> </answer> tags.", "regression": " Please provide the numerical value (e.g., 42 or 3.14) within the <answer> </answer> tags."} # Construct multimodal messagemessages = [ { "role": "user", "content": [ { "type": "video", "video": video_path, "max_pixels": 200704, # max pixels for each frame "nframes": 32 # max frame number }, { "type": "text", "text": QUESTION_TEMPLATE.format(Question=question) + TYPE_TEMPLATE[problem_type] }, ], }] # Convert to prompt stringprompt = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) # Process video inputimage_inputs, video_inputs, video_kwargs = process_vision_info(messages, return_video_kwargs=True) # Prepare vllm inputllm_inputs = [{ "prompt": prompt, "multi_modal_data": {"video": video_inputs[0]}, "mm_processor_kwargs": {key: val[0] for key, val in video_kwargs.items()},}] # Run inferenceoutputs = llm.generate(llm_inputs, sampling_params=sampling_params)output_text = outputs[0].outputs[0].text print(output_text)