InternVL2-2B
OpenGVLab · OpenGVLab/InternVL2-2B
Licença dos pesos: MIT · código e componentes podem ter outros termos
Qualidade
Sem score comparável
Ainda não acompanhamos benchmark público e comparável que cubra este modelo. Nenhuma nota é inferida.
Informado pelo publicador
- DocVQA test
- 86,9InternVL2-2B · Image Benchmarks
- ChartQA test
- 76,2InternVL2-2B · Image Benchmarks
- InfoVQA test
- 58,9InternVL2-2B · Image Benchmarks
- TextVQA val
- 73,4InternVL2-2B · Image Benchmarks
- OCRBench
- 784InternVL2-2B · Image Benchmarks
- MME sum
- 1.876,8InternVL2-2B · Image Benchmarks
- RealWorldQA
- 57,3InternVL2-2B · Image Benchmarks
- AI2D test
- 74,1InternVL2-2B · Image Benchmarks
Todos os resultados informados
- MMMU val
- 36,3InternVL2-2B · Image Benchmarks
- MMBench-EN test
- 73,2InternVL2-2B · Image Benchmarks
- MMBench-CN test
- 70,9InternVL2-2B · Image Benchmarks
- CCBench dev
- 74,7InternVL2-2B · Image Benchmarks
- MMVet GPT-4-0613
- 44,6InternVL2-2B · Image Benchmarks
- MMVet GPT-4-Turbo
- 39,5InternVL2-2B · Image Benchmarks
- SEED-Image
- 71,6InternVL2-2B · Image Benchmarks
- HallBench avg
- 37,9InternVL2-2B · Image Benchmarks
- MathVista testmini
- 46,3InternVL2-2B · Image Benchmarks
- OpenCompass avg
- 54,0InternVL2-2B · Image Benchmarks
- MVBench
- 60,2InternVL2-2B · Video Benchmarks
- MMBench-Video 8f
- 0,97InternVL2-2B · Video Benchmarks
- MMBench-Video 16f
- 1,03InternVL2-2B · Video Benchmarks
- Video-MME w/o subs
- 45,0InternVL2-2B · Video Benchmarks
- Video-MME w subs
- 47,3InternVL2-2B · Video Benchmarks
- avg.
- 77,7InternVL2‑2B · Grounding Benchmarks
- RefCOCO (val)
- 82,3InternVL2‑2B · Grounding Benchmarks
- RefCOCO (testA)
- 88,2InternVL2‑2B · Grounding Benchmarks
- RefCOCO (testB)
- 75,9InternVL2‑2B · Grounding Benchmarks
- RefCOCO+ (val)
- 73,5InternVL2‑2B · Grounding Benchmarks
- RefCOCO+ (testA)
- 82,8InternVL2‑2B · Grounding Benchmarks
- RefCOCO+ (testB)
- 63,3InternVL2‑2B · Grounding Benchmarks
- RefCOCO‑g (val)
- 77,6InternVL2‑2B · Grounding Benchmarks
- RefCOCO‑g (test)
- 78,3InternVL2‑2B · Grounding Benchmarks
Valores como impressos na ficha, com a configuração do próprio publicador (seção da ficha: Performance › Image Benchmarks; Performance › Video Benchmarks). Não comparáveis entre publicadores e fora de qualquer ranking. Ficha do modelo
Roda em
Cabe na memória rápida 119
A estimativa de planejamento cabe na memória que o modelo usa rápido: a VRAM da placa de vídeo, ou a memória unificada dentro da fatia de GPU do macOS.
Mostrar todos os 119
Perfis de hardware agrupados pela estimativa de planejamento Q4 em cada máquina (GPUs dedicadas reservam 2 GB; memória unificada e do sistema, 25%, de 2 a 8 GB): cabe na memória rápida, fica na memória mas fora da parte rápida (camadas na RAM do PC, ou acima da fatia de GPU do macOS), ou falta um degrau de memória. Não é compatibilidade nem velocidade medida.
OpenGVLab lançou InternVL2-2B em 2024-07-04: um modelo de visão e linguagem que responde sobre imagens, com 2,21B parâmetros pela contagem do Hugging Face e licença MIT. Idiomas declarados pelo publicador: Multilíngue.
Licença: uso comercial permitido. Em Q4 os pesos somam cerca de 1,7 GB antes da reserva de cada máquina; resolução da imagem, quadros de vídeo e contexto somam memória, e o encoder visual precisa de suporte no runtime escolhido.
Velocidade medida
Nenhuma medição publicada para este checkpoint
Tokens por segundo só aparecem quando uma fonte os publica com hardware, build e configuração. Nunca são estimados.
Como rodar
Compatibilidade de runtime e tokens por segundo não foram medidos para este registro. Metadados publicados e estimativa de memória não comprovam uma execução local.
Biblioteca na fonte: transformers
Formas documentadas de rodar
Copiado como aparece na fonte vinculada (lida em 2026-10-18); não executado por este site.
import numpy as np
import torch
import torchvision.transforms as T
from decord import VideoReader, cpu
from PIL import Image
from torchvision.transforms.functional import InterpolationMode
from transformers import AutoModel, AutoTokenizer
IMAGENET_MEAN = (0.485, 0.456, 0.406)
IMAGENET_STD = (0.229, 0.224, 0.225)
def build_transform(input_size):
MEAN, STD = IMAGENET_MEAN, IMAGENET_STD
transform = T.Compose([
T.Lambda(lambda img: img.convert('RGB') if img.mode != 'RGB' else img),
T.Resize((input_size, input_size), interpolation=InterpolationMode.BICUBIC),
T.ToTensor(),
T.Normalize(mean=MEAN, std=STD)
])
return transform
def find_closest_aspect_ratio(aspect_ratio, target_ratios, width, height, image_size):
best_ratio_diff = float('inf')
best_ratio = (1, 1)
area = width * height
for ratio in target_ratios:
target_aspect_ratio = ratio[0] / ratio[1]
ratio_diff = abs(aspect_ratio - target_aspect_ratio)
if ratio_diff < best_ratio_diff:
best_ratio_diff = ratio_diff
best_ratio = ratio
elif ratio_diff == best_ratio_diff:
if area > 0.5 * image_size * image_size * ratio[0] * ratio[1]:
best_ratio = ratio
return best_ratio
def dynamic_preprocess(image, min_num=1, max_num=12, image_size=448, use_thumbnail=False):
orig_width, orig_height = image.size
aspect_ratio = orig_width / orig_height
# calculate the existing image aspect ratio
target_ratios = set(
(i, j) for n in range(min_num, max_num + 1) for i in range(1, n + 1) for j in range(1, n + 1) if
i * j <= max_num and i * j >= min_num)
target_ratios = sorted(target_ratios, key=lambda x: x[0] * x[1])
# find the closest aspect ratio to the target
target_aspect_ratio = find_closest_aspect_ratio(
aspect_ratio, target_ratios, orig_width, orig_height, image_size)
# calculate the target width and height
target_width = image_size * target_aspect_ratio[0]
target_height = image_size * target_aspect_ratio[1]
blocks = target_aspect_ratio[0] * target_aspect_ratio[1]
# resize the image
resized_img = image.resize((target_width, target_height))
processed_images = []
for i in range(blocks):
box = (
(i % (target_width // image_size)) * image_size,
(i // (target_width // image_size)) * image_size,
((i % (target_width // image_size)) + 1) * image_size,
((i // (target_width // image_size)) + 1) * image_size
)
# split the image
split_img = resized_img.crop(box)
processed_images.append(split_img)
assert len(processed_images) == blocks
if use_thumbnail and len(processed_images) != 1:
thumbnail_img = image.resize((image_size, image_size))
processed_images.append(thumbnail_img)
return processed_images
def load_image(image_file, input_size=448, max_num=12):
image = Image.open(image_file).convert('RGB')
transform = build_transform(input_size=input_size)
images = dynamic_preprocess(image, image_size=input_size, use_thumbnail=True, max_num=max_num)
pixel_values = [transform(image) for image in images]
pixel_values = torch.stack(pixel_values)
return pixel_values
# If you want to load a model using multiple GPUs, please refer to the `Multiple GPUs` section.
path = 'OpenGVLab/InternVL2-2B'
model = AutoModel.from_pretrained(
path,
torch_dtype=torch.bfloat16,
low_cpu_mem_usage=True,
use_flash_attn=True,
trust_remote_code=True).eval().cuda()
tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True, use_fast=False)
# set the max number of tiles in `max_num`
pixel_values = load_image('./examples/image1.jpg', max_num=12).to(torch.bfloat16).cuda()
generation_config = dict(max_new_tokens=1024, do_sample=True)
# pure-text conversation (纯文本对话)
question = 'Hello, who are you?'
response, history = model.chat(tokenizer, None, question, generation_config, history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'Can you tell me a story?'
response, history = model.chat(tokenizer, None, question, generation_config, history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')
# single-image single-round conversation (单图单轮对话)
question = '<image>\nPlease describe the image shortly.'
response = model.chat(tokenizer, pixel_values, question, generation_config)
print(f'User: {question}\nAssistant: {response}')
# single-image multi-round conversation (单图多轮对话)
question = '<image>\nPlease describe the image in detail.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config, history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'Please write a poem according to the image.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config, history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')
# multi-image multi-round conversation, combined images (多图多轮对话,拼接图像)
pixel_values1 = load_image('./examples/image1.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values2 = load_image('./examples/image2.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0)
question = '<image>\nDescribe the two images in detail.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'What are the similarities and differences between these two images.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')
# multi-image multi-round conversation, separate images (多图多轮对话,独立图像)
pixel_values1 = load_image('./examples/image1.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values2 = load_image('./examples/image2.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0)
num_patches_list = [pixel_values1.size(0), pixel_values2.size(0)]
question = 'Image-1: <image>\nImage-2: <image>\nDescribe the two images in detail.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
num_patches_list=num_patches_list,
history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'What are the similarities and differences between these two images.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
num_patches_list=num_patches_list,
history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')
# batch inference, single image per sample (单图批处理)
pixel_values1 = load_image('./examples/image1.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values2 = load_image('./examples/image2.jpg', max_num=12).to(torch.bfloat16).cuda()
num_patches_list = [pixel_values1.size(0), pixel_values2.size(0)]
pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0)
questions = ['<image>\nDescribe the image in detail.'] * len(num_patches_list)
responses = model.batch_chat(tokenizer, pixel_values,
num_patches_list=num_patches_list,
questions=questions,
generation_config=generation_config)
for question, response in zip(questions, responses):
print(f'User: {question}\nAssistant: {response}')
# video multi-round conversation (视频多轮对话)
def get_index(bound, fps, max_frame, first_idx=0, num_segments=32):
if bound:
start, end = bound[0], bound[1]
else:
start, end = -100000, 100000
start_idx = max(first_idx, round(start * fps))
end_idx = min(round(end * fps), max_frame)
seg_size = float(end_idx - start_idx) / num_segments
frame_indices = np.array([
int(start_idx + (seg_size / 2) + np.round(seg_size * idx))
for idx in range(num_segments)
])
return frame_indices
def load_video(video_path, bound=None, input_size=448, max_num=1, num_segments=32):
vr = VideoReader(video_path, ctx=cpu(0), num_threads=1)
max_frame = len(vr) - 1
fps = float(vr.get_avg_fps())
pixel_values_list, num_patches_list = [], []
transform = build_transform(input_size=input_size)
frame_indices = get_index(bound, fps, max_frame, first_idx=0, num_segments=num_segments)
for frame_index in frame_indices:
img = Image.fromarray(vr[frame_index].asnumpy()).convert('RGB')
img = dynamic_preprocess(img, image_size=input_size, use_thumbnail=True, max_num=max_num)
pixel_values = [transform(tile) for tile in img]
pixel_values = torch.stack(pixel_values)
num_patches_list.append(pixel_values.shape[0])
pixel_values_list.append(pixel_values)
pixel_values = torch.cat(pixel_values_list)
return pixel_values, num_patches_list
video_path = './examples/red-panda.mp4'
pixel_values, num_patches_list = load_video(video_path, num_segments=8, max_num=1)
pixel_values = pixel_values.to(torch.bfloat16).cuda()
video_prefix = ''.join([f'Frame{i+1}: <image>\n' for i in range(len(num_patches_list))])
question = video_prefix + 'What is the red panda doing?'
# Frame1: <image>\nFrame2: <image>\n...\nFrame8: <image>\n{question}
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
num_patches_list=num_patches_list, history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'Describe this video in detail.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
num_patches_list=num_patches_list, history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')from lmdeploy import pipeline, TurbomindEngineConfig
from lmdeploy.vl import load_image
model = 'OpenGVLab/InternVL2-2B'
image = load_image('https://raw.githubusercontent.com/open-mmlab/mmdeploy/main/tests/data/tiger.jpeg')
pipe = pipeline(model, backend_config=TurbomindEngineConfig(session_len=8192))
response = pipe(('describe this image', image))
print(response.text)Detalhes técnicos
O que a fonte e o catálogo guardam deste checkpoint. Campos ausentes ficam de fora, nunca como zero.
- Checkpoint / fonte
- OpenGVLab/InternVL2-2B
- Tarefa
- image-text-to-text
- Arquitetura
- InternVLChatModel
- Biblioteca na fonte
- transformers
- Licença dos pesos
- MITOpenGVLab/InternVL2-2B model card (Hugging Face) · consultado em 2026-10-18fonte
- Uso comercial
- Permitido MIT license text (SPDX) · consultado em 2026-10-18texto da licença
- Idiomas
- MultilíngueMetadados do model card · consultado em 2026-10-25fonte
- Lançamento
- InternVL README news (GitHub) · consultado em 2026-10-18fonte
Todos os campos guardados
- Revisão conferida
- e4f6747bd20f139e637642c6a058c6bd00b36919
- Arquitetura
- InternVLChatModel
- Biblioteca
- transformers
- Tag de tarefa no Hub
- image-text-to-text
- Modelos base declarados
- OpenGVLab/InternViT-300M-448px · internlm/internlm2-chat-1_8b
- Formatos
- transformers, safetensors
- Acesso
- Download aberto
- Tipo de publicador
- Oficial
- Parâmetros no Hub
- 2.205.754.368safetensors.total
- Repositório criado
- 2024-06-27criação não é o lançamento do modelo
- Repositório atualizado
- 2025-03-25um commit não é data de lançamento
- Downloads no Hub (30 dias) · curtidas
- 240.743 · 83só este repositório · lido em 2026-10-25
- Estado no catálogo
- Config ou pesos encontrados no Hub; sem vínculo de benchmark no inventário
Fontes
Identidade e especificações publicadas vêm da página do checkpoint. As notas pertencem à rodada indicada e não medem velocidade no seu computador.
- Ficha do publicador
- Licença: OpenGVLab/InternVL2-2B model card (Hugging Face)
- Data de lançamento: InternVL README news (GitHub)
Fatos curados verificados: 2026-10-18. A licença e a data de lançamento citam as fontes do publicador; parâmetros, arquivos e tarefa vêm do registro do Hugging Face, e a memória é uma estimativa rotulada a partir da contagem de parâmetros. Um valor que nenhuma fonte informa fica vazio; um commit do repositório não é a data de lançamento do modelo.






























































