North Micro Vision Instruct
North Micro Vision Instruct is a 2.4B-parameter open-weight vision-language model from CohereLabs, released under the Apache-2.0 license. It uses an image-text-to-text pipeline and supports native-resolution images while preserving aspect ratios. The model accepts interleaved text and images and produces text. Its language backbone has a 128K-token context window, while multimodal training used an 8K-token context. Model weights are provided in bfloat16 and safetensors format. The model is described as multilingual, conversational, and multimodal, with support for VQA, captioning, grounding, OCR, charts, and documents.
At a glance
- Parameters
- 2.48B
- Architecture
- cohere_compass
- License
- apache-2.0
- Disk space
- 4.6 GB
- Software
- Transformers
Will it run on your machine?
Save your machine to see a personalized rating and its reasoning.
Add my hardwareBest for
- Visual question answering
- Image captioning
- Visual grounding
- OCR from images
- Chart and document understanding
Tradeoffs
- Multimodal prompts are validated up to an 8K-token context; longer multimodal contexts may rely on extrapolation and have not been benchmarked.
- Only the model weights are available under Apache-2.0; the repository does not specify licenses for other assets.
- The model card does not document compatibility beyond the stated PyTorch, Transformers, accelerate, and Pillow requirements.
- No performance benchmarks or numerical hardware requirements are provided in the supplied model facts.
- The model card does not include claims about testing or purchase recommendations.
Ways to run it
Transformers · See official guide
Repository-specific command in publisher documentation; confirm dependencies and hardware in the source.
import torch
from transformers import AutoModelForImageTextToText, AutoProcessor
model_id = "CohereLabs/North-Micro-Vision-Instruct"
processor = AutoProcessor.from_pretrained(
model_id,
)
model = AutoModelForImageTextToText.from_pretrained(
model_id,
dtype="auto",
device_map="auto",
)
# To enable Flash Attention 2, load the model with the following settings:
# model = AutoModelForImageTextToText.from_pretrained(
# model_id,
# dtype=torch.bfloat16,
# attn_implementation="flash_attention_2",
# device_map="auto",
# )
image_url = "https://cdn-uploads.huggingface.co/production/uploads/66d732effe6684fc16b12c28/Io_5OCmftsmH-n158ZtPs.png"
messages = [
{
"role": "user",
"content": [
{"type": "image", "url": image_url},
{"type": "text", "text": "What do you see?"},
],
}
]
inputs = processor.apply_chat_template(
messages,
tokenize=True,
add_generation_prompt=True,
return_tensors="pt",
return_dict=True,
).to(model.device)
outputs = model.generate(
**inputs,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.8,
top_k=20,
)
generated_ids = [
output_ids[len(input_ids) :]
for input_ids, output_ids in zip(inputs.input_ids, outputs)
]
response = processor.batch_decode(
generated_ids,
skip_special_tokens=True,
clean_up_tokenization_spaces=False,
)[0]
print(response)Explore its uses
Keep exploring
Bonsai 2 · 27B
A compressed 27B-class reasoning model with publisher-provided GGUF packs and a dedicated llama.cpp fork for CUDA, Metal, and CPU.
DeepSeek V4.1 Flash
An open-weight model for reasoning, coding, agents, and image understanding. DeepSeek V4.1 Flash combines a 552B backbone with sparse activation and up to one million tokens of context. Choose the original checkpoint or a community uncensored variant below.
Qwen3.8
Explore 4 versions of Qwen3.8. Compare model sizes, packages, hardware requirements, and available setups.