Skip to content
Local AI Wizard
Your machine
Theme

Janus Pro

DeepSeek · https://github.com/deepseek-ai/Janus

Weights license: Publisher terms: MIT (code) / DeepSeek Model License (model) · code and components may use other terms

QualityArtificial Analysis score, rank and sample count excluded: its free tier is internal use only. Source checked 2026-09-22.No comparable score
Published repository size14.9 GBAll published files · not runtime memory
Parameters7 Bmanufacturer model card
Published components1Listed in the model card

Quality

No comparable score

Artificial Analysis score, rank and sample count excluded: its free tier is internal use only. Source checked 2026-09-22. Leaderboard (link only)

Reported by the publisher

This model card has not been read for benchmark results yet. card not read yet

Runs on

Enough memory for the estimate 57

The 16-bit planning estimate fits this memory. It is not a reviewed minimum.

Show all 57

One memory tier short 25

The same machines with the next memory size up would have enough for the estimate.

Show all 25

The publisher documents no minimum for this checkpoint (sources checked 2026-10-13). Hardware profiles are grouped by a labelled 16-bit planning estimate from the published parameter count (2 bytes per parameter, 20% allowance, 2 GB reserve). Precision, resolution, duration, text encoders and offload change the real requirement. Not a measured run.

Plan memory for this model

Consider it for

Consider Janus Pro from DeepSeek for the text-to-image task (7B parameters).

Check first

This catalog has not verified hardware requirements for this entry. Actual memory depends on precision, resolution, duration and offload; consult the official source.

Measured speed

No published measurement for this checkpoint

Generation time per image is shown only when a source publishes it with hardware, build and setup. It is never estimated.

How to run

Hardware requirement not reviewed. This catalog has not verified hardware requirements for this entry. Actual memory depends on precision, resolution, duration and offload; consult the official source.

Documented ways to run it

Copied as printed in the linked source (read 2026-10-18); not run by this site.

Setup stated first by the source · Janus
pip install -e .
Janus · README · Text-to-Image Generation
import os
import PIL.Image
import torch
import numpy as np
from transformers import AutoModelForCausalLM
from janus.models import MultiModalityCausalLM, VLChatProcessor


# specify the path to the model
model_path = "deepseek-ai/Janus-Pro-7B"
vl_chat_processor: VLChatProcessor = VLChatProcessor.from_pretrained(model_path)
tokenizer = vl_chat_processor.tokenizer

vl_gpt: MultiModalityCausalLM = AutoModelForCausalLM.from_pretrained(
    model_path, trust_remote_code=True
)
vl_gpt = vl_gpt.to(torch.bfloat16).cuda().eval()

conversation = [
    {
        "role": "<|User|>",
        "content": "A stunning princess from kabul in red, white traditional clothing, blue eyes, brown hair",
    },
    {"role": "<|Assistant|>", "content": ""},
]

sft_format = vl_chat_processor.apply_sft_template_for_multi_turn_prompts(
    conversations=conversation,
    sft_format=vl_chat_processor.sft_format,
    system_prompt="",
)
prompt = sft_format + vl_chat_processor.image_start_tag


@torch.inference_mode()
def generate(
    mmgpt: MultiModalityCausalLM,
    vl_chat_processor: VLChatProcessor,
    prompt: str,
    temperature: float = 1,
    parallel_size: int = 16,
    cfg_weight: float = 5,
    image_token_num_per_image: int = 576,
    img_size: int = 384,
    patch_size: int = 16,
):
    input_ids = vl_chat_processor.tokenizer.encode(prompt)
    input_ids = torch.LongTensor(input_ids)

    tokens = torch.zeros((parallel_size*2, len(input_ids)), dtype=torch.int).cuda()
    for i in range(parallel_size*2):
        tokens[i, :] = input_ids
        if i % 2 != 0:
            tokens[i, 1:-1] = vl_chat_processor.pad_id

    inputs_embeds = mmgpt.language_model.get_input_embeddings()(tokens)

    generated_tokens = torch.zeros((parallel_size, image_token_num_per_image), dtype=torch.int).cuda()

    for i in range(image_token_num_per_image):
        outputs = mmgpt.language_model.model(inputs_embeds=inputs_embeds, use_cache=True, past_key_values=outputs.past_key_values if i != 0 else None)
        hidden_states = outputs.last_hidden_state
        
        logits = mmgpt.gen_head(hidden_states[:, -1, :])
        logit_cond = logits[0::2, :]
        logit_uncond = logits[1::2, :]
        
        logits = logit_uncond + cfg_weight * (logit_cond-logit_uncond)
        probs = torch.softmax(logits / temperature, dim=-1)

        next_token = torch.multinomial(probs, num_samples=1)
        generated_tokens[:, i] = next_token.squeeze(dim=-1)

        next_token = torch.cat([next_token.unsqueeze(dim=1), next_token.unsqueeze(dim=1)], dim=1).view(-1)
        img_embeds = mmgpt.prepare_gen_img_embeds(next_token)
        inputs_embeds = img_embeds.unsqueeze(dim=1)


    dec = mmgpt.gen_vision_model.decode_code(generated_tokens.to(dtype=torch.int), shape=[parallel_size, 8, img_size//patch_size, img_size//patch_size])
    dec = dec.to(torch.float32).cpu().numpy().transpose(0, 2, 3, 1)

    dec = np.clip((dec + 1) / 2 * 255, 0, 255)

    visual_img = np.zeros((parallel_size, img_size, img_size, 3), dtype=np.uint8)
    visual_img[:, :, :] = dec

    os.makedirs('generated_samples', exist_ok=True)
    for i in range(parallel_size):
        save_path = os.path.join('generated_samples', "img_{}.jpg".format(i))
        PIL.Image.fromarray(visual_img[i]).save(save_path)


generate(
    vl_gpt,
    vl_chat_processor,
    prompt,
)

Evidence level: Generic estimate only; no execution reproduced by this site.

Before downloading
  1. Check the exact checkpoint and license linked here.
  2. Confirm a compatible runtime and precision or quantization for your platform.
  3. Allow for context, batch size and any extra media components beyond the weights.

Technical details

What the source and the catalog store for this checkpoint. Missing fields are left out, never zero.

Task and modality
text-to-image · image
Count and basis
7B · manufacturer model card
Released
deepseek-ai/Janus README (GitHub), News · read 2026-10-09source
Published repository size
14.9 GBNot runtime memory or a minimum download.files
Published components
Technical details pending review
Runtime and precision
No local execution verified.
Commercial use
Allowed DeepSeek Model License (deepseek-ai/DeepSeek-LLM LICENSE-MODEL) · read 2026-10-18license text

License, parameter count and release date come from the publisher's sources listed on this page (a date no source states is left empty). The hardware requirement has not been reviewed, so memory is a labelled estimate. The repository size, when shown, is the full published download.

Sources

Identity and published specifications come from the checkpoint page. Benchmark scores belong to the named release and do not measure speed on your computer.

Catalog reviewed: 2026-09-22.