# Manually generate generate\_ids using BlipForConditionalGeneration

**URL:** <https://discuss.huggingface.co/t/manually-generate-generate-ids-using-blipforconditionalgeneration/82868>\
**Category:** Models\
**Created:** [April 21, 2024, 5:25am UTC](https://discuss.huggingface.co/t/manually-generate-generate-ids-using-blipforconditionalgeneration/82868 "2024-04-21T05:25:00Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![ummehabiba17](https://avatars.discourse-cdn.com/v4/letter/u/c57346/32.png) [@ummehabiba17](https://discuss.huggingface.co/u/ummehabiba17)\
**Post date:** [April 21, 2024, 5:25am UTC](https://discuss.huggingface.co/t/manually-generate-generate-ids-using-blipforconditionalgeneration/82868/1 "2024-04-21T05:25:00Z")

</div>

from transformers import BlipForConditionalGeneration,AutoProcessor,AutoTokenizer  
from PIL import Image  
import requests  
import torch  
processor= AutoProcessor.from\_pretrained(“Salesforce/blip-image-captioning-base”)  
model= BlipForConditionalGeneration.from\_pretrained(“Salesforce/blip-image-captioning-base”)  
tokenizer = AutoTokenizer.from\_pretrained(“Salesforce/blip-image-captioning-base”)

url = “[http://images.cocodataset.org/val2017/000000039769.jpg”](http://images.cocodataset.org/val2017/000000039769.jpg%E2%80%9D)  
image = Image.open(requests.get(url, stream=True).raw)  
text = “A picture of”

decoder\_input\_ids = [model.decoder\_input\_ids]  
print(decoder\_input\_ids)  
inputs.input\_ids = decoder\_input\_ids=torch.tensor([decoder\_input\_ids])  
predicted\_ids =  
for i in range(20):  
outputs = model(\*\*inputs)  
logits = outputs.logits[:,i,:]  
print(outputs.logits.shape)  
predicted\_id = logits.argmax(-1)  
predicted\_ids.append(predicted\_id.item())  
print(tokenizer.decode([predicted\_id.squeeze()]))  
add predicted id to decoder\_input\_ids  
inputs.input\_ids = torch.cat([inputs.input\_ids, predicted\_id.unsqueeze(0)], dim=1)

for this code for every image i found same logits shape [1, 5, 30524] . and also every iteration the logits are remain same. no change in logtis for changing the inputs.input\_ids. why i face this error? and how to solve this?