Instructions to use saakshigupta/blip-finetuned-gradcam-optimized with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use saakshigupta/blip-finetuned-gradcam-optimized with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="saakshigupta/blip-finetuned-gradcam-optimized")

# Load model directly
from transformers import AutoProcessor, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("saakshigupta/blip-finetuned-gradcam-optimized")
model = AutoModelForImageTextToText.from_pretrained("saakshigupta/blip-finetuned-gradcam-optimized")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use saakshigupta/blip-finetuned-gradcam-optimized with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "saakshigupta/blip-finetuned-gradcam-optimized"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "saakshigupta/blip-finetuned-gradcam-optimized",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/saakshigupta/blip-finetuned-gradcam-optimized

SGLang

How to use saakshigupta/blip-finetuned-gradcam-optimized with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "saakshigupta/blip-finetuned-gradcam-optimized" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "saakshigupta/blip-finetuned-gradcam-optimized",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "saakshigupta/blip-finetuned-gradcam-optimized" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "saakshigupta/blip-finetuned-gradcam-optimized",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use saakshigupta/blip-finetuned-gradcam-optimized with Docker Model Runner:
```
docker model run hf.co/saakshigupta/blip-finetuned-gradcam-optimized
```

blip-finetuned-gradcam-optimized

This model is a fine-tuned version of Salesforce/blip-image-captioning-large on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 6.6748

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-06
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 8
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 50
num_epochs: 15
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
9.8985	0.4444	10	9.8608
9.7965	0.8889	20	9.7389
9.7009	1.3111	30	9.6374
9.539	1.7556	40	9.5101
8.2904	2.1778	50	9.1547
8.7882	2.6222	60	8.7516
7.6249	3.0444	70	8.4009
8.2185	3.4889	80	8.1483
8.0129	3.9333	90	7.9948
7.8556	4.3556	100	7.8385
7.7265	4.8	110	7.6990
7.6183	5.2222	120	7.5757
7.5	5.6667	130	7.4770
6.6711	6.0889	140	7.3812
7.3249	6.5333	150	7.3007
7.2369	6.9778	160	7.2224
7.1824	7.4	170	7.1548
7.1173	7.8444	180	7.0876
7.0568	8.2667	190	7.0291
7.0033	8.7111	200	6.9772
6.2525	9.1333	210	6.9274
6.9084	9.5778	220	6.8845
6.1803	10.0	230	6.8465
6.8372	10.4444	240	6.8137
6.807	10.8889	250	6.7842
6.7844	11.3111	260	6.7578
6.7499	11.7556	270	6.7369
6.0823	12.1778	280	6.7185
6.7188	12.6222	290	6.7038
6.036	13.0444	300	6.6918
6.7017	13.4889	310	6.6834
6.7003	13.9333	320	6.6776
6.6899	14.3556	330	6.6748

Framework versions

Transformers 4.50.3
Pytorch 2.6.0+cu124
Datasets 3.5.0
Tokenizers 0.21.1

Downloads last month: 9

Safetensors

Model size

0.5B params

Tensor type

F32

Inference Providers NEW

Image-Text-to-Text

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for saakshigupta/blip-finetuned-gradcam-optimized

Base model

Salesforce/blip-image-captioning-large

Finetuned

(13)

this model