Instructions to use raku10-ghimire/nepali_caption_model_output with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use raku10-ghimire/nepali_caption_model_output with Transformers:

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

pipe = pipeline("image-text-to-text", model="raku10-ghimire/nepali_caption_model_output")

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("raku10-ghimire/nepali_caption_model_output")
model = AutoModelForMultimodalLM.from_pretrained("raku10-ghimire/nepali_caption_model_output", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use raku10-ghimire/nepali_caption_model_output with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "raku10-ghimire/nepali_caption_model_output"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "raku10-ghimire/nepali_caption_model_output",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/raku10-ghimire/nepali_caption_model_output

SGLang

How to use raku10-ghimire/nepali_caption_model_output 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 "raku10-ghimire/nepali_caption_model_output" \
    --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": "raku10-ghimire/nepali_caption_model_output",
		"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 "raku10-ghimire/nepali_caption_model_output" \
        --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": "raku10-ghimire/nepali_caption_model_output",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use raku10-ghimire/nepali_caption_model_output with Docker Model Runner:
```
docker model run hf.co/raku10-ghimire/nepali_caption_model_output
```

nepali_caption_model_output

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

Loss: 1.5684

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: 5e-05
train_batch_size: 2
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 25
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.0445	1.0	450	1.0425
1.0711	2.0	900	0.9059
0.8319	3.0	1350	0.8620
0.6918	4.0	1800	0.8389
0.6192	5.0	2250	0.8225
0.4972	6.0	2700	0.8493
0.4747	7.0	3150	0.9033
0.3015	8.0	3600	0.9647
0.2784	9.0	4050	1.0026
0.2438	10.0	4500	1.0574
0.1858	11.0	4950	1.1057
0.1774	12.0	5400	1.1493
0.1633	13.0	5850	1.1878
0.1348	14.0	6300	1.2216
0.1203	15.0	6750	1.2924
0.1047	16.0	7200	1.2917
0.0939	17.0	7650	1.3597
0.0796	18.0	8100	1.3927
0.0640	19.0	8550	1.4168
0.0510	20.0	9000	1.4512
0.0386	21.0	9450	1.4840
0.0339	22.0	9900	1.5199
0.0194	23.0	10350	1.5498
0.0137	24.0	10800	1.5665
0.0152	25.0	11250	1.5684

Framework versions

Transformers 5.13.1
Pytorch 2.11.0+cu128
Datasets 4.0.0
Tokenizers 0.22.2

Downloads last month: 14

Safetensors

Model size

0.2B 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 raku10-ghimire/nepali_caption_model_output

Base model

Salesforce/blip-image-captioning-base

Finetuned

(59)

this model