Instructions to use moebouassida/medgemma-4b-path-vqa with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use moebouassida/medgemma-4b-path-vqa with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("google/medgemma-4b-it")
model = PeftModel.from_pretrained(base_model, "moebouassida/medgemma-4b-path-vqa")

Transformers

How to use moebouassida/medgemma-4b-path-vqa with Transformers:

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

pipe = pipeline("text-generation", model="moebouassida/medgemma-4b-path-vqa")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("moebouassida/medgemma-4b-path-vqa", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use moebouassida/medgemma-4b-path-vqa with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "moebouassida/medgemma-4b-path-vqa"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "moebouassida/medgemma-4b-path-vqa",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/moebouassida/medgemma-4b-path-vqa

SGLang

How to use moebouassida/medgemma-4b-path-vqa 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 "moebouassida/medgemma-4b-path-vqa" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "moebouassida/medgemma-4b-path-vqa",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "moebouassida/medgemma-4b-path-vqa" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "moebouassida/medgemma-4b-path-vqa",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use moebouassida/medgemma-4b-path-vqa with Docker Model Runner:
```
docker model run hf.co/moebouassida/medgemma-4b-path-vqa
```

medgemma-4b-path-vqa

This model is a fine-tuned version of google/medgemma-4b-it on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.9293

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: 0.0002
train_batch_size: 8
eval_batch_size: 8
seed: 3407
gradient_accumulation_steps: 4
total_train_batch_size: 32
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: cosine
lr_scheduler_warmup_steps: 30
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.1603	0.1628	100	1.1339
1.0356	0.3256	200	1.0636
1.0304	0.4884	300	1.0121
0.9684	0.6512	400	0.9663
0.8976	0.8140	500	0.9384
0.9495	0.9768	600	0.9294
0.9196	1.0	615	0.9293

Framework versions

PEFT 0.19.1
Transformers 5.11.0
Pytorch 2.6.0+cu124
Datasets 5.0.0
Tokenizers 0.22.2

Downloads last month: 5

Model tree for moebouassida/medgemma-4b-path-vqa

Base model

google/gemma-3-4b-pt

Finetuned

google/medgemma-4b-pt

Finetuned

google/medgemma-4b-it

Adapter

(114)

this model