Instructions to use ebrukilic/paligemma2_mix_data with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ebrukilic/paligemma2_mix_data with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("google/paligemma2-3b-pt-448")
model = PeftModel.from_pretrained(base_model, "ebrukilic/paligemma2_mix_data")

Transformers

How to use ebrukilic/paligemma2_mix_data with Transformers:

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

pipe = pipeline("text-generation", model="ebrukilic/paligemma2_mix_data")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("ebrukilic/paligemma2_mix_data", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ebrukilic/paligemma2_mix_data with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ebrukilic/paligemma2_mix_data"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ebrukilic/paligemma2_mix_data",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ebrukilic/paligemma2_mix_data

SGLang

How to use ebrukilic/paligemma2_mix_data 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 "ebrukilic/paligemma2_mix_data" \
    --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": "ebrukilic/paligemma2_mix_data",
		"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 "ebrukilic/paligemma2_mix_data" \
        --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": "ebrukilic/paligemma2_mix_data",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ebrukilic/paligemma2_mix_data with Docker Model Runner:
```
docker model run hf.co/ebrukilic/paligemma2_mix_data
```

paligemma2_mix_data

This model is a fine-tuned version of google/paligemma2-3b-pt-448 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.6083

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-05
train_batch_size: 1
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 4
optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
0.9278	0.0372	200	0.9142
0.7324	0.0745	400	0.7881
0.5464	0.1117	600	0.7628
0.6358	0.1489	800	0.7154
0.6393	0.1862	1000	0.6956
0.5533	0.2234	1200	0.6999
0.5798	0.2606	1400	0.6727
0.5291	0.2979	1600	0.6690
0.5927	0.3351	1800	0.6548
0.6512	0.3723	2000	0.6542
0.6157	0.4096	2200	0.6464
0.6177	0.4468	2400	0.6379
0.5941	0.4840	2600	0.6345
0.5513	0.5213	2800	0.6423
0.6359	0.5585	3000	0.6323
0.5513	0.5957	3200	0.6322
0.4695	0.6330	3400	0.6200
0.5851	0.6702	3600	0.6127
0.5475	0.7074	3800	0.6238
0.5264	0.7447	4000	0.6165
0.5325	0.7819	4200	0.6160
0.5497	0.8191	4400	0.6056
0.5338	0.8564	4600	0.6096
0.5604	0.8936	4800	0.6184
0.527	0.9308	5000	0.6015
0.4486	0.9681	5200	0.6078
0.5502	1.0052	5400	0.5992
0.399	1.0424	5600	0.6141
0.4539	1.0797	5800	0.6069
0.4584	1.1169	6000	0.6120
0.4254	1.1541	6200	0.6083

Framework versions

PEFT 0.18.0
Transformers 4.57.3
Pytorch 2.9.0+cu126
Datasets 4.4.2
Tokenizers 0.22.1

Downloads last month: -

Model tree for ebrukilic/paligemma2_mix_data

Base model

google/paligemma2-3b-pt-448

Adapter

(63)

this model