Instructions to use Ellio98/doctor-gemma-2-2b-it with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Ellio98/doctor-gemma-2-2b-it with Transformers:

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

pipe = pipeline("text-generation", model="Ellio98/doctor-gemma-2-2b-it")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Ellio98/doctor-gemma-2-2b-it")
model = AutoModelForCausalLM.from_pretrained("Ellio98/doctor-gemma-2-2b-it")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Ellio98/doctor-gemma-2-2b-it with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Ellio98/doctor-gemma-2-2b-it"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ellio98/doctor-gemma-2-2b-it",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Ellio98/doctor-gemma-2-2b-it

SGLang

How to use Ellio98/doctor-gemma-2-2b-it 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 "Ellio98/doctor-gemma-2-2b-it" \
    --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": "Ellio98/doctor-gemma-2-2b-it",
		"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 "Ellio98/doctor-gemma-2-2b-it" \
        --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": "Ellio98/doctor-gemma-2-2b-it",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Ellio98/doctor-gemma-2-2b-it with Docker Model Runner:
```
docker model run hf.co/Ellio98/doctor-gemma-2-2b-it
```

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Model Card for Model ID

Model Details

Model Description

Uses

Below we share some code snippets on how to get quickly started with running the model. First, install the Transformers library with:

pip install -U transformers
pip install -U torch

Then, copy the snippet from the section that is relevant for your usecase.

Direct Use

Running the model on Single/Multi GPU

# pip install accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

tokenizer = AutoTokenizer.from_pretrained("Ellio98/doctor-gemma-2-2b-it")
model = AutoModelForCausalLM.from_pretrained(
    "Ellio98/doctor-gemma-2-2b-it",
    device_map="auto",
    torch_dtype=torch.bfloat16,
)

query = "I have a high fever and coughing since morning. What should I do ?"

prompt = [{"role": "user", "content": query}]

model_input = tokenizer(tokenizer.apply_chat_template(prompt, tokenize=False), return_tensors="pt")

outputs = model.generate(
                          input_ids=model_input["input_ids"].to("cuda"),
                          attention_mask=model_input["attention_mask"].to("cuda"),
                          max_new_tokens=32
)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))
# user
# I have a high fever and coughing since morning. What should I do ?
# model
# Hi,From history it seems that you might be having viral infection giving this problem of fever with cough.
# Take paracetamol or ibuprofen for pain due to fever. Take plenty of water.
# If require take one antispasmodic medicine like Meftal spas as needed.
# Ok and take care.

Bias, Risks, and Limitations

The model might not be good at generating the desired output and might suggest some weird medical practices that does not make sense. It is highly recommended to use this model just for research purpose and not as a commercial use. Also, as no data anonymization was performed, the model will be generating random names, addresses or terms that are not intended.

Future Works

Need to perform data anonymization to refine the model output to be generic.

Training Details

Training Data

For fine-tuning the model, 5000 samples were indiced from the sid6i7/patient-doctor dataset, out of which 4500 were used for training, 250 were used for validation and testing.

Training Hyperparameters

Training regime: [More Information Needed]
precision: bfloat16
alpha: 32
rank: 16
lora dropout: 0.05
target modules: 'o_proj', 'v_proj', 'up_proj', 'q_proj', 'down_proj', 'gate_proj', and 'k_proj'
optimizer: paged_adamw_32bit
learning rate: 2e-4

Results

The loss progression during fine-tuning of gemma-2-2b on the dataset is as follows:

Hardware

The model was trained on 2 x T4 GPU. The total training time taken was 6 hours.

Downloads last month: -

Safetensors

Model size

3B params

Tensor type

F32

F16

Model tree for Ellio98/doctor-gemma-2-2b-it

Base model

google/gemma-2-2b

Finetuned

google/gemma-2-2b-it

Quantized

(177)

this model

Ellio98
/

doctor-gemma-2-2b-it