Instructions to use ecorbari/Gemma-2b-it-Psych-Merged with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ecorbari/Gemma-2b-it-Psych-Merged with Transformers:

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

pipe = pipeline("text-generation", model="ecorbari/Gemma-2b-it-Psych-Merged")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ecorbari/Gemma-2b-it-Psych-Merged")
model = AutoModelForCausalLM.from_pretrained("ecorbari/Gemma-2b-it-Psych-Merged")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use ecorbari/Gemma-2b-it-Psych-Merged with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ecorbari/Gemma-2b-it-Psych-Merged"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ecorbari/Gemma-2b-it-Psych-Merged",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ecorbari/Gemma-2b-it-Psych-Merged

SGLang

How to use ecorbari/Gemma-2b-it-Psych-Merged 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 "ecorbari/Gemma-2b-it-Psych-Merged" \
    --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": "ecorbari/Gemma-2b-it-Psych-Merged",
		"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 "ecorbari/Gemma-2b-it-Psych-Merged" \
        --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": "ecorbari/Gemma-2b-it-Psych-Merged",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ecorbari/Gemma-2b-it-Psych-Merged with Docker Model Runner:
```
docker model run hf.co/ecorbari/Gemma-2b-it-Psych-Merged
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Model Card for Gemma-2b-it-Psych-Merged

Model Summary

Gemma-2b-it-Psych-Merged is the full-weight, standalone version of the google/gemma-2b-it model, domain-adapted for psychological contexts. This model integrates the LoRA adapter weights from ecorbari/Gemma-2b-it-Psych directly into the base model using the merge_and_unload() process.

The model is optimized to generate empathetic, supportive, and professionally aligned psychological responses. Unlike the adapter-only version, this repository contains the complete merged weights, meaning it does not require the peft library for standard inference.

Model Details

Model Description

Author: Ederson Corbari (e@NeuroQuest.ai)
Date: February 01, 2026
Model type: Causal Language Model (LLM)
Language(s): English
License: MIT (consistent with Gemma base model terms)
Finetuned from model: google/gemma-2b-it
Merge method: LoRA Weights Integration (merge_and_unload)

Model Sources

Hugging Face Merged: ecorbari/Gemma-2b-it-Psych-Merged
Hugging Face LoRA Adapter: ecorbari/Gemma-2b-it-Psych
Base Model: google/gemma-2b-it

Uses

Direct Use

This merged model is ready for production and simplified inference. It can be loaded directly using standard transformers pipelines. It is intended for:

Research on empathetic AI behavior.
Educational demonstrations of domain-adapted LLMs.
Proof-of-concept psychological support tools.

Out-of-Scope Use

Clinical Use: This model is NOT a substitute for licensed mental health professionals. It must not be used for diagnosis or treatment.
High-Stakes Decision Making: It should not be used in autonomous counseling systems without human oversight.

How to Get Started with the Model

Since the weights are already merged, you can run inference using a simple pipeline:

import torch
from transformers import pipeline

model_id = "ecorbari/Gemma-2b-it-Psych-Merged"

pipe = pipeline(
    "text-generation",
    model=model_id,
    dtype=torch.float16,
    device_map="auto",
)

prompt = "I feel anxious and overwhelmed lately. What should I do?"
result = pipe(prompt, max_new_tokens=200)

print(result[0]["generated_text"])

Bias, Risks, and Limitations

Safety Disclaimer: The model may generate inaccurate information. Empathy in text generation does not imply clinical safety or medical correctness.

Data Bias: Responses may reflect biases inherent in the jkhedri/psychology-dataset.

Human Oversight: Users should apply human judgment, especially in sensitive conversational settings.

Training and Merge Process

The workflow involved loading the google/gemma-2b-it model in float16 precision, attaching the LoRA adapters trained on preference-based psychological data, and merging the weights into a single model for downstream use. This ensures compatibility with environments that do not support PEFT or require lower latency for inference.

Downloads last month: 6

Safetensors

Model size

3B params

Tensor type

F16

Model tree for ecorbari/Gemma-2b-it-Psych-Merged

Base model

google/gemma-2b-it

Finetuned

(118)

this model

Quantizations

1 model

ecorbari
/

Gemma-2b-it-Psych-Merged