46 lines
1.2 KiB
Markdown
46 lines
1.2 KiB
Markdown
---
|
|
tags:
|
|
- qwen
|
|
- tamil
|
|
- therapy
|
|
- mental-health
|
|
- conversational
|
|
library_name: transformers
|
|
base_model: Qwen/Qwen2.5-3B-Instruct
|
|
language:
|
|
- ta
|
|
---
|
|
|
|
# Qwen 2.5-3B Tamil Therapy Model
|
|
|
|
Fine-tuned Qwen 2.5-3B model for empathetic Tamil therapy conversations.
|
|
|
|
## Model Details
|
|
- **Base Model**: Qwen/Qwen2.5-3B-Instruct
|
|
- **Language**: Tamil
|
|
- **Size**: ~6GB
|
|
- **Use Case**: Mental health support chatbot in Tamil
|
|
|
|
## Usage
|
|
```python
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
|
import torch
|
|
|
|
model = AutoModelForCausalLM.from_pretrained(
|
|
"Pavishanth68/qwen-2.5-3b-tamil-therapy-merged",
|
|
torch_dtype=torch.float16,
|
|
device_map="auto"
|
|
)
|
|
tokenizer = AutoTokenizer.from_pretrained("Pavishanth68/qwen-2.5-3b-tamil-therapy-merged")
|
|
|
|
prompt = "System: You are an empathetic Tamil therapist.\n\nUser: எனக்கு பதட்டமாக இருக்கிறது\nTherapist:"
|
|
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
|
|
|
|
outputs = model.generate(**inputs, max_new_tokens=150, temperature=0.7)
|
|
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
|
print(response)
|
|
```
|
|
|
|
## Disclaimer
|
|
This is an AI assistant, not a replacement for professional mental health care.
|