|
|
--- |
|
|
language: en |
|
|
tags: |
|
|
- conversational |
|
|
- roleplay |
|
|
- dippy |
|
|
- dialogpt |
|
|
- bittensor |
|
|
license: mit |
|
|
--- |
|
|
|
|
|
# Dippy DialoGPT Optimized |
|
|
|
|
|
This is a fine-tuned version of microsoft/DialoGPT-medium optimized for conversational AI with Dippy personality. |
|
|
|
|
|
## Model Details |
|
|
- Base model: microsoft/DialoGPT-medium |
|
|
- Fine-tuned for: Conversational AI, roleplay, helpful assistant interactions |
|
|
- Optimized for: Bittensor SN11 Dippy subnet |
|
|
|
|
|
## Usage |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("Gatescrispy/dippy-dialogpt-optimized") |
|
|
model = AutoModelForCausalLM.from_pretrained("Gatescrispy/dippy-dialogpt-optimized") |
|
|
|
|
|
# Generate response |
|
|
inputs = tokenizer.encode("Hello! How are you today?", return_tensors="pt") |
|
|
outputs = model.generate(inputs, max_length=50, pad_token_id=tokenizer.eos_token_id) |
|
|
response = tokenizer.decode(outputs[0], skip_special_tokens=True) |
|
|
print(response) |
|
|
``` |
|
|
|
|
|
## Training |
|
|
- Dataset: Custom Dippy personality conversations |
|
|
- Training: 1 epoch with learning rate scheduling |
|
|
- Hardware: NVIDIA RTX 3090 |
|
|
|
|
|
## Bittensor Integration |
|
|
This model is designed for Bittensor SN11 Dippy subnet integration. |
|
|
|
