53 lines
2.0 KiB
Markdown
53 lines
2.0 KiB
Markdown
---
|
|
language: en
|
|
tags:
|
|
- text-generation
|
|
- transformer
|
|
- mistral
|
|
- fine-tuned
|
|
- uncensored
|
|
- nsfw
|
|
license: apache-2.0
|
|
datasets:
|
|
- open-source-texts
|
|
model-name: Fine-tuned Mistral 7B (Uncensored)
|
|
---
|
|
|
|
# Fine-tuned Mistral 7B (Uncensored)
|
|
|
|
## Model Description
|
|
|
|
This model is a fine-tuned version of the **Mistral 7B**, a dense transformer model, trained on 40,000 datapoints of textual data from a variety of open-source sources. The base model, Mistral 7B, is known for its high efficiency in processing text and generating meaningful, coherent responses.
|
|
|
|
This fine-tuned version has been optimized for tasks involving natural language understanding, generation, and conversation-based interactions. Importantly, this model is **uncensored**, which means it does not filter or restrict content, allowing it to engage in more "spicy" or NSFW conversations.
|
|
|
|
## Fine-tuning Process
|
|
|
|
- **Data**: The model was fine-tuned using a dataset of 40,000 textual datapoints sourced from various open-source repositories.
|
|
- **Training Environment**: Fine-tuning was conducted on two NVIDIA A100 GPUs.
|
|
- **Training Time**: The training process took approximately 16 hours.
|
|
- **Optimizer**: The model was trained using AdamW optimizer with a learning rate of `5e-5`.
|
|
|
|
## Intended Use
|
|
|
|
This fine-tuned model is intended for the following tasks:
|
|
- Text generation
|
|
- Question answering
|
|
- Dialogue systems
|
|
- Content generation for AI-powered interactions, including NSFW or adult-oriented conversations.
|
|
|
|
### How to Use
|
|
|
|
You can easily load and use this model with the `transformers` library in Python:
|
|
|
|
```python
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("your-organization/finetuned-mistral-7b")
|
|
model = AutoModelForCausalLM.from_pretrained("your-organization/finetuned-mistral-7b")
|
|
|
|
inputs = tokenizer("Input your text here.", return_tensors="pt")
|
|
outputs = model.generate(inputs["input_ids"], max_length=50, num_return_sequences=1)
|
|
|
|
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
|