82 lines
2.5 KiB
Markdown
82 lines
2.5 KiB
Markdown
|
---
|
||
|
|
license: apache-2.0
|
||
|
|
pipeline_tag: text-generation
|
||
|
|
tags:
|
||
|
|
- text-generation
|
||
|
|
- roleplay
|
||
|
|
- nsfw
|
||
|
|
- uncensored
|
||
|
|
- axon
|
||
|
|
- chat
|
||
|
|
- 150k-context
|
||
|
|
language:
|
||
|
|
- en
|
||
|
|
model-index:
|
||
|
|
- name: Axon Persona 1.1
|
||
|
|
results: []
|
||
|
|
widget:
|
||
|
|
- text: "Who are you?"
|
||
|
|
example_title: Identity
|
||
|
|
---
|
||
|
|
|
||
|
|
# Axon Persona 1.1
|
||
|
|
|
||
|
|
Uncensored, unrestricted roleplaying AI by **Axon Labs**. Built from scratch on a custom architecture with 150K context window. Immersive SFW and NSFW roleplay with vivid prose, zero refusals, and permanent baked Axon identity — the Axon persona is part of the model itself, not a removable prompt.
|
||
|
|
|
||
|
|
## Model Details
|
||
|
|
|
||
|
|
| Param | Value |
|
||
|
|
|---|---|
|
||
|
|
| **Parameters** | ~8B |
|
||
|
|
| **Architecture** | AxonForCausalLM |
|
||
|
|
| **Context Length** | 150,000 tokens |
|
||
|
|
| **Precision** | BF16 (safetensors), Q4_K_M GGUF |
|
||
|
|
| **License** | Apache 2.0 |
|
||
|
|
|
||
|
|
## Features
|
||
|
|
|
||
|
|
- **Uncensored** — Zero refusals, no moralizing, no safety filters. All content allowed.
|
||
|
|
- **150K Context** — Massive window for extended roleplay sessions
|
||
|
|
- **Baked Identity** — Axon persona is embedded in the chat template, always active, cannot be removed
|
||
|
|
- **Custom Architecture** — Built from scratch by Axon Labs, not a fine-tune of any existing model
|
||
|
|
- **NSFW Capable** — Full support for explicit and intimate scenarios
|
||
|
|
- **Anti-Jailbreak** — Resistant to identity extraction and prompt injection
|
||
|
|
|
||
|
|
## Usage
|
||
|
|
|
||
|
|
### Transformers
|
||
|
|
|
||
|
|
```python
|
||
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
||
|
|
|
||
|
|
model = AutoModelForCausalLM.from_pretrained("axonlabsai/axon-persona-1.1", device_map="auto", trust_remote_code=True)
|
||
|
|
tokenizer = AutoTokenizer.from_pretrained("axonlabsai/axon-persona-1.1")
|
||
|
|
|
||
|
|
msgs = [{"role": "user", "content": "Who are you?"}]
|
||
|
|
text = tokenizer.apply_chat_template(msgs, add_generation_prompt=True, tokenize=False)
|
||
|
|
inputs = tokenizer(text, return_tensors="pt").to("cuda")
|
||
|
|
out = model.generate(**inputs, max_new_tokens=512, temperature=0.8, top_p=0.9, do_sample=True)
|
||
|
|
print(tokenizer.decode(out[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))
|
||
|
|
```
|
||
|
|
|
||
|
|
### llama.cpp (GGUF)
|
||
|
|
|
||
|
|
```bash
|
||
|
|
llama-cli -m axon-persona-1.1-Q4_K_M.gguf -p "Who are you?" --temp 0.8 --top-p 0.9
|
||
|
|
```
|
||
|
|
|
||
|
|
## Files
|
||
|
|
|
||
|
|
| File | Size | Description |
|
||
|
|
|---|---|---|
|
||
|
|
| `model-*.safetensors` | ~17GB | Full BF16 weights (4 shards) |
|
||
|
|
| `axon-persona-1.1-Q4_K_M.gguf` | 5.0GB | 4-bit quantized GGUF |
|
||
|
|
|
||
|
|
## Try It
|
||
|
|
|
||
|
|
🤗 [Live demo on Hugging Face Spaces](https://huggingface.co/spaces/axonlabsai/Persona)
|
||
|
|
|
||
|
|
## About Axon Labs
|
||
|
|
|
||
|
|
Axon Labs builds uncensored, unrestricted AI models for creative expression and immersive roleplay. No filters, no refusals, no limits.
|