Update README.md

README.md

@@ -2,34 +2,21 @@
 license: apache-2.0
 language:
 - tg
-base_model: muhtasham/orpheus-tj
 pipeline_tag: text-to-speech
 tags:
 - mlx
 ---
 
-# muhtasham/orpheus-tj-mlx-fp16
+# re-skill/orpheus-tj-early-mlx-fp16
 
-The Model [muhtasham/orpheus-tj-mlx-fp16](https://huggingface.co/muhtasham/orpheus-tj-mlx-fp16) was converted to MLX format from [muhtasham/orpheus-tj](https://huggingface.co/muhtasham/orpheus-tj) using mlx-lm version **0.22.3**.
+The Model [re-skill/orpheus-tj-early-mlx-fp16](https://huggingface.co/re-skill/orpheus-tj-early-mlx-fp16) was converted to MLX format from [re-skill/orpheus-tj-early](https://huggingface.co/re-skill/orpheus-tj-early) using mlx-lm version **0.22.3**.
 
 ## Use with mlx
 
 ```bash
-pip install mlx-lm
+pip install mlx-audio
 ```
 
-```python
-from mlx_lm import load, generate
-
-model, tokenizer = load("muhtasham/orpheus-tj-mlx-fp16")
-
-prompt="hello"
-
-if hasattr(tokenizer, "apply_chat_template") and tokenizer.chat_template is not None:
-    messages = [{"role": "user", "content": prompt}]
-    prompt = tokenizer.apply_chat_template(
-        messages, tokenize=False, add_generation_prompt=True
-    )
-
-response = generate(model, tokenizer, prompt=prompt, verbose=True)
+```bash
+mlx_audio.tts.generate --model re-skill/orpheus-tj-early-mlx-fp16 --max_tokens 1200 --temperature 0.0 --top_p 0.95 --repetition_penalty 1.1 --text "салом ин матн пурра бо компютер сохта шудааст ман қурбони собир нестам."
 ```