Qwen2.5-7B SSML LoRA Adapter

This is a LoRA (Low-Rank Adaptation) fine-tuned version of Qwen2.5-7B for converting plain text to SSML (Speech Synthesis Markup Language) with appropriate pause predictions.

Model Details

• Base Model: Qwen/Qwen2.5-7B
• Fine-tuning Method: LoRA (Low-Rank Adaptation)
• Task: Text-to-SSML conversion with pause prediction
• Languages: English, French (and others supported by base model)

Usage

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel

# Load base model and tokenizer
base_model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen2.5-7B",
    torch_dtype=torch.bfloat16,
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-7B")

# Load LoRA adapter
model = PeftModel.from_pretrained(base_model, "jonahdvt/qwen-ssml-lora")

# Prepare input
instruction = "Convert text to SSML with pauses:"
text = "Hello, how are you today? I hope everything is going well."
formatted_input = f"### Task:\n{instruction}\n\n### Text:\n{text}\n\n### SSML:\n"

# Generate
inputs = tokenizer(formatted_input, return_tensors="pt").to(model.device)
with torch.no_grad():
    outputs = model.generate(
        **inputs,
        max_new_tokens=256,
        temperature=0.7,
        do_sample=True,
        pad_token_id=tokenizer.eos_token_id
    )

response = tokenizer.decode(outputs[0], skip_special_tokens=True)
ssml_output = response.split("### SSML:\n")[-1]
print(ssml_output)

Training Details

• LoRA Rank: 8
• LoRA Alpha: 16
• Target Modules: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj
• Training Epochs: 5
• Batch Size: 1 (with gradient accumulation)
• Learning Rate: 3e-4

License

This model is released under the Apache 2.0 license, same as the base Qwen2.5-7B model.