--- library_name: transformers tags: - voice-cloning - audio - unsloth - orpheus - snac - hypaai - lora - merged license: apache-2.0 datasets: - hypaai/Hypa_Fleurs - MrDragonFox/Elise - canopylabs/zac-sample-dataset - google/fleurs language: - ig - yo - ha - en base_model: - canopylabs/orpheus-3b-0.1-ft pipeline_tag: text-to-speech --- # Hypa_Orpheus-3b-0.1-ft (merged 16-bit) A 16-bit quantized and merged, memory-efficient fine-tuned version of [`canopylabs/orpheus-3b-0.1-ft`](https://huggingface.co/canopylabs/orpheus-3b-0.1-ft), optimized with Unsloth and LoRA, for expressive multilingual TTS, especially in low-resource African languages. This model provides the following capabilities: * Text-to-Speech generation * Speech synthesis for under-represented accents * Voice cloning & emotion synthesis * Research on multilingual low-resource voice AI --- ## Model Details ### Model Summary This model was trained on a parallel text-speech dataset totaling over 300 hours (75k samples) of Nigerian-accented and low-resource language audio (Igbo, Yoruba, Hausa). A key part of the dataset comes from AfroVoices' transcription of real-world YouTube data (denoted as Random speaker, ~100+ hrs). To preserve and enhance multilingual capabilities while avoiding catastrophic forgetting, we included synthetic speech-text data sampled from the original 8 Orpheus voices using the default emotional prompts. The final training set also included new speakers like: * Eniola (40 hrs) – Female, bold, clear * Moyo (40 hrs) – Female, professional, articulate * Lovelyn (35 hrs) – Female, warm, shy * Precious (30 hrs) – Female, friendly, gentle This model sets state-of-the-art performance on low-resource Multilingual TTS tasks across African languages (see training stats below) ### Base Model Details The default Orpheus-TTS model released by [`Canopy Labs`](https://canopylabs.ai/releases/orpheus_can_speak_any_language) supports the below voices and emotions: Voices: `tara`, `leah`, `jess`, `leo`, `dan`, `mia`, `zac`, and `zoe`. Emotions: ``, ``, ``, ``, ``, ``, ``, and ``. Through synthetic data generation and addition, our finetuned model also maintains these voices and emotions. Please visit the default model's card for mor info on voices and emotions. ### Our Model Sample Generations 🎧 Listen to samples generated by Hypa Orpheus-TTS

Text Input	Audio Output	Language	Voice
I am cooking for guests tomorrow and need to know how to make aioli. Can you give me a step-by-step recipe.		English	Emmanuel
Ina dafa abinci don bakin gobe kuma ina bukatar sanin yadda ake yin ailoli. Za ka iya ba ni girke-gireken matakan daya bayan daya?		Hausa	Emmanuel
Ina dafa abinci don bakin gobe kuma ina bukatar sanin yadda ake yin ailoli. Za ka iya ba ni girke-gireken matakan daya bayan daya?		Hausa	Eniola
Èmi máa se oúnjẹ fún àwọn àlejò l'ọ́la mo sì nílò láti mọ bí wọn ti ńṣe aioli. Ṣe o lè fún mi ni àwọn ìlànà ìdáná ẹlẹ́sẹẹsẹ?		Yoruba	Eniola
I am cooking for guests tomorrow and need to know how to make aioli. Can you give me a step-by-step recipe.		English	Eniola
M na-esi nri maka ndị ọbịa echi ma achọ ịmata otú esi esi aioli. Ị nwere ike inye m usoro ntụziaka?		Igbo	Eniola
M na-esi nri maka ndị ọbịa echi ma achọ ịmata otú esi esi aioli. Ị nwere ike inye m usoro ntụziaka?		Igbo	Lovelyn
I am cooking for guests tomorrow and need to know how to make aioli. Can you give me a step-by-step recipe.		English	Lovelyn

--- ## Training Details ### Training Summary * Base model: canopylabs/orpheus-3b-0.1-ft * Training engine: Unsloth + LoRA * LoRA config: r=1024, alpha=1024, dropout=0.0, full attention + FFN adaptation * Quantization: 4-bit (bnb) for training; final model is highly memory-efficient * Total steps: 18,014 (1 epoch) * Batch size: 1 × 4 (grad accum) * GPU: A100 40GB (max 55% VRAM used)

Step	Training Loss	Validation Loss
5,000	3.9496	3.8790
10,000	3.8863	3.79497
15,000	3.8544	3.75323

### Dataset Summary * Sources: * ✅ Manually aligned YouTube transcriptions (aka Random) * ✅ Synthetic voice generation from Orpheus TTS * ✅ Parallel text-audio pairs for African-English, Igbo, Yoruba, Hausa * Total Hours: 300+ (multi-accent) * Key Speakers: 45+ unique voices (see speaker distribution chart below) ![image/png](https://huggingface.co/hypaai/Hypa_Orpheus-3b-0.1-ft-unsloth-bnb-4bit/resolve/main/assets/stats.png) We plan to open-source the full dataset shortly similar to the [Hypa_Fleurs](https://huggingface.co/datasets/hypaai/Hypa_Fleurs) initiative. --- ## Licensing and Citation This model is released under an [Open Source License](./LICENSE) (apache-2.0). Please refer to the LICENSE file for full details. When using this model in your work, please cite both this model as well as the base [`canopylabs/orpheus-3b-0.1-ft`](https://huggingface.co/canopylabs/orpheus-3b-0.1-ft) model as follows: ```bibtex @misc{canopylabsorpheus, title={Orpheus-3b-0.1-ft: A Multilingual Text-to-Speech Model}, author={Canopy Labs}, year={2025}, publisher={Hugging Face}, howpublished={\url{https://huggingface.co/canopylabs/orpheus-3b-0.1-ft}}, note={Fine-tuned version of Orpheus for expressive TTS} } @misc{hypaorpheus4bit, title={Hypa_Orpheus-3b-0.1-ft (LoRA-4bit)}, author={Hypa AI}, year={2025}, note={Fine-tuned Orpheus TTS on African languages}, url={https://huggingface.co/hypaai/Hypa_Orpheus-3b-0.1-ft-unsloth-bnb-4bit} } ``` --- ## Acknowledgements - **Canopy Labs Team:** For creating the foundational model and opensourcing it. - **AfroVoices Experts:** For their translation expertise and high-quality datasets. - **Community Support:** We thank all supporters, contributors, and users. --- ## Contact and Contributions For any questions, issues, or contributions, please open an issue in this repository or contact [hypa.ai.ng@gmail.com](mailto:hypa.ai.ng@gmail.com). Contributions are welcome! --- ## Closing Remarks By making Hypa_Orpheus available, we hope to empower research and development in multilingual speech technologies for African languages. Hypa AI remains steadfast in its mission to pioneer intelligent solutions that are not just technologically advanced but are also culturally aware, ensuring that the future of AI is as diverse and inclusive as the world it serves. AfroVoices, a subsidiary of Hypa AI, is dedicated to amplifying African voices, languages, and cultures in the intelligence age. Focused on bridging the digital representation gap, AfroVoices curates datasets and resources for African languages, promoting inclusivity and cultural appreciation in AI technologies. Their mission goes beyond technological innovation, aiming to celebrate the richness of African linguistic diversity on a global stage. --- ## Usage ### Unsloth Inference Download the nedded packages. ```python %%capture import os if "COLAB_" not in "".join(os.environ.keys()): !pip install unsloth else: # Do this only in Colab notebooks! Otherwise use pip install unsloth !pip install --no-deps bitsandbytes accelerate xformers==0.0.29.post3 peft trl==0.15.2 triton cut_cross_entropy unsloth_zoo !pip install sentencepiece protobuf datasets huggingface_hub hf_transfer !pip install --no-deps unsloth !pip install snac ``` Download the models (both the SNAC encoder/decoder as well as our finetuned Hypa_Orpheus). ```python import torch from snac import SNAC from unsloth import FastLanguageModel dtype = None # None for auto detection. Float16 for Tesla T4, V100, Bfloat16 for Ampere+ load_in_4bit = True # Use 4bit quantization to reduce memory usage. Can be False. model, tokenizer = FastLanguageModel.from_pretrained( model_name = "hypaai/Hypa_Orpheus-3b-0.1-ft-unsloth-merged_16bit", max_seq_length= 2048, # Choose any for long context! dtype = dtype, load_in_4bit = load_in_4bit, #token = "hf_...", # use one if using gated models like meta-llama/Llama-2-7b-hf ) snac_model = SNAC.from_pretrained("hubertsiuzdak/snac_24khz") snac_model = snac_model.to("cuda") ``` Create your text prompt, select the voice, and pass through the Model. ```python prompts = [ """Mo nífẹ̀ẹ́sí láti ṣe Ph.D sùgbọ́n mi ò ì tíì pinnu ẹ̀ka tí màá ṣe. Àwọn anfaani tí óń dé oríṣiríṣi àwọn olùgbọ́ káàkiri àgbáyé wo ni mo ní""", ] chosen_voice = "Eniola" # None for single-speaker FastLanguageModel.for_inference(model) # Enable native 2x faster inference snac_model.to("cpu")# Moving snac_model cuda to cpu prompts_ = [(f"{chosen_voice}: " + p) if chosen_voice else p for p in prompts] all_input_ids = [] for prompt in prompts_: input_ids = tokenizer(prompt, return_tensors="pt").input_ids all_input_ids.append(input_ids) start_token = torch.tensor([[ 128259]], dtype=torch.int64) # Start of human end_tokens = torch.tensor([[128009, 128260]], dtype=torch.int64) # End of text, End of human all_modified_input_ids = [] for input_ids in all_input_ids: modified_input_ids = torch.cat([start_token, input_ids, end_tokens], dim=1) # SOH SOT Text EOT EOH all_modified_input_ids.append(modified_input_ids) all_padded_tensors = [] all_attention_masks = [] max_length = max([modified_input_ids.shape[1] for modified_input_ids in all_modified_input_ids]) for modified_input_ids in all_modified_input_ids: padding = max_length - modified_input_ids.shape[1] padded_tensor = torch.cat([torch.full((1, padding), 128263, dtype=torch.int64), modified_input_ids], dim=1) attention_mask = torch.cat([torch.zeros((1, padding), dtype=torch.int64), torch.ones((1, modified_input_ids.shape[1]), dtype=torch.int64)], dim=1) all_padded_tensors.append(padded_tensor) all_attention_masks.append(attention_mask) all_padded_tensors = torch.cat(all_padded_tensors, dim=0) all_attention_masks = torch.cat(all_attention_masks, dim=0) input_ids = all_padded_tensors.to("cuda") attention_mask = all_attention_masks.to("cuda") generated_ids = model.generate( input_ids=input_ids, attention_mask=attention_mask, max_new_tokens=1200, do_sample=True, temperature=0.6, top_p=0.95, repetition_penalty=1.1, num_return_sequences=1, eos_token_id=128258, use_cache = True ) token_to_find = 128257 token_to_remove = 128258 token_indices = (generated_ids == token_to_find).nonzero(as_tuple=True) if len(token_indices[1]) > 0: last_occurrence_idx = token_indices[1][-1].item() cropped_tensor = generated_ids[:, last_occurrence_idx+1:] else: cropped_tensor = generated_ids mask = cropped_tensor != token_to_remove processed_rows = [] for row in cropped_tensor: masked_row = row[row != token_to_remove] processed_rows.append(masked_row) code_lists = [] for row in processed_rows: row_length = row.size(0) new_length = (row_length // 7) * 7 trimmed_row = row[:new_length] trimmed_row = [t - 128266 for t in trimmed_row] code_lists.append(trimmed_row) def redistribute_codes(code_list): layer_1 = [] layer_2 = [] layer_3 = [] for i in range((len(code_list)+1)//7): layer_1.append(code_list[7*i]) layer_2.append(code_list[7*i+1]-4096) layer_3.append(code_list[7*i+2]-(2*4096)) layer_3.append(code_list[7*i+3]-(3*4096)) layer_2.append(code_list[7*i+4]-(4*4096)) layer_3.append(code_list[7*i+5]-(5*4096)) layer_3.append(code_list[7*i+6]-(6*4096)) codes = [torch.tensor(layer_1).unsqueeze(0), torch.tensor(layer_2).unsqueeze(0), torch.tensor(layer_3).unsqueeze(0)] # codes = [c.to("cuda") for c in codes] audio_hat = snac_model.decode(codes) return audio_hat my_samples = [] for code_list in code_lists: samples = redistribute_codes(code_list) my_samples.append(samples) from IPython.display import display, Audio if len(prompts) != len(my_samples): raise Exception("Number of prompts and samples do not match") else: for i in range(len(my_samples)): print(prompts[i]) samples = my_samples[i] display(Audio(samples.detach().squeeze().to("cpu").numpy(), rate=24000)) # Clean up to save RAM del my_samples,samples ``` ### Standard Inference Download the nedded packages. ```python %%capture !pip install snac ipywebrtc ``` Download the Models (SNAC & Hypa_Orpheus) ```python import torch from transformers import AutoModelForCausalLM, Trainer, TrainingArguments, AutoTokenizer from snac import SNAC # Loads the pre-trained SNAC model and moves it to the CPU. snac_model = SNAC.from_pretrained("hubertsiuzdak/snac_24khz") snac_model = snac_model #.to("cpu") print("We have loaded the Encoder/Decoder model to the cpu, to use vram - use the gpu for faster inference") # Loading the Orpheus Model and Tokenizer, moving the model to the GPU for faster inference model_name = "hypaai/Hypa_Orpheus-3b-0.1-ft-unsloth-merged_16bit" model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16) model.cuda() tokenizer = AutoTokenizer.from_pretrained(model_name) ``` Create Prompt(s) and Select Voice & Emotions as needed. ```python # List of supported voices in Orpheus-TTS voices = [ "Eniola", "tara", # Female, conversational, clear "Moyo", "leah", # Female, warm, gentle "Gift", "jess", # Female, energetic, youthful "Prince", "leo", # Male, authoritative, deep "Emmanuel", "dan", # Male, friendly, casual "Cynthia", "mia", # Female, professional, articulate "Kolade", "zac", # Male, enthusiastic, dynamic "Lovelyn", "zoe" # Female, calm, soothing ] # List of supported emotion tags in Orpheus-TTS emotions = [ "", # Laughter "", # Soft chuckle "", # Sighing "", # Coughing "", # Sniffling "", # Groaning "", # Yawning "" # Gasping ] # Creating Prompts prompts = [ "Hey there my name is Eniola 9000, and I'm a speech generation model that can sound like a person.", # "I've also been taught to understand and produce paralinguistic things like sighing, or chuckling, or yawning!", # "I live in San Francisco, and have, uhm let's see, 3 billion 7 hundred ... well, lets just say a lot of parameters.", ] chosen_voice = "Eniola" # "tara" # see github for other voices prompts = [f"{chosen_voice}: " + p for p in prompts] # Creating the prompts (as a batch) print(prompts) ``` Tokenize prompt(s) into inputIDs, pad, and create attention masks. ```python # Tokenizing each prompt into input IDs. all_input_ids = [] for prompt in prompts: input_ids = tokenizer(prompt, return_tensors="pt").input_ids all_input_ids.append(input_ids) # Adds special tokens to mark the beginning and end of each prompt start_token = torch.tensor([[128259]], dtype=torch.int64) # Start of human end_tokens = torch.tensor([[128009, 128260]], dtype=torch.int64) # End of text, End of human all_modified_input_ids = [] for input_ids in all_input_ids: modified_input_ids = torch.cat([start_token, input_ids, end_tokens], dim=1) # SOH SOT Text EOT EOH all_modified_input_ids.append(modified_input_ids) # Padding All sequences to same length and creating corresponding attention masks all_padded_tensors = [] all_attention_masks = [] max_length = max([modified_input_ids.shape[1] for modified_input_ids in all_modified_input_ids]) for modified_input_ids in all_modified_input_ids: padding = max_length - modified_input_ids.shape[1] # Left Padding padded_tensor = torch.cat([torch.full((1, padding), 128263, dtype=torch.int64), modified_input_ids], dim=1) attention_mask = torch.cat([torch.zeros((1, padding), dtype=torch.int64), torch.ones((1, modified_input_ids.shape[1]), dtype=torch.int64)], dim=1) all_padded_tensors.append(padded_tensor) all_attention_masks.append(attention_mask) all_padded_tensors = torch.cat(all_padded_tensors, dim=0) all_attention_masks = torch.cat(all_attention_masks, dim=0) # Moving all padded sequences to GPU for Faster computation input_ids = all_padded_tensors.to("cuda") attention_mask = all_attention_masks.to("cuda") ``` Generate Output Tokens from the and Parse output tokens as speech ```python print("*** Model.generate is slow - see vllm implementation on github for realtime streaming and inference") print("*** Increase/decrease inference params for more expressive less stable generations") # Generating Output Tokens with torch.no_grad(): generated_ids = model.generate( input_ids=input_ids, attention_mask=attention_mask, max_new_tokens=1200, do_sample=True, temperature=0.6, top_p=0.95, repetition_penalty=1.1, num_return_sequences=1, eos_token_id=128258, ) # Processing Generated Tokens (Parse Output as speech) token_to_find = 128257 # Start of Audio token (relevant output) token_to_remove = 128258 # End/ Terminal Token (End of Audio/ relevant output) token_indices = (generated_ids == token_to_find).nonzero(as_tuple=True) print(token_indices) # Slices the tensor to exclude unwanted tokens. if len(token_indices[1]) > 0: last_occurrence_idx = token_indices[1][-1].item() cropped_tensor = generated_ids[:, last_occurrence_idx+1:] else: cropped_tensor = generated_ids # mask = cropped_tensor != token_to_remove # Storing the cleaned-up token sequences# processed_rows = [] for row in cropped_tensor: masked_row = row[row != token_to_remove] processed_rows.append(masked_row) # Preparing (Audio Codes) the token sequences for audio decoding by trimming and adjusting token values. code_lists = [] for row in processed_rows: row_length = row.size(0) # Determines the length of the token sequence. new_length = (row_length // 7) * 7 # Ensures the sequence length is a multiple of 7, as required by the decoder. trimmed_row = row[:new_length] trimmed_row = [t - 128266 for t in trimmed_row] # Adjusts token values to match the expected input range for the decoder. code_lists.append(trimmed_row) ``` Decode Outputs with SNAC Decoder ```python # Processes the token sequences into the format expected by the SNAC decoder: def redistribute_codes(code_list): """ Reorganizes the flattened token list into three separate layers, adjusting each token's value to align with the decoder's expectations""" layer_1 = [] # Coarsest layer layer_2 = [] # Intermediate layer layer_3 = [] # Finest layer num_groups = (len(code_list) + 1) // 7 #Calculate the number of complete 7-token groups in the code_list for i in range(num_groups): idx = 7 * i # starting index for the current group # Layer 1 receives the first token of the group layer_1.append(code_list[idx]) # Layer 2 receives the second token, adjusted by subtracting 4096 layer_2.append(code_list[idx + 1] - 4096) # Layer 3 receives the third and fourth tokens, adjusted by subtracting 8192 and 12288 respectively layer_3.append(code_list[idx+2]-(2*4096)) layer_3.append(code_list[idx+3]-(3*4096)) # Layer 2 receives the fifth token, adjusted by subtracting 16384 layer_2.append(code_list[idx+4]-(4*4096)) # Layer 3 receives the sixth and seventh tokens, adjusted by subtracting 20480 and 24576 respectively layer_3.append(code_list[idx+5]-(5*4096)) layer_3.append(code_list[idx+6]-(6*4096)) codes = [ torch.tensor(layer_1).unsqueeze(0), # Shape: (1, len(layer_1)) torch.tensor(layer_2).unsqueeze(0), # Shape: (1, len(layer_2)) torch.tensor(layer_3).unsqueeze(0) # Shape: (1, len(layer_3)) ] # Convert the lists to PyTorch tensors and add a batch dimension audio_hat = snac_model.decode(codes) # Decode the structured codes into an audio waveform using the SNAC model return audio_hat my_samples = [] for code_list in code_lists: samples = redistribute_codes(code_list) # Generates audio samples from the processed token sequences my_samples.append(samples) # Display Audio from IPython.display import display, Audio if len(prompts) != len(my_samples): raise Exception("Number of prompts and samples do not match") else: for i in range(len(my_samples)): print(prompts[i]) samples = my_samples[i] display(Audio(samples.detach().squeeze().to("cpu").numpy(), rate=24000)) ``` - **Repository:** [N/A] - **Paper:** [N/A] - **Demo:** [N/A] This llama based model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library. [

](https://github.com/unslothai/unsloth)