Text-to-Speech
Transformers
Safetensors
PyTorch
English
Chinese
breeze
text-generation
cuda
english
chinese
speech-generation
voice-clone
voice-design
voice-direction
comfyui
int8
convrot
quantization
quantised
hadamard
Instructions to use drbaph/Breeze-TTS-2-comfyui with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use drbaph/Breeze-TTS-2-comfyui with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-to-speech", model="drbaph/Breeze-TTS-2-comfyui")# Load model directly from transformers import AutoModelForSeq2SeqLM model = AutoModelForSeq2SeqLM.from_pretrained("drbaph/Breeze-TTS-2-comfyui", device_map="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "bf16": { | |
| "description": "all linears bf16 (Breeze-TTS-2-bf16.safetensors)", | |
| "load_s": 7.6, | |
| "avg_rtf": 5.681, | |
| "max_peak_vram_gib": 7.51, | |
| "prompts": { | |
| "design_en": { | |
| "audio_s": 4.16, | |
| "gen_s": 24.1, | |
| "rtf": 5.797, | |
| "peak_vram_gib": 7.32, | |
| "logmel_l1_vs_bf16": 0.0, | |
| "wer": 0.0, | |
| "whisper": "It is good to hear your voice again after all this time." | |
| }, | |
| "design_zh": { | |
| "audio_s": 2.8, | |
| "gen_s": 14.9, | |
| "rtf": 5.324, | |
| "peak_vram_gib": 7.25, | |
| "logmel_l1_vs_bf16": 0.0, | |
| "wer": 0.0, | |
| "whisper": "欢迎来到今晚的故事时间让我们一起开始吧" | |
| }, | |
| "clone_en": { | |
| "audio_s": 6.24, | |
| "gen_s": 37.0, | |
| "rtf": 5.922, | |
| "peak_vram_gib": 7.51, | |
| "logmel_l1_vs_bf16": 0.0, | |
| "wer": 0.1667, | |
| "whisper": "This is a clone test of the Breeze TTS2 Comfy UI node pack. Now running with quantized weights." | |
| }, | |
| "direction_en": { | |
| "audio_s": 4.8, | |
| "gen_s": 27.3, | |
| "rtf": 5.68, | |
| "peak_vram_gib": 7.42, | |
| "logmel_l1_vs_bf16": 0.0, | |
| "wer": 0.0, | |
| "whisper": "We need to discuss what happened last night." | |
| } | |
| } | |
| }, | |
| "int8-convrot": { | |
| "description": "all 462 transformer linears int8 convrot (Breeze-TTS-2-int8-convrot.safetensors)", | |
| "load_s": 9.6, | |
| "avg_rtf": 9.08, | |
| "max_peak_vram_gib": 5.21, | |
| "prompts": { | |
| "design_en": { | |
| "audio_s": 4.72, | |
| "gen_s": 36.7, | |
| "rtf": 7.778, | |
| "peak_vram_gib": 5.06, | |
| "logmel_l1_vs_bf16": 1.796, | |
| "wer": 0.0833, | |
| "whisper": "Ah, it is good to hear your voice again, after all this time." | |
| }, | |
| "design_zh": { | |
| "audio_s": 4.24, | |
| "gen_s": 41.0, | |
| "rtf": 9.664, | |
| "peak_vram_gib": 5.04, | |
| "logmel_l1_vs_bf16": 3.4857, | |
| "wer": 0.0, | |
| "whisper": "欢迎来到今晚的故事时间让我们一起开始吧" | |
| }, | |
| "clone_en": { | |
| "audio_s": 6.24, | |
| "gen_s": 57.8, | |
| "rtf": 9.262, | |
| "peak_vram_gib": 5.21, | |
| "logmel_l1_vs_bf16": 1.3277, | |
| "wer": 0.2222, | |
| "whisper": "This is a clone test of the Breeze TTS2 Comfy UI node pack, now running with Quantite weights." | |
| }, | |
| "direction_en": { | |
| "audio_s": 3.68, | |
| "gen_s": 35.4, | |
| "rtf": 9.617, | |
| "peak_vram_gib": 5.07, | |
| "logmel_l1_vs_bf16": 2.1019, | |
| "wer": 0.0, | |
| "whisper": "We need to discuss what happened last night." | |
| } | |
| } | |
| }, | |
| "int8-hybrid": { | |
| "description": "backbone+text_encoder int8 convrot, depth decoder bf16 (Breeze-TTS-2-int8-hybrid.safetensors)", | |
| "load_s": 9.2, | |
| "avg_rtf": 5.642, | |
| "max_peak_vram_gib": 5.53, | |
| "prompts": { | |
| "design_en": { | |
| "audio_s": 3.52, | |
| "gen_s": 18.8, | |
| "rtf": 5.338, | |
| "peak_vram_gib": 5.3, | |
| "logmel_l1_vs_bf16": 1.634, | |
| "wer": 0.0, | |
| "whisper": "It is good to hear your voice again after all this time." | |
| }, | |
| "design_zh": { | |
| "audio_s": 4.16, | |
| "gen_s": 22.6, | |
| "rtf": 5.421, | |
| "peak_vram_gib": 5.34, | |
| "logmel_l1_vs_bf16": 3.5299, | |
| "wer": 0.0, | |
| "whisper": "欢迎来到今晚的故事时间让我们一起开始吧" | |
| }, | |
| "clone_en": { | |
| "audio_s": 6.32, | |
| "gen_s": 33.5, | |
| "rtf": 5.304, | |
| "peak_vram_gib": 5.53, | |
| "logmel_l1_vs_bf16": 1.0783, | |
| "wer": 0.2222, | |
| "whisper": "This is a clone test of the Breeze TTS2 Comfy UI node pack. Now running with quantize weights." | |
| }, | |
| "direction_en": { | |
| "audio_s": 3.84, | |
| "gen_s": 25.0, | |
| "rtf": 6.506, | |
| "peak_vram_gib": 5.38, | |
| "logmel_l1_vs_bf16": 1.9786, | |
| "wer": 0.0, | |
| "whisper": "We need to discuss what happened last night." | |
| } | |
| } | |
| }, | |
| "int8-te": { | |
| "description": "text_encoder int8 convrot only (Breeze-TTS-2-int8-text-encoder.safetensors)", | |
| "load_s": 8.4, | |
| "avg_rtf": 5.342, | |
| "max_peak_vram_gib": 6.82, | |
| "prompts": { | |
| "design_en": { | |
| "audio_s": 4.24, | |
| "gen_s": 25.1, | |
| "rtf": 5.928, | |
| "peak_vram_gib": 6.66, | |
| "logmel_l1_vs_bf16": 1.1453, | |
| "wer": 0.0, | |
| "whisper": "It is good to hear your voice again after all this time." | |
| }, | |
| "design_zh": { | |
| "audio_s": 3.44, | |
| "gen_s": 17.5, | |
| "rtf": 5.097, | |
| "peak_vram_gib": 6.61, | |
| "logmel_l1_vs_bf16": 2.3411, | |
| "wer": 0.0, | |
| "whisper": "欢迎来到今晚的故事时间,让我们一起开始吧。" | |
| }, | |
| "clone_en": { | |
| "audio_s": 6.08, | |
| "gen_s": 32.1, | |
| "rtf": 5.275, | |
| "peak_vram_gib": 6.82, | |
| "logmel_l1_vs_bf16": 1.1573, | |
| "wer": 0.2222, | |
| "whisper": "This is a clone test of the Breeze TTS2 Comfy UI node pack. Now running with quantize weights." | |
| }, | |
| "direction_en": { | |
| "audio_s": 4.48, | |
| "gen_s": 22.7, | |
| "rtf": 5.068, | |
| "peak_vram_gib": 6.73, | |
| "logmel_l1_vs_bf16": 2.6358, | |
| "wer": 0.0, | |
| "whisper": "We need to discuss what happened last night." | |
| } | |
| } | |
| } | |
| } |