Views
No views yet
| Model | Original (FP32) | Q4 Quantized |
|---|---|---|
| speech_encoder.onnx | 564 MB | 172 MB |
| embed_tokens.onnx | 66 MB | 65 MB |
| language_model.onnx | 2.0 GB | 338 MB |
| conditional_decoder.onnx | 510 MB | 215 MB |
| Total | 3.2 GB | 790 MB |
1import onnxruntime
2
3# Load Q4 models - single files, no external data needed
4speech_encoder = onnxruntime.InferenceSession("onnx/speech_encoder.onnx")
5embed_tokens = onnxruntime.InferenceSession("onnx/embed_tokens.onnx")
6language_model = onnxruntime.InferenceSession("onnx/language_model.onnx")
7conditional_decoder = onnxruntime.InferenceSession("onnx/conditional_decoder.onnx")1// Models are single-file ONNX format, compatible with ONNX Runtime Web
2import { AutoTokenizer } from '@huggingface/transformers';
3
4const tokenizer = await AutoTokenizer.from_pretrained('ipsilondev/chatterbox-multilingual-ONNX-q4');MatMulNBitsQuantizer1repetition_penalty = 1.2 # CRITICAL: Do NOT use 2.0 - causes infinite loops
2temperature = 0.8
3top_p = 0.95
4min_p = 0.05| Code | Language | Code | Language |
|---|---|---|---|
| ar | Arabic | ko | Korean |
| da | Danish | ms | Malay |
| de | German | nl | Dutch |
| el | Greek | no | Norwegian |
| en | English | pl | Polish |
| es | Spanish | pt | Portuguese |
| fi | Finnish | ru | Russian |
| fr | French | sv | Swedish |
| he | Hebrew | sw | Swahili |
| hi | Hindi | tr | Turkish |
| it | Italian | zh | Chinese |
| ja | Japanese |