| | |
|---|
| | | # for url use "utt_id" as key |
|---|
| | | rec_result = inference_sv_pipline( |
|---|
| | | audio_in='https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/sv_example_enroll.wav') |
|---|
| | | enroll = rec_result["utt_id"] |
|---|
| | | enroll = rec_result["spk_embedding"] |
|---|
| | | |
|---|
| | | # for local file use "utt_id" as key |
|---|
| | | rec_result = inference_sv_pipline(audio_in='sv_example_same.wav')["test1"] |
|---|
| | | same = rec_result["test1"] |
|---|
| | | same = rec_result["spk_embedding"] |
|---|
| | | |
|---|
| | | import soundfile |
|---|
| | | wav = soundfile.read('sv_example_enroll.wav')[0] |
|---|
| | | # for raw inputs use "utt_id" as key |
|---|
| | | spk_embedding = inference_sv_pipline(audio_in=wav)["utt_id"] |
|---|
| | | spk_embedding = inference_sv_pipline(audio_in=wav)["spk_embedding"] |
|---|
| | | |
|---|
| | | rec_result = inference_sv_pipline( |
|---|
| | | audio_in='https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/sv_example_different.wav') |
|---|
| | | different = rec_result["utt_id"] |
|---|
| | | different = rec_result["spk_embedding"] |
|---|
| | | |
|---|
| | | # 对相同的说话人计算余弦相似度 |
|---|
| | | sv_threshold = 0.9465 |
|---|
