| | |
|---|
| | | |
|---|
| | | meta_data = {} |
|---|
| | | # meta_data["batch_data_time"] = -1 |
|---|
| | | |
|---|
| | | sp_prompt = "<|startoftranscription|><|en|><|transcribe|><|en|><|notimestamps|><|wo_itn|>" |
|---|
| | | query = f"<audio>{data_in[0]}</audio>{sp_prompt}" |
|---|
| | | prompt = kwargs.get("prompt", "<|startoftranscription|><|en|><|transcribe|><|en|><|notimestamps|><|wo_itn|>") |
|---|
| | | query = f"<audio>{data_in[0]}</audio>{prompt}" |
|---|
| | | audio_info = self.tokenizer.process_audio(query) |
|---|
| | | inputs = self.tokenizer(query, return_tensors='pt', audio_info=audio_info) |
|---|
| | | inputs = inputs.to(self.model.device) |
|---|
| | | pred = self.model.generate(**inputs, audio_info=audio_info) |
|---|
| | | response = tokenizer.decode(pred.cpu()[0], skip_special_tokens=False, audio_info=audio_info) |
|---|
| | | response = self.tokenizer.decode(pred.cpu()[0], skip_special_tokens=False, audio_info=audio_info) |
|---|
| | | |
|---|
| | | results = [] |
|---|
| | | result_i = {"key": key[0], "text": response} |
|---|
