| | |
|---|
| | | self.batch_size = kwargs.get("batch_size") |
|---|
| | | self.batch_type = kwargs.get("batch_type") |
|---|
| | | self.prompt_ids_len = 0 |
|---|
| | | self.retry = kwargs.get("retry", 10) |
|---|
| | | self.retry = kwargs.get("retry", 100) |
|---|
| | | |
|---|
| | | self.permute = False |
|---|
| | | from funasr.frontends.whisper_frontend import WhisperFrontend |
|---|
| | |
|---|
| | | # self.kwargs = kwargs |
|---|
| | | self.max_token_length = kwargs.get("max_token_length", 1024) |
|---|
| | | self.batch_size_scale_ratio_max = kwargs.get("batch_size_scale_ratio_max", 1.5) |
|---|
| | | self.batch_size_token_max = kwargs.get("batch_size_token_max", 2500) |
|---|
| | | |
|---|
| | | def get_source_len(self, index): |
|---|
| | | item = self.index_ds[index] |
|---|
| | |
|---|
| | | |
|---|
| | | if self.batch_type != "example": |
|---|
| | | b, t = outputs["input_ids"].shape |
|---|
| | | if b * t > self.batch_size * self.batch_size_scale_ratio_max: |
|---|
| | | beg = torch.randint(0, 2, ()).item() |
|---|
| | | if b < 2: |
|---|
| | | beg = 0 |
|---|
| | | if b > 1 and b * t > self.batch_size_token_max: |
|---|
| | | logging.info( |
|---|
| | | f"Warning, b * t: {b * t} > {self.batch_size_scale_ratio_max} * {self.batch_size}, b: {b}, t: {t}, drop half data {idx}th, beg:{beg}" |
|---|
| | | f"Warning, {idx}th, b*t: {b}*{t}={b * t} > batch_size_sample_max: {self.batch_size_token_max}, drop last data" |
|---|
| | | ) |
|---|
| | | samples = samples[beg : beg + b : 2] |
|---|
| | | samples = samples[:-1] |
|---|
| | | continue |
|---|
| | | |
|---|
| | | break |
|---|
