auto frontend

游雁

2024-06-07 870fa4a37a194fb7919b53930064129b6d616cb7

 funasr/models/llm_asr/model.py

@@ -468,7 +468,7 @@
        if len(speech_lengths.size()) > 1:
            speech_lengths = speech_lengths[:, 0]

        batch_size = speech.shape[0]
        batch_size, frames, _ = speech.shape

        # audio encoder
        encoder_out, encoder_out_lens = self.audio_encoder(speech.permute(0, 2, 1), speech_lengths)
@@ -476,23 +476,18 @@
        # audio_adaptor
        encoder_out, encoder_out_lens = self.audio_adaptor(encoder_out, encoder_out_lens)

        input_ids[input_ids == -1] = 0
        input_ids[input_ids == -100] = 0
        if hasattr(self.llm.model, "embed_tokens"):
            inputs_embeds = self.llm.model.embed_tokens(input_ids)
        elif hasattr(self.llm.model.model, "embed_tokens"):
            inputs_embeds = self.llm.model.model.embed_tokens(input_ids)
        else:
            inputs_embeds = self.llm.model.model.model.embed_tokens(input_ids)
        input_ids[input_ids < 0] = 0
        inputs_embeds = self.llm.model.get_input_embeddings()(input_ids)

        batch_size, token_num, dims = inputs_embeds.shape
        _, l, _ = encoder_out.shape
        for batch_idx in range(batch_size):
            fbank_beg_idx = fbank_beg[batch_idx, 0].item()
            inputs_embeds[batch_idx, fbank_beg_idx : fbank_beg_idx + l, :] = encoder_out[
                batch_idx, :l, :
            min_len = min(l, inputs_embeds.shape[1] - fbank_beg_idx)
            inputs_embeds[batch_idx, fbank_beg_idx : fbank_beg_idx + min_len, :] = encoder_out[
                batch_idx, :min_len, :
            ]

        labels_ids[labels_ids == -1] = -100
        model_outputs = self.llm(
            inputs_embeds=inputs_embeds, attention_mask=attention_mask, labels=labels_ids
        )
@@ -505,6 +500,13 @@
            stats["acc"] = acc_att

        stats["loss"] = torch.clone(loss.detach())
        stats["batch_size"] = batch_size
        stats["batch_size_x_frames"] = frames * batch_size
        stats["batch_size_real_frames"] = speech_lengths.sum().item()
        stats["padding_frames"] = stats["batch_size_x_frames"] - stats["batch_size_real_frames"]
        stats["batch_size_x_tokens"] = token_num * batch_size
        stats["batch_size_real_tokens"] = attention_mask.sum().item()
        stats["padding_tokens"] = stats["batch_size_x_tokens"] - stats["batch_size_real_tokens"]

        # force_gatherable: to-device and to-tensor if scalar for DataParallel
        if self.length_normalized_loss: