fix requirements

egs/speech_llm/ASR_LLM/debug.sh

@@ -17,8 +17,8 @@ export CUDA_VISIBLE_DEVICES=0,1
 
 python3 ./whisper_llm_zh/decode.py \
   --max-duration 80 \
-  --exp-dir ./whisper_llm_zh/exp_test \
-  --speech-encoder-path-or-name tiny \
+  --exp-dir ./whisper_llm_zh/exp_qwen_0.5B \
+  --speech-encoder-path-or-name /mnt/samsung-t7/yuekai/asr/v1.1/whisper-large-v2-multi-hans-zh-epoch-3-avg-10.pt  \
   --llm-path-or-name Qwen/Qwen1.5-0.5B-Chat \
   --epoch 1 --avg 1 \
   --manifest-dir data/fbank \

egs/speech_llm/ASR_LLM/whisper_llm_zh/decode.py

@@ -291,7 +291,7 @@ def decode_one_batch(
 
     generated_ids = model.decode(feature, input_ids.to(device, dtype=torch.long), attention_mask.to(device))
     hyps = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
-
+    hyps = tokenizer.batch_decode(generated_ids, skip_special_tokens=False)[0]
     # hyps = remove_punctuation(hyps)
     # hyps = to_simple(hyps)
     # hyps = [params.normalizer.normalize(hyp) for hyp in hyps]
@@ -377,7 +377,7 @@ def decode_dataset(
 
         for lm_scale, hyps in hyps_dict.items():
             this_batch = []
-            assert len(hyps) == len(texts)
+            # assert len(hyps) == len(texts)
             for cut_id, hyp_words, ref_text in zip(cut_ids, hyps, texts):
                 ref_text = normalize_text_alimeeting(ref_text)
                 ref_words = ref_text.split()

egs/speech_llm/ASR_LLM/whisper_llm_zh/model.py

@@ -169,7 +169,8 @@ class SPEECH_LLM(nn.Module):
             eos_token_id=self.llm.config.eos_token_id,
             pad_token_id=self.llm.config.pad_token_id
         )
-        generated_ids = [
-            output_ids[len(input_ids):] for input_ids, output_ids in zip(input_ids, generated_ids)
-        ]
+        # print(generated_ids, input_ids)
+        # generated_ids = [
+        #     output_ids[len(input_ids):] for input_ids, output_ids in zip(input_ids, generated_ids)
+        # ]
         return generated_ids

egs/speech_llm/ASR_LLM/whisper_llm_zh/requirements.txt

@@ -1,7 +1,8 @@
 k2
 kaldialign
 git+https://github.com/lhotse-speech/lhotse
-# sentencepiece
+sentencepiece
+pypinyin
 tensorboard
 librosa
 # git+https://github.com/yuekaizhang/whisper.git