funasr1.0

游雁

2024-01-04 36702d28297839f2b41be9a60e822a4059dbe3f6

funasr1.0

 examples/industrial_data_pretraining/paraformer-long/infer.sh

@@ -22,5 +22,6 @@
+device="cpu" \
+batch_size_s=300 \
+batch_size_threshold_s=60 \
+debug="true"
+debug="true" \
+"hotword='达摩院 魔搭'"


 examples/industrial_data_pretraining/paraformer/demo.py

@@ -5,17 +5,17 @@

from funasr import AutoModel

model = AutoModel(model="/Users/zhifu/Downloads/modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch")
model = AutoModel(model="../modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch")

res = model(input="/Users/zhifu/Downloads/modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/example/asr_example.wav")
res = model(input="../modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/example/asr_example.wav")
print(res)


from funasr import AutoFrontend

frontend = AutoFrontend(model="/Users/zhifu/Downloads/modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch")
frontend = AutoFrontend(model="../modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch")

fbanks = frontend(input="/Users/zhifu/funasr_github/test_local/wav.scp", batch_size=2)
fbanks = frontend(input="../modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/example/asr_example.wav", batch_size=2)

for batch_idx, fbank_dict in enumerate(fbanks):
    res = model(**fbank_dict)

 examples/industrial_data_pretraining/paraformer/finetune.sh

@@ -1,12 +1,14 @@

cmd="funasr/bin/train.py"
# download model
local_path_root=../modelscope_models
mkdir -p ${local_path_root}
local_path=${local_path_root}/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch
git clone https://www.modelscope.cn/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch.git ${local_path}

python $cmd \
+model="/Users/zhifu/modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch" \
+token_list="/Users/zhifu/.cache/modelscope/hub/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/tokens.txt" \
+train_data_set_list="/Users/zhifu/funasr_github/test_local/aishell2_dev_ios/asr_task_debug_len.jsonl" \
+output_dir="/Users/zhifu/Downloads/ckpt/funasr2/exp2" \

python funasr/bin/train.py \
+model="../modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch" \
+token_list="../modelscope_models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/tokens.txt" \
+train_data_set_list="data/list/audio_datasets.jsonl" \
+output_dir="outputs/debug/ckpt/funasr2/exp2" \
+device="cpu"

#--config-path "/Users/zhifu/funasr_github/examples/industrial_data_pretraining/paraformer-large/conf" \
#--config-name "finetune.yaml" \

 examples/industrial_data_pretraining/paraformer/infer.sh

@@ -1,9 +1,9 @@

# download model
local_path_root=./modelscope_models
local_path_root=../modelscope_models
mkdir -p ${local_path_root}
local_path=${local_path_root}/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch
git clone https://www.modelscope.cn/damo/speech_paraformer-large-contextual_asr_nat-zh-cn-16k-common-vocab8404.git ${local_path}
git clone https://www.modelscope.cn/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch.git ${local_path}


python funasr/bin/inference.py \

 funasr/version.txt

@@ -1 +1 @@
0.8.7
1.0.0