执行构建agent代码的时候报错如下：Qwen2TokenizerFast has no attribute "get _token_id" . Did you mean:"sep_token_id"

[qxevtnuko]

Self Checks

• This template is only for bug reports. For questions, please visit Discussions.
• I have thoroughly reviewed the project documentation (installation, training, inference) but couldn't find information to solve my problem. English 中文 日本語 Portuguese (Brazil)
• I have searched for existing issues, including closed ones. Search issues
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
• Please do not modify this template and fill in all required fields.

Cloud or Self Hosted

Self Hosted (Source)

Environment Details

python 3.10 fish-speech-1.5 windows

Steps to Reproduce

When I python -m tools.api --llama-checkpoint-path checkpoints/fish-agent-v0.1-3b/ --mode agent --compile ,
AttributeError: Qwen2TokenizerFast has no attribute "get _token_id" . Did you mean:"sep_token_id"

✔️ Expected Behavior

No response

❌ Actual Behavior

No response

[bellkjtt]

same error occured.

[AnyaCoder]

先用1.4版本

[None1024Outlook]

Self Checks

• This template is only for bug reports. For questions, please visit Discussions.
• I have thoroughly reviewed the project documentation (installation, training, inference) but couldn't find information to solve my problem. English 中文 日本語 Portuguese (Brazil)
• I have searched for existing issues, including closed ones. Search issues
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
• Please do not modify this template and fill in all required fields.

Cloud or Self Hosted

Self Hosted (Source)

Environment Details

python 3.10 fish-speech-1.5 windows

Steps to Reproduce

When I python -m tools.api --llama-checkpoint-path checkpoints/fish-agent-v0.1-3b/ --mode agent --compile , AttributeError: Qwen2TokenizerFast has no attribute "get _token_id" . Did you mean:"sep_token_id"

✔️ Expected Behavior

No response

❌ Actual Behavior

No response

Please use version 1.4 of the code, or you can refer to version 1.4 to modify the fish_stpeech/tokenizer. py file

[zhouzhouwen]

Hello, fish-speech is an amazing library, and I truly appreciate your work. I have also encountered the same error with the latest version.

(fish-speech) root@3aac063e8786:/app/data/zhouwen/fish-speech# python -m tools.api_server --llama-checkpoint-path /app/data/zhouwen/fish-speech/checkpoints/fish-agent-v0.1-3b --mode agent --compile python -m tools.api_server --llama-checkpoint-path /app/data/zhouwen/fish-speech/checkpoints/fish-agent-v0.1-3b --mode agent --compile
INFO: Started server process [3919003]
INFO: Waiting for application startup.
Exception in thread Thread-1 (worker):
Traceback (most recent call last):
File "/app/data/zhouwen/Anaconda3/envs/fish-speech/lib/python3.10/threading.py", line 1016, in _bootstrap_inner
self.run()
File "/app/data/zhouwen/Anaconda3/envs/fish-speech/lib/python3.10/threading.py", line 953, in run
self._target(*self._args, **self._kwargs)
File "/app/data/zhouwen/fish-speech/tools/llama/generate.py", line 960, in worker
model, decode_one_token = load_model(
File "/app/data/zhouwen/fish-speech/tools/llama/generate.py", line 677, in load_model
model: Union[NaiveTransformer, DualARTransformer] = BaseTransformer.from_pretrained(
File "/app/data/zhouwen/fish-speech/fish_speech/models/text2semantic/llama.py", line 411, in from_pretrained
model = model_cls(config, tokenizer=tokenizer)
File "/app/data/zhouwen/fish-speech/fish_speech/models/text2semantic/llama.py", line 542, in init
super().init(config, init_weights=False, tokenizer=tokenizer)
File "/app/data/zhouwen/fish-speech/fish_speech/models/text2semantic/llama.py", line 176, in init
self.semantic_token_ids = [
File "/app/data/zhouwen/fish-speech/fish_speech/models/text2semantic/llama.py", line 177, in
tokenizer.get_token_id(SEMANTIC_TOKEN) for SEMANTIC_TOKEN in SEMANTIC_TOKENS
File "/app/data/zhouwen/Anaconda3/envs/fish-speech/lib/python3.10/site-packages/transformers/tokenization_utils_base.py", line 1104, in getattr
raise AttributeError(f"{self.class.name} has no attribute {key}")
AttributeError: Qwen2TokenizerFast has no attribute get_token_id

Following your previous responses, I tried using version 1.4.0, but the error still persists.

(fish-speech) root@3aac063e8786:/app/data/zhouwen/fish-speech-1.4.0# python -m tools.api --llama-checkpoint-path /app/data/zhouwen/fish-speech-1.4.0/checkpoints/fish-agent-v0.1-3b --compile
2024-12-09 06:17:49.674 | INFO | main::403 - Loading Llama model...
Exception in thread Thread-1 (worker):
Traceback (most recent call last):
File "/app/data/zhouwen/Anaconda3/envs/fish-speech/lib/python3.10/threading.py", line 1016, in _bootstrap_inner
self.run()
File "/app/data/zhouwen/Anaconda3/envs/fish-speech/lib/python3.10/threading.py", line 953, in run
self._target(*self._args, **self._kwargs)
File "/app/data/zhouwen/fish-speech-1.4.0/tools/llama/generate.py", line 561, in worker
model, decode_one_token = load_model(
File "/app/data/zhouwen/fish-speech-1.4.0/tools/llama/generate.py", line 343, in load_model
model: Union[NaiveTransformer, DualARTransformer] = BaseTransformer.from_pretrained(
File "/app/data/zhouwen/fish-speech-1.4.0/fish_speech/models/text2semantic/llama.py", line 325, in from_pretrained
config = BaseModelArgs.from_pretrained(str(path))
File "/app/data/zhouwen/fish-speech-1.4.0/fish_speech/models/text2semantic/llama.py", line 87, in from_pretrained
return cls(**data)
TypeError: DualARModelArgs.init() got an unexpected keyword argument 'fast_attention_qkv_bias'

I look forward to your reply.

[AnyaCoder]

1.4.3

[zhouzhouwen]

Thank you so much!!! It works now.