AlibabaResearch / DAMO-ConvAI

DAMO-ConvAI: The official repository which contains the codebase for Alibaba DAMO Conversational AI.
MIT License
1.15k stars 185 forks source link

请问STAR有LARGE的中文模型吗? #22

Closed oushu1zhangxiangxuan1 closed 1 year ago

oushu1zhangxiangxuan1 commented 1 year ago
  1. SAPCE-T的中文base模型貌似效果不太好,join无法支持
  2. SPACE-T的英文large模型,按modelscope给的方式运行报错(dgl版本试过最新nightly的也不行)
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ /home/johnsaxon/NL2SQL/SPACE-T/star/star.py:29 in <module>                                       │
│                                                                                                  │
│   26 │   │   │   "last_sql": last_sql,                                                           │
│   27 │   │   │   "database_id": test_case['database_id'],                                        │
│   28 │   │   │   'local_db_path': test_case['local_db_path']}                                    │
│ ❱ 29 │   results = pipeline(case)                                                                │
│   30 │   print(results)                                                                          │
│   31 │   history.append(item)                                                                    │
│   32                                                                                             │
│                                                                                                  │
│ /home/johnsaxon/aidd/localcolabfold/colabfold_batch/colabfold-conda/lib/python3.8/site-packages/ │
│ modelscope/pipelines/base.py:212 in __call__                                                     │
│                                                                                                  │
│   209 │   │   │   return self._process_iterator(input, *args, **kwargs)                          │
│   210 │   │                                                                                      │
│   211 │   │   else:                                                                              │
│ ❱ 212 │   │   │   output = self._process_single(input, *args, **kwargs)                          │
│   213 │   │   return output                                                                      │
│   214 │                                                                                          │
│   215 │   def _sanitize_parameters(self, **pipeline_parameters):                                 │
│                                                                                                  │
│ /home/johnsaxon/aidd/localcolabfold/colabfold_batch/colabfold-conda/lib/python3.8/site-packages/ │
│ modelscope/pipelines/base.py:240 in _process_single                                              │
│                                                                                                  │
│   237 │   │   forward_params = kwargs.get('forward_params', {})                                  │
│   238 │   │   postprocess_params = kwargs.get('postprocess_params', {})                          │
│   239 │   │   self._check_input(input)                                                           │
│ ❱ 240 │   │   out = self.preprocess(input, **preprocess_params)                                  │
│   241 │   │                                                                                      │
│   242 │   │   with device_placement(self.framework, self.device_name):                           │
│   243 │   │   │   if self.framework == Frameworks.torch:                                         │
│                                                                                                  │
│ /home/johnsaxon/aidd/localcolabfold/colabfold_batch/colabfold-conda/lib/python3.8/site-packages/ │
│ modelscope/pipelines/base.py:369 in preprocess                                                   │
│                                                                                                  │
│   366 │   │   assert self.preprocessor is not None, 'preprocess method should be implemented'    │
│   367 │   │   assert not isinstance(self.preprocessor, List),\                                   │
│   368 │   │   │   'default implementation does not support using multiple preprocessors.'        │
│ ❱ 369 │   │   return self.preprocessor(inputs, **preprocess_params)                              │
│   370 │                                                                                          │
│   371 │   def forward(self, inputs: Dict[str, Any],                                              │
│   372 │   │   │   │   **forward_params) -> Dict[str, Any]:                                       │
│                                                                                                  │
│ /home/johnsaxon/aidd/localcolabfold/colabfold_batch/colabfold-conda/lib/python3.8/site-packages/ │
│ modelscope/utils/type_assert.py:48 in wrapper                                                    │
│                                                                                                  │
│   45 │   │   │   │   │   if not isinstance(value, bound_types[name]):                            │
│   46 │   │   │   │   │   │   raise TypeError('Argument {} must be {}'.format(                    │
│   47 │   │   │   │   │   │   │   name, bound_types[name]))                                       │
│ ❱ 48 │   │   │   return func(*args, **kwargs)                                                    │
│   49 │   │                                                                                       │
│   50 │   │   return wrapper                                                                      │
│   51                                                                                             │
│                                                                                                  │
│ /home/johnsaxon/aidd/localcolabfold/colabfold_batch/colabfold-conda/lib/python3.8/site-packages/ │
│ modelscope/preprocessors/nlp/space_T_en/conversational_text_to_sql_preprocessor.py:113 in        │
│ __call__                                                                                         │
│                                                                                                  │
│   110 │   │   output_dataset = process_dataset(self.model_dir, self.processor,                   │
│   111 │   │   │   │   │   │   │   │   │   │    theresult, self.output_tables)                    │
│   112 │   │   output_dataset = \                                                                 │
│ ❱ 113 │   │   │   process_dataset_graph(                                                         │
│   114 │   │   │   │   self.graph_processor,                                                      │
│   115 │   │   │   │   output_dataset,                                                            │
│   116 │   │   │   │   self.output_tables,                                                        │
│                                                                                                  │
│ /home/johnsaxon/aidd/localcolabfold/colabfold_batch/colabfold-conda/lib/python3.8/site-packages/ │
│ text2sql_lgesql/preprocess/process_graphs.py:14 in process_dataset_graph                         │
│                                                                                                  │
│   11 │   │   │   continue                                                                        │
│   12 │   │   if (idx + 1) % 500 == 0:                                                            │
│   13 │   │   │   print('Processing the %d-th example ...' % (idx + 1))                           │
│ ❱ 14 │   │   entry = processor.process_graph_utils(entry, db, method=method)                     │
│   15 │   │   processed_dataset.append(entry)                                                     │
│   16 │   # print('In total, process %d samples, skip %d samples .' % (len(processed_dataset),    │
│   17 │   if output_path is not None:                                                             │
│                                                                                                  │
│ /home/johnsaxon/aidd/localcolabfold/colabfold_batch/colabfold-conda/lib/python3.8/site-packages/ │
│ text2sql_lgesql/preprocess/graph_utils.py:112 in process_graph_utils                             │
│                                                                                                  │
│   109 │   │   if method == 'rgatsql':                                                            │
│   110 │   │   │   ex = self.process_rgatsql(ex, db, relation)                                    │
│   111 │   │   elif method == 'lgesql':                                                           │
│ ❱ 112 │   │   │   ex = self.process_lgesql(ex, db, relation)                                     │
│   113 │   │   return ex                                                                          │
│   114                                                                                            │
│                                                                                                  │
│ /home/johnsaxon/aidd/localcolabfold/colabfold_batch/colabfold-conda/lib/python3.8/site-packages/ │
│ text2sql_lgesql/preprocess/graph_utils.py:93 in process_lgesql                                   │
│                                                                                                  │
│    90 │   │   match_ids = [idx for idx, r in enumerate(graph.global_edges) if 'match' in r[2]]   │
│    91 │   │   src, dst, eids = lg.edges(form='all', order='eid')                                 │
│    92 │   │   eids = [e for u, v, e in zip(src.tolist(), dst.tolist(), eids.tolist()) if not (   │
│ ❱  93 │   │   graph.lg = lg.edge_subgraph(eids, preserve_nodes=True).remove_self_loop().add_se   │
│    94 │   │   ex['graph'] = graph                                                                │
│    95 │   │   return ex                                                                          │
│    96                                                                                            │
│                                                                                                  │
│ /home/johnsaxon/aidd/localcolabfold/colabfold_batch/colabfold-conda/lib/python3.8/site-packages/ │
│ dgl/utils/internal.py:1052 in _fn                                                                │
│                                                                                                  │
│   1049 │                                                                                         │
│   1050 │   @wraps(func)                                                                          │
│   1051 │   def _fn(*args, **kwargs):                                                             │
│ ❱ 1052 │   │   return func(*args, **kwargs)                                                      │
│   1053 │                                                                                         │
│   1054 │   _fn.__doc__ = """Alias of :func:`dgl.{}`.""".format(func.__name__)                    │
│   1055 │   return _fn                                                                            │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
TypeError: edge_subgraph() got an unexpected keyword argument 'preserve_nodes'
huybery commented 1 year ago

你好,关于 modelscope 的疑问,麻烦在 https://developer.aliyun.com/space/modelscope/nlp?spm=a2c6h.17735062.J_1779720920.4.197072d7rbAf6F 提问,会有专人支持。

oushu1zhangxiangxuan1 commented 1 year ago

你好,关于 modelscope 的疑问,麻烦在 https://developer.aliyun.com/space/modelscope/nlp?spm=a2c6h.17735062.J_1779720920.4.197072d7rbAf6F 提问,会有专人支持。

请问STAR有中文Large模型吗

huybery commented 1 year ago

@oushu1zhangxiangxuan1 抱歉,没有中文 Large