sh1gechan commited on
Commit
29b13b8
·
verified ·
1 Parent(s): 8a9dbef

Update src/leaderboard/read_evals.py

Browse files
Files changed (1) hide show
  1. src/leaderboard/read_evals.py +4 -34
src/leaderboard/read_evals.py CHANGED
@@ -108,30 +108,6 @@ class EvalResult:
108
  except Exception:
109
  print(f"Could not find request file for {self.org}/{self.model} with precision {self.precision.value.name}")
110
 
111
- # def to_dict(self):
112
- # """Converts the Eval Result to a dict compatible with our dataframe display"""
113
- # average = sum([v for v in self.results.values() if v is not None]) / len(Tasks)
114
- # data_dict = {
115
- # "eval_name": self.eval_name, # not a column, just a save name,
116
- # AutoEvalColumn.precision.name: self.precision.value.name,
117
- # AutoEvalColumn.model_type.name: self.model_type.value.name,
118
- # AutoEvalColumn.model_type_symbol.name: self.model_type.value.symbol,
119
- # AutoEvalColumn.weight_type.name: self.weight_type.value.name,
120
- # AutoEvalColumn.architecture.name: self.architecture,
121
- # AutoEvalColumn.model.name: make_clickable_model(self.full_model),
122
- # AutoEvalColumn.revision.name: self.revision,
123
- # AutoEvalColumn.average.name: average,
124
- # AutoEvalColumn.license.name: self.license,
125
- # AutoEvalColumn.likes.name: self.likes,
126
- # AutoEvalColumn.params.name: self.num_params,
127
- # AutoEvalColumn.still_on_hub.name: self.still_on_hub,
128
- # }
129
-
130
- # for task in Tasks:
131
- # data_dict[task.value.col_name] = self.results.get(task.value.benchmark, None)
132
- # print(f"Debug: {task.value.col_name} = {self.results.get(task.value.benchmark, 'N/A')}")
133
-
134
- # return data_dict
135
  def to_dict(self):
136
  """Converts the Eval Result to a dict compatible with our dataframe display"""
137
  average = sum([v for v in self.results.values() if v is not None]) / len(self.results)
@@ -150,18 +126,12 @@ class EvalResult:
150
  AutoEvalColumn.params.name: self.num_params,
151
  AutoEvalColumn.still_on_hub.name: self.still_on_hub,
152
  }
153
-
154
- for task in Tasks:
155
- task_value = self.results.get(task.value.benchmark)
156
- if task_value is not None:
157
- data_dict[task.value.col_name] = task_value
158
- else:
159
- data_dict[task.value.col_name] = 'N/A' # スコアが存在しない場合の処理
160
- print(f"Debug: {task.value.col_name} = {data_dict[task.value.col_name]}")
161
-
162
- return data_dict
163
 
 
 
 
164
 
 
165
 
166
  def get_request_file_for_model(requests_path, model_name, precision):
167
  """Selects the correct request file for a given model. Only keeps runs tagged as FINISHED"""
 
108
  except Exception:
109
  print(f"Could not find request file for {self.org}/{self.model} with precision {self.precision.value.name}")
110
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
111
  def to_dict(self):
112
  """Converts the Eval Result to a dict compatible with our dataframe display"""
113
  average = sum([v for v in self.results.values() if v is not None]) / len(self.results)
 
126
  AutoEvalColumn.params.name: self.num_params,
127
  AutoEvalColumn.still_on_hub.name: self.still_on_hub,
128
  }
 
 
 
 
 
 
 
 
 
 
129
 
130
+ for task in Tasks:
131
+ data_dict[task.value.col_name] = self.results.get(task.value.benchmark, None)
132
+ print(f"Debug: {task.value.col_name} = {self.results.get(task.value.benchmark, 'N/A')}")
133
 
134
+ return data_dict
135
 
136
  def get_request_file_for_model(requests_path, model_name, precision):
137
  """Selects the correct request file for a given model. Only keeps runs tagged as FINISHED"""