Spaces:
Running
Running
ZeroCommand
commited on
Commit
•
9e212de
1
Parent(s):
1aa43b4
updated version of ui
Browse files- app.py +74 -81
- text_classification.py +14 -17
app.py
CHANGED
@@ -59,26 +59,28 @@ def check_dataset(dataset_id, dataset_config="default", dataset_split="test"):
|
|
59 |
return dataset_id, None, None
|
60 |
return dataset_id, dataset_config, dataset_split
|
61 |
|
62 |
-
def try_validate(model_id, dataset_id, dataset_config, dataset_split, column_mapping):
|
63 |
# Validate model
|
64 |
m_id, ppl = check_model(model_id=model_id)
|
65 |
if m_id is None:
|
66 |
gr.Warning(f'Model "{model_id}" is not accessible. Please set your HF_TOKEN if it is a private model.')
|
67 |
return (
|
68 |
-
dataset_config, dataset_split,
|
69 |
gr.update(interactive=False), # Submit button
|
|
|
|
|
|
|
70 |
gr.update(visible=False), # Model prediction preview
|
71 |
gr.update(visible=False), # Label mapping preview
|
72 |
-
gr.update(visible=True), # Column mapping
|
73 |
)
|
74 |
if isinstance(ppl, Exception):
|
75 |
gr.Warning(f'Failed to load "{model_id} model": {ppl}')
|
76 |
return (
|
77 |
-
dataset_config, dataset_split,
|
78 |
gr.update(interactive=False), # Submit button
|
|
|
|
|
|
|
79 |
gr.update(visible=False), # Model prediction preview
|
80 |
gr.update(visible=False), # Label mapping preview
|
81 |
-
gr.update(visible=True), # Column mapping
|
82 |
)
|
83 |
|
84 |
# Validate dataset
|
@@ -98,11 +100,13 @@ def try_validate(model_id, dataset_id, dataset_config, dataset_split, column_map
|
|
98 |
|
99 |
if not dataset_ok:
|
100 |
return (
|
101 |
-
config, split,
|
102 |
gr.update(interactive=False), # Submit button
|
|
|
|
|
|
|
103 |
gr.update(visible=False), # Model prediction preview
|
104 |
gr.update(visible=False), # Label mapping preview
|
105 |
-
gr.update(visible=True), # Column mapping
|
106 |
)
|
107 |
|
108 |
# TODO: Validate column mapping by running once
|
@@ -110,11 +114,12 @@ def try_validate(model_id, dataset_id, dataset_config, dataset_split, column_map
|
|
110 |
id2label_df = None
|
111 |
if isinstance(ppl, TextClassificationPipeline):
|
112 |
try:
|
|
|
113 |
column_mapping = json.loads(column_mapping)
|
114 |
except Exception:
|
115 |
column_mapping = {}
|
116 |
|
117 |
-
column_mapping, prediction_result, id2label_df = \
|
118 |
text_classification_fix_column_mapping(column_mapping, ppl, d_id, config, split)
|
119 |
|
120 |
column_mapping = json.dumps(column_mapping, indent=2)
|
@@ -124,31 +129,35 @@ def try_validate(model_id, dataset_id, dataset_config, dataset_split, column_map
|
|
124 |
if prediction_result is None:
|
125 |
gr.Warning('The model failed to predict with the first row in the dataset. Please provide column mappings in "Advance" settings.')
|
126 |
return (
|
127 |
-
config, split,
|
128 |
gr.update(interactive=False), # Submit button
|
|
|
|
|
|
|
129 |
gr.update(visible=False), # Model prediction preview
|
130 |
gr.update(visible=False), # Label mapping preview
|
131 |
-
gr.update(value=column_mapping, visible=True, interactive=True), # Column mapping
|
132 |
)
|
133 |
elif id2label_df is None:
|
134 |
gr.Warning('The prediction result does not conform the labels in the dataset. Please provide label mappings in "Advance" settings.')
|
135 |
return (
|
136 |
-
config, split,
|
137 |
gr.update(interactive=False), # Submit button
|
|
|
|
|
|
|
138 |
gr.update(value=prediction_result, visible=True), # Model prediction preview
|
139 |
gr.update(visible=False), # Label mapping preview
|
140 |
-
gr.update(value=column_mapping, visible=True, interactive=True), # Column mapping
|
141 |
)
|
142 |
|
143 |
gr.Info("Model and dataset validations passed. Your can submit the evaluation task.")
|
144 |
|
145 |
return (
|
|
|
146 |
gr.update(visible=False), # Loading row
|
147 |
gr.update(visible=True), # Preview row
|
148 |
-
gr.update(
|
149 |
gr.update(value=prediction_result, visible=True), # Model prediction preview
|
150 |
-
gr.update(value=id2label_df, visible=True), # Label mapping preview
|
151 |
-
gr.update(value=column_mapping, visible=True, interactive=True), # Column mapping
|
152 |
)
|
153 |
|
154 |
|
@@ -200,10 +209,7 @@ def try_submit(m_id, d_id, config, split, column_mappings, local):
|
|
200 |
|
201 |
with gr.Blocks(theme=theme) as iface:
|
202 |
with gr.Tab("Text Classification"):
|
203 |
-
global_ds_id = gr.State('ds')
|
204 |
-
|
205 |
def check_dataset_and_get_config(dataset_id):
|
206 |
-
global_ds_id.value = dataset_id
|
207 |
try:
|
208 |
configs = datasets.get_dataset_config_names(dataset_id)
|
209 |
print(configs)
|
@@ -212,10 +218,9 @@ with gr.Blocks(theme=theme) as iface:
|
|
212 |
# Dataset may not exist
|
213 |
pass
|
214 |
|
215 |
-
def check_dataset_and_get_split(
|
216 |
-
print('choice: ',choice, global_ds_id.value)
|
217 |
try:
|
218 |
-
splits = list(datasets.load_dataset(
|
219 |
print('splits: ',splits)
|
220 |
return gr.Dropdown(splits, value=splits[0], visible=True)
|
221 |
except Exception as e:
|
@@ -223,12 +228,20 @@ with gr.Blocks(theme=theme) as iface:
|
|
223 |
print(e)
|
224 |
pass
|
225 |
|
226 |
-
def gate_validate_btn(model_id, dataset_id, dataset_config, dataset_split):
|
227 |
print('model_id: ',model_id)
|
|
|
|
|
|
|
|
|
228 |
if model_id and dataset_id and dataset_config and dataset_split:
|
229 |
-
return
|
230 |
else:
|
231 |
-
return gr.update(interactive=False)
|
|
|
|
|
|
|
|
|
232 |
|
233 |
with gr.Row():
|
234 |
model_id_input = gr.Textbox(
|
@@ -245,22 +258,10 @@ with gr.Blocks(theme=theme) as iface:
|
|
245 |
dataset_split_input = gr.Dropdown(['default'], value=['default'], label='Dataset Split', visible=False)
|
246 |
|
247 |
dataset_id_input.change(check_dataset_and_get_config, dataset_id_input, dataset_config_input)
|
248 |
-
dataset_config_input.change(
|
249 |
-
|
250 |
-
|
251 |
-
|
252 |
-
model_id_input.change(gate_validate_btn,
|
253 |
-
inputs=[model_id_input, dataset_id_input, dataset_config_input, dataset_split_input],
|
254 |
-
outputs=[validate_btn])
|
255 |
-
dataset_id_input.change(gate_validate_btn,
|
256 |
-
inputs=[model_id_input, dataset_id_input, dataset_config_input, dataset_split_input],
|
257 |
-
outputs=[validate_btn])
|
258 |
-
dataset_config_input.change(gate_validate_btn,
|
259 |
-
inputs=[model_id_input, dataset_id_input, dataset_config_input, dataset_split_input],
|
260 |
-
outputs=[validate_btn])
|
261 |
-
dataset_split_input.change(gate_validate_btn,
|
262 |
-
inputs=[model_id_input, dataset_id_input, dataset_config_input, dataset_split_input],
|
263 |
-
outputs=[validate_btn])
|
264 |
|
265 |
with gr.Row(visible=True) as loading_row:
|
266 |
gr.Markdown('''
|
@@ -270,51 +271,45 @@ with gr.Blocks(theme=theme) as iface:
|
|
270 |
''')
|
271 |
|
272 |
with gr.Row(visible=False) as preview_row:
|
273 |
-
|
274 |
-
|
275 |
-
|
276 |
-
|
277 |
-
|
278 |
-
|
279 |
-
|
280 |
-
|
281 |
-
|
282 |
-
|
283 |
-
|
284 |
-
|
285 |
-
|
286 |
-
|
287 |
-
|
288 |
-
label="Column mapping",
|
289 |
-
placeholder="Description of mapping of columns in model to dataset, in json format, e.g.:\n"
|
290 |
-
'{\n'
|
291 |
-
' "text": "context",\n'
|
292 |
-
' "label": {0: "Positive", 1: "Negative"}\n'
|
293 |
-
'}',
|
294 |
-
)
|
295 |
|
|
|
296 |
run_btn = gr.Button(
|
297 |
"Get Evaluation Result",
|
298 |
variant="primary",
|
299 |
interactive=False,
|
|
|
300 |
)
|
301 |
-
|
302 |
-
|
303 |
-
|
304 |
-
|
305 |
-
|
306 |
-
|
307 |
-
|
308 |
-
|
309 |
-
|
310 |
-
|
311 |
-
|
312 |
-
|
313 |
-
|
314 |
-
|
315 |
-
|
316 |
-
|
317 |
-
)
|
318 |
|
319 |
run_btn.click(
|
320 |
try_submit,
|
@@ -323,8 +318,6 @@ with gr.Blocks(theme=theme) as iface:
|
|
323 |
dataset_id_input,
|
324 |
dataset_config_input,
|
325 |
dataset_split_input,
|
326 |
-
column_mapping_input,
|
327 |
-
run_local,
|
328 |
],
|
329 |
outputs=[
|
330 |
run_btn,
|
|
|
59 |
return dataset_id, None, None
|
60 |
return dataset_id, dataset_config, dataset_split
|
61 |
|
62 |
+
def try_validate(model_id, dataset_id, dataset_config, dataset_split, column_mapping='{}'):
|
63 |
# Validate model
|
64 |
m_id, ppl = check_model(model_id=model_id)
|
65 |
if m_id is None:
|
66 |
gr.Warning(f'Model "{model_id}" is not accessible. Please set your HF_TOKEN if it is a private model.')
|
67 |
return (
|
|
|
68 |
gr.update(interactive=False), # Submit button
|
69 |
+
gr.update(visible=True), # Loading row
|
70 |
+
gr.update(visible=False), # Preview row
|
71 |
+
gr.update(visible=False), # Model prediction input
|
72 |
gr.update(visible=False), # Model prediction preview
|
73 |
gr.update(visible=False), # Label mapping preview
|
|
|
74 |
)
|
75 |
if isinstance(ppl, Exception):
|
76 |
gr.Warning(f'Failed to load "{model_id} model": {ppl}')
|
77 |
return (
|
|
|
78 |
gr.update(interactive=False), # Submit button
|
79 |
+
gr.update(visible=True), # Loading row
|
80 |
+
gr.update(visible=False), # Preview row
|
81 |
+
gr.update(visible=False), # Model prediction input
|
82 |
gr.update(visible=False), # Model prediction preview
|
83 |
gr.update(visible=False), # Label mapping preview
|
|
|
84 |
)
|
85 |
|
86 |
# Validate dataset
|
|
|
100 |
|
101 |
if not dataset_ok:
|
102 |
return (
|
|
|
103 |
gr.update(interactive=False), # Submit button
|
104 |
+
gr.update(visible=True), # Loading row
|
105 |
+
gr.update(visible=False), # Preview row
|
106 |
+
gr.update(visible=False), # Model prediction input
|
107 |
gr.update(visible=False), # Model prediction preview
|
108 |
gr.update(visible=False), # Label mapping preview
|
109 |
+
# gr.update(visible=True), # Column mapping
|
110 |
)
|
111 |
|
112 |
# TODO: Validate column mapping by running once
|
|
|
114 |
id2label_df = None
|
115 |
if isinstance(ppl, TextClassificationPipeline):
|
116 |
try:
|
117 |
+
print('validating phase, ', column_mapping)
|
118 |
column_mapping = json.loads(column_mapping)
|
119 |
except Exception:
|
120 |
column_mapping = {}
|
121 |
|
122 |
+
column_mapping, prediction_input, prediction_result, id2label_df = \
|
123 |
text_classification_fix_column_mapping(column_mapping, ppl, d_id, config, split)
|
124 |
|
125 |
column_mapping = json.dumps(column_mapping, indent=2)
|
|
|
129 |
if prediction_result is None:
|
130 |
gr.Warning('The model failed to predict with the first row in the dataset. Please provide column mappings in "Advance" settings.')
|
131 |
return (
|
|
|
132 |
gr.update(interactive=False), # Submit button
|
133 |
+
gr.update(visible=True), # Loading row
|
134 |
+
gr.update(visible=False), # Preview row
|
135 |
+
gr.update(visible=False), # Model prediction input
|
136 |
gr.update(visible=False), # Model prediction preview
|
137 |
gr.update(visible=False), # Label mapping preview
|
138 |
+
# gr.update(value=column_mapping, visible=True, interactive=True), # Column mapping
|
139 |
)
|
140 |
elif id2label_df is None:
|
141 |
gr.Warning('The prediction result does not conform the labels in the dataset. Please provide label mappings in "Advance" settings.')
|
142 |
return (
|
|
|
143 |
gr.update(interactive=False), # Submit button
|
144 |
+
gr.update(visible=False), # Loading row
|
145 |
+
gr.update(visible=True), # Preview row
|
146 |
+
gr.update(value=f'**Sample Input**: {prediction_input}', visible=True), # Model prediction input
|
147 |
gr.update(value=prediction_result, visible=True), # Model prediction preview
|
148 |
gr.update(visible=False), # Label mapping preview
|
149 |
+
# gr.update(value=column_mapping, visible=True, interactive=True), # Column mapping
|
150 |
)
|
151 |
|
152 |
gr.Info("Model and dataset validations passed. Your can submit the evaluation task.")
|
153 |
|
154 |
return (
|
155 |
+
gr.update(interactive=True), # Submit button
|
156 |
gr.update(visible=False), # Loading row
|
157 |
gr.update(visible=True), # Preview row
|
158 |
+
gr.update(value=f'**Sample Input**: {prediction_input}', visible=True), # Model prediction input
|
159 |
gr.update(value=prediction_result, visible=True), # Model prediction preview
|
160 |
+
gr.update(value=id2label_df, visible=True, interactive=True), # Label mapping preview
|
|
|
161 |
)
|
162 |
|
163 |
|
|
|
209 |
|
210 |
with gr.Blocks(theme=theme) as iface:
|
211 |
with gr.Tab("Text Classification"):
|
|
|
|
|
212 |
def check_dataset_and_get_config(dataset_id):
|
|
|
213 |
try:
|
214 |
configs = datasets.get_dataset_config_names(dataset_id)
|
215 |
print(configs)
|
|
|
218 |
# Dataset may not exist
|
219 |
pass
|
220 |
|
221 |
+
def check_dataset_and_get_split(dataset_config, dataset_id):
|
|
|
222 |
try:
|
223 |
+
splits = list(datasets.load_dataset(dataset_id, dataset_config).keys())
|
224 |
print('splits: ',splits)
|
225 |
return gr.Dropdown(splits, value=splits[0], visible=True)
|
226 |
except Exception as e:
|
|
|
228 |
print(e)
|
229 |
pass
|
230 |
|
231 |
+
def gate_validate_btn(model_id, dataset_id, dataset_config, dataset_split, id2label_mapping_dataframe=None):
|
232 |
print('model_id: ',model_id)
|
233 |
+
column_mapping = '{}'
|
234 |
+
if id2label_mapping_dataframe is not None:
|
235 |
+
column_mapping = id2label_mapping_dataframe.to_json(orient="split")
|
236 |
+
print(column_mapping)
|
237 |
if model_id and dataset_id and dataset_config and dataset_split:
|
238 |
+
return try_validate(model_id, dataset_id, dataset_config, dataset_split, column_mapping)
|
239 |
else:
|
240 |
+
return (gr.update(interactive=False),
|
241 |
+
gr.update(visible=True),
|
242 |
+
gr.update(visible=False),
|
243 |
+
gr.update(visible=False),
|
244 |
+
gr.update(visible=False))
|
245 |
|
246 |
with gr.Row():
|
247 |
model_id_input = gr.Textbox(
|
|
|
258 |
dataset_split_input = gr.Dropdown(['default'], value=['default'], label='Dataset Split', visible=False)
|
259 |
|
260 |
dataset_id_input.change(check_dataset_and_get_config, dataset_id_input, dataset_config_input)
|
261 |
+
dataset_config_input.change(
|
262 |
+
check_dataset_and_get_split,
|
263 |
+
inputs=[dataset_config_input, dataset_id_input],
|
264 |
+
outputs=[dataset_split_input])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
265 |
|
266 |
with gr.Row(visible=True) as loading_row:
|
267 |
gr.Markdown('''
|
|
|
271 |
''')
|
272 |
|
273 |
with gr.Row(visible=False) as preview_row:
|
274 |
+
gr.Markdown('''
|
275 |
+
<h1 style="text-align: center;">
|
276 |
+
Confirm Label Details
|
277 |
+
</h1>
|
278 |
+
Base on your model and dataset, we inferred this label mapping. **If the mapping is incorrect, please modify it in the table below.**
|
279 |
+
''')
|
280 |
+
|
281 |
+
with gr.Row():
|
282 |
+
id2label_mapping_dataframe = gr.DataFrame(label="Preview of label mapping", interactive=True, visible=False)
|
283 |
+
|
284 |
+
with gr.Row():
|
285 |
+
example_input = gr.Markdown('Sample Input: ', visible=False)
|
286 |
+
|
287 |
+
with gr.Row():
|
288 |
+
example_labels = gr.Label(label='Model Prediction Sample', visible=False)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
289 |
|
290 |
+
|
291 |
run_btn = gr.Button(
|
292 |
"Get Evaluation Result",
|
293 |
variant="primary",
|
294 |
interactive=False,
|
295 |
+
size="lg",
|
296 |
)
|
297 |
+
|
298 |
+
model_id_input.change(gate_validate_btn,
|
299 |
+
inputs=[model_id_input, dataset_id_input, dataset_config_input, dataset_split_input],
|
300 |
+
outputs=[run_btn, loading_row, preview_row, example_input, example_labels, id2label_mapping_dataframe])
|
301 |
+
dataset_id_input.change(gate_validate_btn,
|
302 |
+
inputs=[model_id_input, dataset_id_input, dataset_config_input, dataset_split_input],
|
303 |
+
outputs=[run_btn, loading_row, preview_row, example_input, example_labels, id2label_mapping_dataframe])
|
304 |
+
dataset_config_input.change(gate_validate_btn,
|
305 |
+
inputs=[model_id_input, dataset_id_input, dataset_config_input, dataset_split_input],
|
306 |
+
outputs=[run_btn, loading_row, preview_row, example_input, example_labels, id2label_mapping_dataframe])
|
307 |
+
dataset_split_input.change(gate_validate_btn,
|
308 |
+
inputs=[model_id_input, dataset_id_input, dataset_config_input, dataset_split_input],
|
309 |
+
outputs=[run_btn, loading_row, preview_row, example_input, example_labels, id2label_mapping_dataframe])
|
310 |
+
id2label_mapping_dataframe.input(gate_validate_btn,
|
311 |
+
inputs=[model_id_input, dataset_id_input, dataset_config_input, dataset_split_input, id2label_mapping_dataframe],
|
312 |
+
outputs=[run_btn, loading_row, preview_row, example_input, example_labels, id2label_mapping_dataframe])
|
|
|
313 |
|
314 |
run_btn.click(
|
315 |
try_submit,
|
|
|
318 |
dataset_id_input,
|
319 |
dataset_config_input,
|
320 |
dataset_split_input,
|
|
|
|
|
321 |
],
|
322 |
outputs=[
|
323 |
run_btn,
|
text_classification.py
CHANGED
@@ -72,10 +72,12 @@ def text_classification_fix_column_mapping(column_mapping, ppl, d_id, config, sp
|
|
72 |
id2label_mapping = {}
|
73 |
id2label = ppl.model.config.id2label
|
74 |
label2id = {v: k for k, v in id2label.items()}
|
|
|
75 |
prediction_result = None
|
76 |
try:
|
77 |
# Use the first item to test prediction
|
78 |
-
|
|
|
79 |
prediction_result = {
|
80 |
f'{result["label"]}({label2id[result["label"]]})': result["score"] for result in results
|
81 |
}
|
@@ -85,33 +87,28 @@ def text_classification_fix_column_mapping(column_mapping, ppl, d_id, config, sp
|
|
85 |
|
86 |
# Infer labels
|
87 |
id2label_mapping, dataset_labels = text_classification_map_model_and_dataset_labels(id2label, dataset_features)
|
88 |
-
if "
|
89 |
-
if
|
90 |
-
logging.warning(f'Provided {column_mapping["label"]} does not match labels in Dataset')
|
91 |
-
return column_mapping, prediction_result, None
|
92 |
-
|
93 |
-
if isinstance(column_mapping["label"], dict):
|
94 |
# Use the column mapping passed by user
|
95 |
-
for
|
96 |
-
|
|
|
97 |
elif None in id2label_mapping.values():
|
98 |
column_mapping["label"] = {
|
99 |
i: None for i in id2label.keys()
|
100 |
}
|
101 |
return column_mapping, prediction_result, None
|
102 |
|
103 |
-
id2label_mapping
|
104 |
-
v: k for k, v in id2label_mapping.items()
|
105 |
-
}
|
106 |
id2label_df = pd.DataFrame({
|
107 |
-
"
|
108 |
-
"Labels": dataset_labels,
|
109 |
-
"Labels in original model": [f"{id2label_mapping[label]}({label2id[id2label_mapping[label]]})" for label in dataset_labels],
|
110 |
})
|
111 |
-
|
|
|
112 |
# Column mapping should contain original model labels
|
113 |
column_mapping["label"] = {
|
114 |
str(i): id2label_mapping[label] for i, label in zip(id2label.keys(), dataset_labels)
|
115 |
}
|
116 |
|
117 |
-
return column_mapping, prediction_result, id2label_df
|
|
|
72 |
id2label_mapping = {}
|
73 |
id2label = ppl.model.config.id2label
|
74 |
label2id = {v: k for k, v in id2label.items()}
|
75 |
+
prediction_input = None
|
76 |
prediction_result = None
|
77 |
try:
|
78 |
# Use the first item to test prediction
|
79 |
+
prediction_input = df.head(1).at[0, column_mapping["text"]]
|
80 |
+
results = ppl({"text": prediction_input}, top_k=None)
|
81 |
prediction_result = {
|
82 |
f'{result["label"]}({label2id[result["label"]]})': result["score"] for result in results
|
83 |
}
|
|
|
87 |
|
88 |
# Infer labels
|
89 |
id2label_mapping, dataset_labels = text_classification_map_model_and_dataset_labels(id2label, dataset_features)
|
90 |
+
if "data" in column_mapping.keys():
|
91 |
+
if isinstance(column_mapping["data"], list):
|
|
|
|
|
|
|
|
|
92 |
# Use the column mapping passed by user
|
93 |
+
for user_label, model_label in column_mapping["data"]:
|
94 |
+
print(user_label, model_label)
|
95 |
+
id2label_mapping[model_label] = user_label
|
96 |
elif None in id2label_mapping.values():
|
97 |
column_mapping["label"] = {
|
98 |
i: None for i in id2label.keys()
|
99 |
}
|
100 |
return column_mapping, prediction_result, None
|
101 |
|
102 |
+
print(id2label_mapping)
|
|
|
|
|
103 |
id2label_df = pd.DataFrame({
|
104 |
+
"Dataset Labels": dataset_labels,
|
105 |
+
"Model Prediction Labels": [id2label_mapping[label] for label in dataset_labels],
|
|
|
106 |
})
|
107 |
+
|
108 |
+
if "data" not in column_mapping.keys():
|
109 |
# Column mapping should contain original model labels
|
110 |
column_mapping["label"] = {
|
111 |
str(i): id2label_mapping[label] for i, label in zip(id2label.keys(), dataset_labels)
|
112 |
}
|
113 |
|
114 |
+
return column_mapping, prediction_input, prediction_result, id2label_df
|