Luis J Camargo commited on
Commit
7f30859
·
1 Parent(s): f72e6b3

feat(ui): single ranked prediction list, metadata-only family/superlanguage

Browse files

The model emits one softmax over 128 language codes. Family and
superlanguage were never separate heads, so per-group percentages were
just re-aggregations of the same distribution and misleading. Drop those
tables and present a single ranked list instead.

- Results are now rendered as pills: language name + code, with
superlanguage and family as small italic context underneath. The top
match is highlighted (larger, accented), the rest are muted.
- No table header row, so the bilingual "Prediction / Predicción"
headers are gone; the section title comes from i18n instead.
- Renamed i18n keys to the ones actually used; removed the family,
superlanguage and code-table keys.
- Advanced Options collapsed from three groups to one (Top-K + threshold
for the ranked list).
- Subtitle now states the real scope: this model covers N of the 364
catalogue variants of the 68 indigenous languages of Mexico, instead
of implying Mexico has N variants.
- Curated cards keep only the ground-truth language label.
- requirements.txt: pandas no longer needed now that Dataframe is gone.

Files changed (3) hide show
  1. README.md +5 -3
  2. app.py +82 -88
  3. requirements.txt +0 -1
README.md CHANGED
@@ -18,7 +18,8 @@ Identify language family, superlanguage, and specific language code from audio f
18
  ## 🚀 How to use
19
  1. Upload an audio file (WAV or MP3) or use the microphone to record.
20
  2. Click **Classify Language**.
21
- 3. View the results in the hierarchical classification dashboard.
 
22
 
23
  ## 🧠 What changed in 2.0
24
  Version 1 used a Whisper encoder-only classifier. Version 2 uses a much smaller
@@ -29,8 +30,9 @@ parameters (~85 MB) which makes it viable on CPU-only hardware, where the Space
29
  ## 📊 Model Details
30
  - **Architecture**: ECAPA-TDNN (SpeechBrain) + cosine head, temperature-calibrated
31
  - **Input**: 16 kHz mono audio, 6 s sliding windows for clips longer than 10 s
32
- - **Output**: 128 ISO 639-3 language codes; family and superlanguage are derived from
33
- the label metadata, not predicted by separate heads
 
34
  - **Scope**: indigenous languages of Mexico (Mesoamerica)
35
  - **Metrics** (test split, calibrated):
36
  - Family: ~96.6%
 
18
  ## 🚀 How to use
19
  1. Upload an audio file (WAV or MP3) or use the microphone to record.
20
  2. Click **Classify Language**.
21
+ 3. View the ranked predictions: the best match is highlighted, with its superlanguage and
22
+ family shown as metadata. Top-K and the confidence threshold live in *Advanced Options*.
23
 
24
  ## 🧠 What changed in 2.0
25
  Version 1 used a Whisper encoder-only classifier. Version 2 uses a much smaller
 
30
  ## 📊 Model Details
31
  - **Architecture**: ECAPA-TDNN (SpeechBrain) + cosine head, temperature-calibrated
32
  - **Input**: 16 kHz mono audio, 6 s sliding windows for clips longer than 10 s
33
+ - **Output**: 128 ISO 639-3 language codes. The model predicts the *code only*;
34
+ superlanguage and family are attached from the label metadata as context, not predicted
35
+ separately, so no separate percentages are shown for them.
36
  - **Scope**: indigenous languages of Mexico (Mesoamerica)
37
  - **Metrics** (test split, calibrated):
38
  - Family: ~96.6%
app.py CHANGED
@@ -1,13 +1,13 @@
1
  import os
2
  import gc
3
  import base64
 
4
  import json
5
  import re
6
 
7
  import gradio as gr
8
  import librosa
9
  import numpy as np
10
- import pandas as pd
11
  import torch
12
  import psutil
13
 
@@ -15,62 +15,30 @@ from lid_ecapa import TachiwinLID, LID_MODEL_REPO
15
 
16
  i18n = gr.I18n(
17
  en={
18
- "subtitle": "Identify the indigenous language variants of Mexico covered by this model",
19
  "input_audio_header": "### 🎙️ 1. Input Audio",
20
  "upload_or_record_label": "Upload or Record",
21
  "advanced_options": "⚙️ Advanced Options",
22
- "language_family": "#### Language Family",
23
  "top_k_label": "Top-K",
24
  "threshold_label": "Threshold",
25
- "superlanguage": "#### Superlanguage",
26
- "language_code": "#### Language Code",
27
  "clear_btn": "🗑️ Clear",
28
  "classify_btn": "🚀 Classify",
29
  "results_header": "### 📊 2. Classification Results",
30
- "pred_fam_label": "🌍 Language Family",
31
- "pred_super_label": "🗣️ Superlanguage",
32
- "pred_code_label": "🔤 Language Code",
33
  "curated_header": "### 📋 Curated Samples for Reference",
34
  "curated_desc": "Select any of the sample results below to load the corresponding audio for inference.",
35
- "prediction_header": "Prediction",
36
- "confidence_header": "Confidence",
37
- "col_sample": "🔊 Sample",
38
- "col_pred_fam": "Pred Family",
39
- "col_pred_super": "Pred Super",
40
- "col_pred_code": "Pred Code",
41
- "col_act_fam": "Actual Family",
42
- "col_act_super": "Actual Super",
43
- "col_act_code": "Actual Code",
44
- "col_match": "Match"
45
  },
46
  es={
47
- "subtitle": "Identifica las variantes de lenguas originarias de México cubiertas por este modelo",
48
  "input_audio_header": "### 🎙️ 1. Entrada de Audio",
49
  "upload_or_record_label": "Subir o Grabar",
50
  "advanced_options": "⚙️ Opciones Avanzadas",
51
- "language_family": "#### Familia Lingüística",
52
  "top_k_label": "Top-K",
53
  "threshold_label": "Umbral",
54
- "superlanguage": "#### Superlengua",
55
- "language_code": "#### Código de Lengua",
56
  "clear_btn": "🗑️ Limpiar",
57
  "classify_btn": "🚀 Clasificar",
58
  "results_header": "### 📊 2. Resultados de Clasificación",
59
- "pred_fam_label": "🌍 Familia Lingüística",
60
- "pred_super_label": "🗣️ Superlengua",
61
- "pred_code_label": "🔤 Código de Lengua",
62
  "curated_header": "### 📋 Muestras Curadas de Referencia",
63
  "curated_desc": "Seleccione cualquiera de los resultados de muestra a continuación para cargar el audio correspondiente para la clasificación.",
64
- "prediction_header": "Predicción",
65
- "confidence_header": "Confianza",
66
- "col_sample": "🔊 Muestra",
67
- "col_pred_fam": "Familia Pred",
68
- "col_pred_super": "Super Pred",
69
- "col_pred_code": "Cód Pred",
70
- "col_act_fam": "Familia Real",
71
- "col_act_super": "Super Real",
72
- "col_act_code": "Cód Real",
73
- "col_match": "Coincidencia"
74
  }
75
  )
76
 
@@ -79,6 +47,10 @@ MAX_AUDIO_SECONDS = 30
79
  SAMPLE_RATE = 16000
80
  UNKNOWN_GROUP = "Sin especificar / Unspecified"
81
 
 
 
 
 
82
  torch.set_num_threads(1)
83
 
84
  # === LOAD MODEL ===
@@ -91,7 +63,8 @@ INFO = lid.info
91
  N_CLASSES = len(CLASSES)
92
 
93
  # === LABEL METADATA ===
94
- # languages.json (INALI table) is kept as a secondary source of names / family / superlanguage.
 
95
  try:
96
  with open("languages.json", "r", encoding="utf-8") as f:
97
  _languages_data = json.load(f)
@@ -111,7 +84,7 @@ except Exception as e:
111
  CODE_TO_META = {}
112
 
113
  # Manual completion for entries where labels.json ships an empty family/superlanguage.
114
- # (Metadata gaps in the model repo -- purely display-side, probabilities are unaffected.)
115
  META_FALLBACK = {
116
  "chq": ("Otomangue", "Chinanteco"),
117
  "hue": ("Huave", "Huave"),
@@ -127,7 +100,7 @@ META_FALLBACK = {
127
 
128
 
129
  def _meta(code):
130
- """(family, superlanguage, language_name) for a code, with graceful fallbacks."""
131
  entry = INFO.get(code, {})
132
  family = entry.get("family") or ""
133
  superlanguage = entry.get("superlanguage") or ""
@@ -140,17 +113,16 @@ def _meta(code):
140
  family = family or db_family or UNKNOWN_GROUP
141
  superlanguage = superlanguage or db_super or UNKNOWN_GROUP
142
  name = entry.get("language") or CODE_TO_NAME.get(code, code)
143
- return family, superlanguage, name
144
 
145
 
146
- # code -> group, used to marginalise the code distribution into family / superlanguage
147
- FAMILY_OF = {}
148
  SUPERLANGUAGE_OF = {}
 
149
  DISPLAY_NAME = {}
150
  for _code in CLASSES:
151
- _family, _super, _name = _meta(_code)
152
- FAMILY_OF[_code] = _family
153
  SUPERLANGUAGE_OF[_code] = _super
 
154
  DISPLAY_NAME[_code] = _name
155
 
156
  print(
@@ -164,22 +136,61 @@ def get_mem_usage():
164
  return process.memory_info().rss / (1024 ** 2)
165
 
166
 
167
- def _marginalize(probs, group_of):
168
- """Sum the per-code probabilities into group probabilities (keeps 1.0 total mass)."""
169
- totals = {}
170
- for code, p in zip(CLASSES, probs):
171
- group = group_of[code]
172
- totals[group] = totals.get(group, 0.0) + float(p)
173
- return sorted(totals.items(), key=lambda kv: kv[1], reverse=True)
174
-
175
-
176
- def _format_table(rows, k, thresh):
177
- table_data = [[name, f"{score:.2%}"] for name, score in rows[: int(k)] if score >= thresh]
178
- return pd.DataFrame(table_data, columns=["Prediction / Predicción", "Confidence / Confianza"])
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
179
 
180
 
181
  # === INFERENCE FUNCTION ===
182
- def predict_language(audio_path, fam_k=1, fam_thresh=0.0, super_k=1, super_thresh=0.0, code_k=3, code_thresh=0.0):
183
  if not audio_path:
184
  raise gr.Error("No audio provided! Please upload or record an audio file.")
185
 
@@ -211,20 +222,14 @@ def predict_language(audio_path, fam_k=1, fam_thresh=0.0, super_k=1, super_thres
211
  print(f"[LOG] Memory after inference: {get_mem_usage():.2f} MB")
212
 
213
  print("[LOG] Step 3: Post-processing results...")
214
- df_fam = _format_table(_marginalize(probs, FAMILY_OF), fam_k, fam_thresh)
215
- df_super = _format_table(_marginalize(probs, SUPERLANGUAGE_OF), super_k, super_thresh)
216
-
217
- code_rows = sorted(
218
- ((f"{DISPLAY_NAME[code]} ({code})", float(p)) for code, p in zip(CLASSES, probs)),
219
- key=lambda kv: kv[1],
220
- reverse=True,
221
- )
222
- df_code = _format_table(code_rows, code_k, code_thresh)
223
 
224
  print(f"[LOG] Final Memory: {get_mem_usage():.2f} MB")
225
  print("--- [LOG] Request Finished ---\n")
226
 
227
- return df_fam, df_super, df_code
228
  except gr.Error:
229
  raise
230
  except Exception as e:
@@ -273,12 +278,13 @@ except Exception as e:
273
  # === UI COMPONENTS ===
274
  with gr.Blocks(theme=gr.themes.Ocean()) as demo:
275
  gr.HTML(
276
- f"""
 
277
  <div style="text-align: center; padding: 30px; background: linear-gradient(120deg, rgb(2, 132, 199) 0%, rgb(16, 185, 129) 60%, rgb(5, 150, 105) 100%); color: white; border-radius: 15px; margin-bottom: 25px; box-shadow: 0 4px 6px -1px rgba(0, 0, 0, 0.1);">
278
  {icon_html}
279
  <h1 style="color: white; margin: 0; font-size: 2.5em;">🦡 Tachiwin AudioId 2.0 🦡</h1>
280
- <p style="font-size: 1.2em; opacity: 0.9; margin-top: 10px;">Identify the {N_CLASSES} indigenous language variants of Mexico covered by this model<br/>
281
- <i>Identifica las {N_CLASSES} variantes de lenguas originarias de México cubiertas por este modelo</i></p>
282
  </div>
283
  """
284
  )
@@ -293,17 +299,7 @@ with gr.Blocks(theme=gr.themes.Ocean()) as demo:
293
  )
294
  with gr.Accordion(i18n("advanced_options"), open=False):
295
  with gr.Group():
296
- gr.Markdown(i18n("language_family"))
297
- with gr.Row():
298
- fam_k = gr.Slider(minimum=1, maximum=10, step=1, value=1, label=i18n("top_k_label"))
299
- fam_thresh = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, value=0.0, label=i18n("threshold_label"))
300
- with gr.Group():
301
- gr.Markdown(i18n("superlanguage"))
302
- with gr.Row():
303
- super_k = gr.Slider(minimum=1, maximum=10, step=1, value=1, label=i18n("top_k_label"))
304
- super_thresh = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, value=0.0, label=i18n("threshold_label"))
305
- with gr.Group():
306
- gr.Markdown(i18n("language_code"))
307
  with gr.Row():
308
  code_k = gr.Slider(minimum=1, maximum=10, step=1, value=3, label=i18n("top_k_label"))
309
  code_thresh = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, value=0.0, label=i18n("threshold_label"))
@@ -314,20 +310,18 @@ with gr.Blocks(theme=gr.themes.Ocean()) as demo:
314
 
315
  with gr.Column(scale=1):
316
  gr.Markdown(i18n("results_header"))
317
- fam_table = gr.Dataframe(headers=["Prediction / Predicción", "Confidence / Confianza"], datatype=["str", "str"], label=i18n("pred_fam_label"), interactive=False, wrap=True)
318
- super_table = gr.Dataframe(headers=["Prediction / Predicción", "Confidence / Confianza"], datatype=["str", "str"], label=i18n("pred_super_label"), interactive=False, wrap=True)
319
- code_table = gr.Dataframe(headers=["Prediction / Predicción", "Confidence / Confianza"], datatype=["str", "str"], label=i18n("pred_code_label"), interactive=False, wrap=True)
320
 
321
  submit_btn.click(
322
  fn=predict_language,
323
- inputs=[audio_input, fam_k, fam_thresh, super_k, super_thresh, code_k, code_thresh],
324
- outputs=[fam_table, super_table, code_table]
325
  )
326
 
327
  clear_btn.click(
328
- fn=lambda: (None, None, None, None),
329
  inputs=None,
330
- outputs=[audio_input, fam_table, super_table, code_table]
331
  )
332
 
333
  gr.Markdown(i18n("curated_header"))
@@ -339,7 +333,7 @@ with gr.Blocks(theme=gr.themes.Ocean()) as demo:
339
  with gr.Row():
340
  for i in range(row_idx, min(row_idx+columns, len(curated_audio))):
341
  with gr.Column(scale=1):
342
- gr.Markdown(f"```text\n{curated_labels[i]}\n```")
343
  gr.Audio(value=curated_audio[i], interactive=False, show_label=False)
344
  btn = gr.Button("Load / Cargar", size="sm")
345
  btn.click(fn=lambda url=curated_audio[i]: url, inputs=[], outputs=[audio_input])
 
1
  import os
2
  import gc
3
  import base64
4
+ import html
5
  import json
6
  import re
7
 
8
  import gradio as gr
9
  import librosa
10
  import numpy as np
 
11
  import torch
12
  import psutil
13
 
 
15
 
16
  i18n = gr.I18n(
17
  en={
 
18
  "input_audio_header": "### 🎙️ 1. Input Audio",
19
  "upload_or_record_label": "Upload or Record",
20
  "advanced_options": "⚙️ Advanced Options",
21
+ "prediction_options": "#### Ranked Predictions",
22
  "top_k_label": "Top-K",
23
  "threshold_label": "Threshold",
 
 
24
  "clear_btn": "🗑️ Clear",
25
  "classify_btn": "🚀 Classify",
26
  "results_header": "### 📊 2. Classification Results",
 
 
 
27
  "curated_header": "### 📋 Curated Samples for Reference",
28
  "curated_desc": "Select any of the sample results below to load the corresponding audio for inference.",
 
 
 
 
 
 
 
 
 
 
29
  },
30
  es={
 
31
  "input_audio_header": "### 🎙️ 1. Entrada de Audio",
32
  "upload_or_record_label": "Subir o Grabar",
33
  "advanced_options": "⚙️ Opciones Avanzadas",
34
+ "prediction_options": "#### Predicciones Ordenadas",
35
  "top_k_label": "Top-K",
36
  "threshold_label": "Umbral",
 
 
37
  "clear_btn": "🗑️ Limpiar",
38
  "classify_btn": "🚀 Clasificar",
39
  "results_header": "### 📊 2. Resultados de Clasificación",
 
 
 
40
  "curated_header": "### 📋 Muestras Curadas de Referencia",
41
  "curated_desc": "Seleccione cualquiera de los resultados de muestra a continuación para cargar el audio correspondiente para la clasificación.",
 
 
 
 
 
 
 
 
 
 
42
  }
43
  )
44
 
 
47
  SAMPLE_RATE = 16000
48
  UNKNOWN_GROUP = "Sin especificar / Unspecified"
49
 
50
+ # INALI catalogue: 68 language groups / 364 variants. Used only for the honest subtitle.
51
+ INALI_VARIANTS = 364
52
+ INALI_LANGUAGES = 68
53
+
54
  torch.set_num_threads(1)
55
 
56
  # === LOAD MODEL ===
 
63
  N_CLASSES = len(CLASSES)
64
 
65
  # === LABEL METADATA ===
66
+ # The model predicts a language code only; superlanguage and family are *metadata*
67
+ # attached to that code, shown as context, never as a separate prediction.
68
  try:
69
  with open("languages.json", "r", encoding="utf-8") as f:
70
  _languages_data = json.load(f)
 
84
  CODE_TO_META = {}
85
 
86
  # Manual completion for entries where labels.json ships an empty family/superlanguage.
87
+ # (Metadata gaps in the model repo -- display-side only, probabilities are unaffected.)
88
  META_FALLBACK = {
89
  "chq": ("Otomangue", "Chinanteco"),
90
  "hue": ("Huave", "Huave"),
 
100
 
101
 
102
  def _meta(code):
103
+ """(superlanguage, family, language_name) for a code, with graceful fallbacks."""
104
  entry = INFO.get(code, {})
105
  family = entry.get("family") or ""
106
  superlanguage = entry.get("superlanguage") or ""
 
113
  family = family or db_family or UNKNOWN_GROUP
114
  superlanguage = superlanguage or db_super or UNKNOWN_GROUP
115
  name = entry.get("language") or CODE_TO_NAME.get(code, code)
116
+ return superlanguage, family, name
117
 
118
 
 
 
119
  SUPERLANGUAGE_OF = {}
120
+ FAMILY_OF = {}
121
  DISPLAY_NAME = {}
122
  for _code in CLASSES:
123
+ _super, _family, _name = _meta(_code)
 
124
  SUPERLANGUAGE_OF[_code] = _super
125
+ FAMILY_OF[_code] = _family
126
  DISPLAY_NAME[_code] = _name
127
 
128
  print(
 
136
  return process.memory_info().rss / (1024 ** 2)
137
 
138
 
139
+ # === RESULT RENDERING ===
140
+ RESULT_STYLES = """
141
+ <style>
142
+ .tw-lid { display: flex; flex-direction: column; gap: 8px; margin-top: 4px; }
143
+ .tw-row { display: flex; align-items: center; justify-content: space-between; gap: 14px;
144
+ padding: 10px 14px; border-radius: 12px;
145
+ background: rgba(2, 132, 199, 0.06);
146
+ border: 1px solid rgba(2, 132, 199, 0.18); }
147
+ .tw-row.tw-first { padding: 14px 16px;
148
+ background: linear-gradient(120deg, rgba(16, 185, 129, 0.14) 0%, rgba(2, 132, 199, 0.10) 100%);
149
+ border: 1px solid rgba(5, 150, 105, 0.45);
150
+ box-shadow: 0 1px 3px rgba(0, 0, 0, 0.08); }
151
+ .tw-main { min-width: 0; }
152
+ .tw-title { display: flex; align-items: baseline; gap: 8px; flex-wrap: wrap; }
153
+ .tw-name { font-size: 1.02rem; font-weight: 600; line-height: 1.25; }
154
+ .tw-first .tw-name { font-size: 1.3rem; }
155
+ .tw-code { font-family: ui-monospace, SFMono-Regular, Menlo, monospace; font-size: 0.72rem;
156
+ letter-spacing: 0.04em; padding: 2px 8px; border-radius: 999px;
157
+ background: rgba(2, 132, 199, 0.14); color: #0369a1; white-space: nowrap; }
158
+ .tw-meta { font-size: 0.74rem; font-style: italic; opacity: 0.72; margin-top: 4px; }
159
+ .tw-conf { font-variant-numeric: tabular-nums; font-weight: 600; font-size: 1rem;
160
+ white-space: nowrap; opacity: 0.85; }
161
+ .tw-first .tw-conf { font-size: 1.4rem; color: #047857; opacity: 1; }
162
+ .dark .tw-code { color: #7dd3fc; }
163
+ .dark .tw-first .tw-conf { color: #6ee7b7; }
164
+ </style>
165
+ """
166
+
167
+
168
+ def render_results(rows):
169
+ """Render ranked predictions as pills. `rows` is [(code, probability), ...]."""
170
+ if not rows:
171
+ return ""
172
+
173
+ blocks = []
174
+ for index, (code, prob) in enumerate(rows):
175
+ css = "tw-row tw-first" if index == 0 else "tw-row"
176
+ blocks.append(
177
+ f'<div class="{css}">'
178
+ f'<div class="tw-main">'
179
+ f'<div class="tw-title">'
180
+ f'<span class="tw-name">{html.escape(DISPLAY_NAME[code])}</span>'
181
+ f'<span class="tw-code">{html.escape(code)}</span>'
182
+ f'</div>'
183
+ f'<div class="tw-meta">{html.escape(SUPERLANGUAGE_OF[code])} · '
184
+ f'{html.escape(FAMILY_OF[code])}</div>'
185
+ f'</div>'
186
+ f'<div class="tw-conf">{prob:.2%}</div>'
187
+ f'</div>'
188
+ )
189
+ return '<div class="tw-lid">' + "".join(blocks) + "</div>"
190
 
191
 
192
  # === INFERENCE FUNCTION ===
193
+ def predict_language(audio_path, code_k=3, code_thresh=0.0):
194
  if not audio_path:
195
  raise gr.Error("No audio provided! Please upload or record an audio file.")
196
 
 
222
  print(f"[LOG] Memory after inference: {get_mem_usage():.2f} MB")
223
 
224
  print("[LOG] Step 3: Post-processing results...")
225
+ order = np.argsort(-probs)[: int(code_k)]
226
+ rows = [(CLASSES[int(i)], float(probs[int(i)])) for i in order]
227
+ rows = [(code, prob) for code, prob in rows if prob >= code_thresh]
 
 
 
 
 
 
228
 
229
  print(f"[LOG] Final Memory: {get_mem_usage():.2f} MB")
230
  print("--- [LOG] Request Finished ---\n")
231
 
232
+ return render_results(rows)
233
  except gr.Error:
234
  raise
235
  except Exception as e:
 
278
  # === UI COMPONENTS ===
279
  with gr.Blocks(theme=gr.themes.Ocean()) as demo:
280
  gr.HTML(
281
+ RESULT_STYLES
282
+ + f"""
283
  <div style="text-align: center; padding: 30px; background: linear-gradient(120deg, rgb(2, 132, 199) 0%, rgb(16, 185, 129) 60%, rgb(5, 150, 105) 100%); color: white; border-radius: 15px; margin-bottom: 25px; box-shadow: 0 4px 6px -1px rgba(0, 0, 0, 0.1);">
284
  {icon_html}
285
  <h1 style="color: white; margin: 0; font-size: 2.5em;">🦡 Tachiwin AudioId 2.0 🦡</h1>
286
+ <p style="font-size: 1.2em; opacity: 0.9; margin-top: 10px;">Identify variants of the {INALI_LANGUAGES} indigenous languages of Mexico — this model covers {N_CLASSES} of the {INALI_VARIANTS} catalogued variants<br/>
287
+ <i>Identifica variantes de las {INALI_LANGUAGES} lenguas originarias de México — este modelo cubre {N_CLASSES} de las {INALI_VARIANTS} variantes catalogadas</i></p>
288
  </div>
289
  """
290
  )
 
299
  )
300
  with gr.Accordion(i18n("advanced_options"), open=False):
301
  with gr.Group():
302
+ gr.Markdown(i18n("prediction_options"))
 
 
 
 
 
 
 
 
 
 
303
  with gr.Row():
304
  code_k = gr.Slider(minimum=1, maximum=10, step=1, value=3, label=i18n("top_k_label"))
305
  code_thresh = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, value=0.0, label=i18n("threshold_label"))
 
310
 
311
  with gr.Column(scale=1):
312
  gr.Markdown(i18n("results_header"))
313
+ results_html = gr.HTML(value="")
 
 
314
 
315
  submit_btn.click(
316
  fn=predict_language,
317
+ inputs=[audio_input, code_k, code_thresh],
318
+ outputs=[results_html]
319
  )
320
 
321
  clear_btn.click(
322
+ fn=lambda: (None, ""),
323
  inputs=None,
324
+ outputs=[audio_input, results_html]
325
  )
326
 
327
  gr.Markdown(i18n("curated_header"))
 
333
  with gr.Row():
334
  for i in range(row_idx, min(row_idx+columns, len(curated_audio))):
335
  with gr.Column(scale=1):
336
+ gr.Markdown(f"**{curated_labels[i]}**")
337
  gr.Audio(value=curated_audio[i], interactive=False, show_label=False)
338
  btn = gr.Button("Load / Cargar", size="sm")
339
  btn.click(fn=lambda url=curated_audio[i]: url, inputs=[], outputs=[audio_input])
requirements.txt CHANGED
@@ -17,6 +17,5 @@ huggingface_hub
17
  soundfile
18
  librosa
19
  numpy
20
- pandas
21
  psutil
22
  gradio
 
17
  soundfile
18
  librosa
19
  numpy
 
20
  psutil
21
  gradio