Update app.py

app.py

@@ -66,86 +66,156 @@ def ClassifyMIDI(input_midi):
 
     print('Done!')
     print('=' * 70)
-    seed_melody = seed_melodies_data[input_melody_seed_number]
-    print('Input melody seed number:', input_melody_seed_number)
-    print('-' * 70)
 
     #==================================================================
 
     print('=' * 70)
-    
-    print('Sample output events', seed_melody[:16])
-    print('=' * 70)
-    print('Generating...')
 
-    x = (torch.tensor(seed_melody, dtype=torch.long, device='cuda')[None, ...])
+    fn = os.path.basename(input_midi.name)
+    fn1 = fn.split('.')[0]
+
+    print('-' * 70)
+    print('Input file name:', fn)
 
-    with ctx:
-      out = model.generate(x,
-                          1536,
-                          temperature=0.9,
-                          return_prime=False,
-                          verbose=False)
-    
-    output = out[0].tolist()
-        
     print('=' * 70)
+    print('Loading MIDI file...')
+    
+    midi_name = fn
+    
+    raw_score = TMIDIX.midi2single_track_ms_score(open(input_midi.name, 'rb').read())
+    
+    escore_notes = TMIDIX.advanced_score_processor(raw_score, return_enhanced_score_notes=True)[0]
+    
+    escore = [e for e in TMIDIX.augment_enhanced_score_notes(escore_notes, timings_divider=32) if e[6] < 80]
+    
+    cscore = TMIDIX.chordify_score([1000, escore])
+    
+    #=======================================================
+    # MAIN PROCESSING CYCLE
+    #=======================================================
+    
+    melody_chords = []
+    
+    pe = cscore[0][0]
+    
+    for c in cscore:
+    
+        pitches = []
+    
+        for e in c:
+    
+          if e[4] not in pitches:
+    
+            dtime = max(0, min(127, e[1]-pe[1]))
+    
+            dur = max(1, min(127, e[2]))
+            ptc = max(1, min(127, e[4]))
+    
+            melody_chords.append([dtime, dur, ptc])
+    
+            pitches.append(ptc)
+    
+            pe = e
+    
+    #==============================================================
+    
+    seq = []
+    input_data = []
+    
+    notes_counter = 0
+    
+    for mm in melody_chords:
+    
+        time = mm[0]
+        dur = mm[1]
+        ptc = mm[2]
+    
+        seq.extend([time, dur+128, ptc+256])
+        notes_counter += 1
+    
+    for i in range(0, len(seq)-SEQ_LEN-4, (SEQ_LEN-4) // 2):
+      schunk = seq[i:i+SEQ_LEN-4]
+      input_data.append([14624] + schunk + [14625])
+    
     print('Done!')
     print('=' * 70)
     
-    #===============================================================================
-    print('Rendering results...')
+    #==============================================================
     
+    print('Composition has', notes_counter, 'notes')
     print('=' * 70)
-    print('Sample INTs', output[:15])
+    print('Composition was split into' , len(input_data), 'chunks', 'of 340 notes each with 170 notes overlap')
+    print('Number of notes in all composition chunks:', len(input_data) * 340)
+
+    number_of_batches = 100 # @param {type:"slider", min:1, max:100, step:1}
+    
+    # @markdown NOTE: You can increase the number of batches on high-ram GPUs for better classification
+    
+    print('=' * 70)
+    print('Annotated MIDI Dataset Classifier')
     print('=' * 70)
+    print('Classifying...')
     
-    out1 = output
-
-    if len(out1) != 0:
+    torch.cuda.empty_cache()
+    
+    model.eval()
     
-        song = out1
-        song_f = []
+    results = []
     
-        time = 0
-        dur = 0
-        vel = 90
-        pitch = 0
-        channel = 0
+    for input in input_data:
     
-        patches = [0] * 16
-        patches[3] = 40
+        x = torch.tensor([input[:1022]] * number_of_batches, dtype=torch.long, device='cuda')
     
-        for ss in song:
+        with ctx:
+          out = model.generate(x,
+                              1,
+                              temperature=0.3,
+                              return_prime=False,
+                              verbose=False)
     
-            if 0 < ss < 128:
+        y = out.tolist()
     
-                time += (ss * 32)
+        output = [l[0] for l in y]
+        result = statistics.mode(output)
     
-            if 128 < ss < 256:
+        results.append(result)
     
-                dur = (ss-128) * 32
+    all_results_labels = [classifier_labels[0][r-384] for r in results]
+    final_result = statistics.mode(results)
     
-            if 256 < ss < 512:
+    print('Done!')
+    print('=' * 70)
     
-                pitch = (ss-256) % 128
+    print('Most common classification label:', classifier_labels[0][final_result-384])
+    print('Most common classification label ratio:' , results.count(final_result) / len(results))
+    print('Most common classification label index', final_result)
+    print('=' * 70)
     
-                channel = (ss-256) // 128
+    print('All classification labels summary:')
+    print('=' * 70)
     
-                if channel == 1:
-                    channel = 3
-                    vel = 110 + (pitch % 12)
-                    song_f.append(['note', time, dur, channel, pitch, vel, 40])
-                    
-                else:
-                    vel = 80 + (pitch % 12)
-                    channel = 0
-                    song_f.append(['note', time, dur, channel, pitch, vel, 0])
+    for i, a in enumerate(all_results_labels):
+      print('Notes', i*170, '-', (i*170)+340, '===', a)
+    
+    print('=' * 70)
+        
+    print('=' * 70)
+    print('Done!')
+    print('=' * 70)
+    
+    #===============================================================================
+    print('Rendering results...')
+    
+    print('=' * 70)
+    print('Sample INTs', output[:15])
+    print('=' * 70)
+    
+    out1 = output
 
     fn1 = "Melody2Song-Seq2Seq-Music-Transformer-Composition"
     
     detailed_stats = TMIDIX.Tegridy_ms_SONG_to_MIDI_Converter(song_f,
-                                                              output_signature = 'Melody2Song Seq2Seq Music Transformer',
+                                                              output_signature = 'Advanced MIDI Classifier',
                                                               output_file_name = fn1,
                                                               track_name='Project Los Angeles',
                                                               list_of_MIDI_patches=patches