deleted old recordings, updated gpu for training,

.gitignore

@@ -5,4 +5,5 @@ __pycache__/
 *.h5
 *.ckpt
 *.ipynb
-*.onnx
+*.onnx
+*.npy

conf/app.yaml

@@ -20,9 +20,7 @@ training:
   use_gpu: true
 
 inference:
-  model_path: checkpoints/inference_recognition_model.ckpt
   num_classes: 4
-  output_path: onnx_files/inference_recognition_model.onnx
 
 app:
   build_dir: dist

convert_to_onnx.py

@@ -79,5 +79,7 @@ if __name__ == "__main__":
     convert_to_onnx(
         ckpt_path=os.path.join(CHECKPOINTS_DIR, model_checkpoint), fp16=False
     )
-
+    
-    print("Conversion complete stored at: ", os.path.join(ONNX_DIR, model_checkpoint))
+    output_file = "inference_recognition_model.onnx"
+    
+    print("Conversion complete stored at: ", os.path.join(ONNX_DIR, output_file))

data/scripts/data_gen.py

@@ -13,56 +13,56 @@ mods = {
 def generate_modulated_signals():
     for modulation in ["bpsk", "qpsk", "qam16", "qam64"]:
         for snr in np.arange(-6, 13, 3):
+            for i in range(100):
+                recording_length = 1024
+                beta = 0.3  # the rolloff factor, can be changed to add variety
+                sps = 4  # samples per symbol, or the relative bandwidth of the digital signal. Can also be changed.
 
-            recording_length = 1024
+                # blocks don't directly take the string 'qpsk' so we use the dict 'mods' to get parameters
-            beta = 0.3  # the rolloff factor, can be changed to add variety
+                constellation_type = mods[modulation]["constellation_type"]
-            sps = 4  # samples per symbol, or the relative bandwidth of the digital signal. Can also be changed.
+                num_bits_per_symbol = mods[modulation]["num_bits_per_symbol"]
 
-            # blocks don't directly take the string 'qpsk' so we use the dict 'mods' to get parameters
+                # construct the digital modulation blocks with these parameters
-            constellation_type = mods[modulation]["constellation_type"]
+                # we have bit source -> mapper -> upsampling -> pulse shaping
-            num_bits_per_symbol = mods[modulation]["num_bits_per_symbol"]
 
-            # construct the digital modulation blocks with these parameters
+                bit_source = block_generator.RandomBinarySource()
-            # we have bit source -> mapper -> upsampling -> pulse shaping
+                mapper = block_generator.Mapper(
+                    constellation_type=constellation_type,
+                    num_bits_per_symbol=num_bits_per_symbol,
+                )
+                upsampler = block_generator.Upsampling(factor=sps)
+                pulse_shaping_filter = block_generator.RaisedCosineFilter(
+                    upsampling_factor=sps, beta=beta
+                )
 
-            bit_source = block_generator.RandomBinarySource()
+                pulse_shaping_filter.connect_input([upsampler])
-            mapper = block_generator.Mapper(
+                upsampler.connect_input([mapper])
-                constellation_type=constellation_type,
+                mapper.connect_input([bit_source])
-                num_bits_per_symbol=num_bits_per_symbol,
-            )
-            upsampler = block_generator.Upsampling(factor=sps)
-            pulse_shaping_filter = block_generator.RaisedCosineFilter(
-                upsampling_factor=sps, beta=beta
-            )
 
-            pulse_shaping_filter.connect_input([upsampler])
+                modulation_recording = pulse_shaping_filter.record(
-            upsampler.connect_input([mapper])
+                    num_samples=recording_length
-            mapper.connect_input([bit_source])
+                )
 
-            modulation_recording = pulse_shaping_filter.record(
+                # add noise by calculating the power of the modulation recording and generating AWGN from the snr parameter
-                num_samples=recording_length
+                signal_power = np.mean(np.abs(modulation_recording.data[0] ** 2))
-            )
+                awgn_source = block_generator.AWGNSource(
+                    variance=(signal_power / 2) * (10 ** (((-1 * snr) / 20)))
+                )
+                noise = awgn_source.record(num_samples=recording_length)
+                samples_with_noise = modulation_recording.data + noise.data
+                output_recording = Recording(data=samples_with_noise)
 
-            # add noise by calculating the power of the modulation recording and generating AWGN from the snr parameter
+                # add metadata for ML later
-            signal_power = np.mean(np.abs(modulation_recording.data[0] ** 2))
+                output_recording.add_to_metadata(key="modulation", value=modulation)
-            awgn_source = block_generator.AWGNSource(
+                output_recording.add_to_metadata(key="snr", value=int(snr))
-                variance=(signal_power / 2) * (10 ** (((-1 * snr) / 20)))
+                output_recording.add_to_metadata(key="beta", value=beta)
-            )
+                output_recording.add_to_metadata(key="sps", value=sps)
-            noise = awgn_source.record(num_samples=recording_length)
-            samples_with_noise = modulation_recording.data + noise.data
-            output_recording = Recording(data=samples_with_noise)
 
-            # add metadata for ML later
+                # view if you want
-            output_recording.add_to_metadata(key="modulation", value=modulation)
+                # output_recording.view()
-            output_recording.add_to_metadata(key="snr", value=int(snr))
-            output_recording.add_to_metadata(key="beta", value=beta)
-            output_recording.add_to_metadata(key="sps", value=sps)
 
-            # view if you want
+                # save to file
-            # output_recording.view()
+                output_recording.to_npy()  # optionally add path and filename parameters
-
-            # save to file
-            output_recording.to_npy()  # optionally add path and filename parameters
 
 
 if __name__ == "__main__":

data/training/train.py

@@ -31,7 +31,7 @@ def train_model():
 
     train_flag = True
     batch_size = 128
-    epochs = 1
+    epochs = 50
 
     checkpoint_dir = training_cfg.checkpoint_dir
     checkpoint_filename = training_cfg.checkpoint_filename
@@ -76,8 +76,8 @@ def train_model():
     hparams = {
         "drop_path_rate": 0.2,
         "drop_rate": 0.5,
-        "learning_rate": 3e-4,
+        "learning_rate": 1e-4,
-        "wd": 0.2,
+        "wd": 0.01,
     }
 
     class RFClassifier(L.LightningModule):

helpers/app_settings.py

@@ -33,9 +33,7 @@ class TrainingConfig:
 
 @dataclass
 class InferenceConfig:
-    model_path: str
     num_classes: int
-    output_path: str
 
 
 @dataclass