LCNN all data aug

.gitignore

@@ -4,7 +4,7 @@ checkpoints/
 debug.sh
 data_test/
 
-.idea
+.idea/
 
 # Byte-compiled / optimized / DLL files
 __pycache__/
@@ -180,3 +180,4 @@ cython_debug/
 
 # PyPI configuration file
 .pypirc
+/checkpoints/

script.py

@@ -2,10 +2,10 @@ import io
 import time
 import os
 
-# os.environ['OMP_NUM_THREADS'] = '1'
-# os.environ['OPENBLAS_NUM_THREADS'] = '1'
-# os.environ['MKL_NUM_THREADS'] = '1'
-# os.environ['NUMEXPR_NUM_THREADS'] = '1'
+os.environ['OMP_NUM_THREADS'] = '1'
+os.environ['OPENBLAS_NUM_THREADS'] = '1'
+os.environ['MKL_NUM_THREADS'] = '1'
+os.environ['NUMEXPR_NUM_THREADS'] = '1'
 
 import torch
 import tqdm.auto as tqdm
@@ -51,30 +51,31 @@ print('Define Model')
 # model_path = './checkpoints/RAWNET_ASVSPOOF_FOR_INTHEWILD_PURDUE.pth'
 # model.load_state_dict(torch.load(model_path, map_location=device))
 
-# # LCNN MODEL
-# model = LCNN(return_emb=False).to(device)
-# # model_path = './checkpoints/LCNN_ASVSPOOF_FOR_INTHEWILD_PURDUE.pth'
+# LCNN MODEL
+model = LCNN(return_emb=False).to(device)
+# model_path = './checkpoints/LCNN_ASVSPOOF_FOR_INTHEWILD_PURDUE.pth'
 # model_path = './checkpoints/LCNN_ALL_DATA.pth'
-# model.load_state_dict(torch.load(model_path, map_location=device))
-
-# MOE MODEL
-expert_1 = LCNN(return_emb=True).to(device)
-expert_2 = LCNN(return_emb=True).to(device)
-expert_3 = LCNN(return_emb=True).to(device)
-expert_4 = LCNN(return_emb=True).to(device)
-expert_5 = LCNN(return_emb=True).to(device)
-expert_6 = LCNN(return_emb=True).to(device)
+model_path = './checkpoints/LCNN_ALL_DATA_AUG.pth'
+model.load_state_dict(torch.load(model_path, map_location=device))
 
-# # model = UltimateMOE(experts=[expert_1, expert_2, expert_3, expert_4])
-# # model_path = './checkpoints/MOE_ULTIMATE.pth'
+# # MOE MODEL
+# expert_1 = LCNN(return_emb=True).to(device)
+# expert_2 = LCNN(return_emb=True).to(device)
+# expert_3 = LCNN(return_emb=True).to(device)
+# expert_4 = LCNN(return_emb=True).to(device)
+# expert_5 = LCNN(return_emb=True).to(device)
+# expert_6 = LCNN(return_emb=True).to(device)
 #
-# model = MOE_attention(experts=[expert_1, expert_2, expert_3, expert_4, expert_5, expert_6], device=device)
-# # model_path = './checkpoints/MOE_ATTENTION.pth'
-# model_path = './checkpoints/MOE_TRANSF.pth'
-
-expert_7 = LCNN(return_emb=True).to(device)
-model = MOE_attention(experts=[expert_1, expert_2, expert_3, expert_4, expert_5, expert_6, expert_7], device=device, freezing=True)
-model_path = './checkpoints/MOE_TRANSF_7EXP.pth'
+# # # model = UltimateMOE(experts=[expert_1, expert_2, expert_3, expert_4])
+# # # model_path = './checkpoints/MOE_ULTIMATE.pth'
+#
+# # model = MOE_attention(experts=[expert_1, expert_2, expert_3, expert_4, expert_5, expert_6], device=device)
+# # # model_path = './checkpoints/MOE_ATTENTION.pth'
+# # model_path = './checkpoints/MOE_TRANSF.pth'
+#
+# expert_7 = LCNN(return_emb=True).to(device)
+# model = MOE_attention(experts=[expert_1, expert_2, expert_3, expert_4, expert_5, expert_6, expert_7], device=device, freezing=True)
+# model_path = './checkpoints/MOE_TRANSF_7EXP.pth'
 
 
 model = (model).to(device)
@@ -102,44 +103,44 @@ for el in tqdm.tqdm(dataset_remote):
 
     # if you are using libraries that expect a file. You can use BytesIO object
 
-    try:
-
-        # RUNNING ON HUGGINGFACE
-        file_like = io.BytesIO(el["audio"]["bytes"])
-        tensor = preprocess(file_like)
-        # # RUNNING LOCALLY
-        # tensor = preprocess(el)
-
-        with torch.no_grad():
-            # soft decision (such as log likelihood score)
-            # positive score correspond to synthetic prediction
-            # negative score correspond to pristine prediction
-
-            # # OLD MODEL
-            # score = model(tensor.to(device)).cpu().item()
-            # CUSTOM MODEL
-            score = model(tensor.to(device))[:, 1].cpu()
-            print(f'SCORE OUT: {score}')
-            score = score.mean().item()
-            print(f'SCORE FINAL: {score}')
-
-            # we require a hard decision to be submited. so you need to pick a threshold
-            pred = "generated" if score > model.threshold else "pristine"
-
-        # append your prediction
-        # "id" and "pred" are required. "score" will not be used in scoring but we encourage you to include it. We'll use it for analysis of the results
-
-        # RUNNING ON HUGGINGFACE
-        out.append(dict(id=el["id"], pred=pred, score=score, time=time.time() - start_time))
-        ## RUNNING LOCALLY
-        # out.append(dict(id=el, pred=pred, score=score, time=time.time() - start_time))
-
-    except Exception as e:
-        print(e)
-        print("failed", el["id"])
-        out.append(dict(id=el["id"], pred="none", score=None))
-        # print("failed", el)
-        # out.append(dict(id=el, pred="none", score=None))
+    # try:
+
+    # RUNNING ON HUGGINGFACE
+    file_like = io.BytesIO(el["audio"]["bytes"])
+    tensor = preprocess(file_like)
+    # # RUNNING LOCALLY
+    # tensor = preprocess(el)
+
+    with torch.no_grad():
+        # soft decision (such as log likelihood score)
+        # positive score correspond to synthetic prediction
+        # negative score correspond to pristine prediction
+
+        # # OLD MODEL
+        # score = model(tensor.to(device)).cpu().item()
+        # CUSTOM MODEL
+        score = model(tensor.to(device))[:, 1].cpu()
+        print(f'SCORE OUT: {score}')
+        score = score.mean().item()
+        print(f'SCORE FINAL: {score}')
+
+        # we require a hard decision to be submited. so you need to pick a threshold
+        pred = "generated" if score > model.threshold else "pristine"
+
+    # append your prediction
+    # "id" and "pred" are required. "score" will not be used in scoring but we encourage you to include it. We'll use it for analysis of the results
+
+    # RUNNING ON HUGGINGFACE
+    out.append(dict(id=el["id"], pred=pred, score=score, time=time.time() - start_time))
+    # # RUNNING LOCALLY
+    # out.append(dict(id=el, pred=pred, score=score, time=time.time() - start_time))
+
+    # except Exception as e:
+    #     print(e)
+    #     print("failed", el["id"])
+    #     out.append(dict(id=el["id"], pred="none", score=None))
+    #     # print("failed", el)
+    #     # out.append(dict(id=el, pred="none", score=None))
 
 # save the final result and that's it
 pd.DataFrame(out).to_csv("submission.csv", index=False)