pullproper number of spectra

to saturate batch size

casanovo/denovo/dataloaders.py

@@ -158,9 +158,14 @@ def _make_db_loader(
         torch.utils.data.DataLoader
             A PyTorch DataLoader.
         """
+        # Calculate new batch size to saturate previous batch size with PSMs
+        pep_per_spec = []
+        for i in range(min(10, len(dataset))):
+            pep_per_spec.append(len(dataset[i][3].split(",")))
+        new_batch_size = int(self.batch_size // np.mean(pep_per_spec))
         return torch.utils.data.DataLoader(
             dataset,
-            batch_size=self.batch_size,
+            batch_size=new_batch_size,
             collate_fn=prepare_db_batch,
             pin_memory=True,
             num_workers=self.n_workers,

casanovo/denovo/model.py

@@ -962,11 +962,9 @@ def predict_step(self, batch, *args):
             precursors,
             encoded_ms,
         ) in self.smart_batch_gen(batch):
-            with torch.set_grad_enabled(True):  # Fixes NaN!?
-                pred, truth = self.decoder(
-                    peptides, precursors, *encoded_ms
-                )  #! get detailed breakdown of this speed
-                sm = torch.nn.Softmax(dim=2)  # dim=2 is very important!
+            with torch.set_grad_enabled(True):
+                pred, truth = self.decoder(peptides, precursors, *encoded_ms)
+                sm = torch.nn.Softmax(dim=2)
                 pred = sm(pred)
                 score_result, per_aa_score = calc_match_score(
                     pred, truth
@@ -1026,8 +1024,9 @@ def on_predict_epoch_end(self, results) -> None:
                     for index, t_or_d, peptide, score, per_aa_scores in list(
                         zip(*batch)
                     ):
+                        # Remove scores of 0 (padding)
                         per_aa_scores = per_aa_scores.numpy()
-                        per_aa_scores = per_aa_scores[per_aa_scores != 0]
+                        per_aa_scores = list(per_aa_scores[per_aa_scores != 0])
                         score = score.numpy()
                         csv_writer.writerow(
                             (
@@ -1061,17 +1060,19 @@ def calc_match_score(
             The score between the input spectra and associated peptide (for an entire batch)
             a list of lists of per amino acid scores (for an entire batch)
     """
-    batch_all_aa_scores = batch_all_aa_scores[
-        :, :-1, :  # -2
-    ]  # Remove trailing tokens from predictions, change to -1 to keep stop token
-    truth_aa_indicies = truth_aa_indicies[
-        :, :  # -1
-    ]  # Remove trailing tokens from label, remove -1 to keep stop token
-    per_aa_scores = batch_all_aa_scores[
-        torch.arange(batch_all_aa_scores.shape[0])[:, None],
-        torch.arange(0, batch_all_aa_scores.shape[1]),
-        truth_aa_indicies,
-    ]
+    # Remove trailing tokens from predictions,
+    batch_all_aa_scores = batch_all_aa_scores[:, :-1]
+
+    # Vectorized scoring using efficient indexing.
+    rows = (
+        torch.arange(batch_all_aa_scores.shape[0])
+        .unsqueeze(-1)
+        .expand(-1, batch_all_aa_scores.shape[1])
+    )
+    cols = torch.arange(0, batch_all_aa_scores.shape[1]).expand_as(rows)
+
+    per_aa_scores = batch_all_aa_scores[rows, cols, truth_aa_indicies]
+
     score_mask = truth_aa_indicies != 0
     masked_per_aa_scores = per_aa_scores * score_mask
     all_scores = masked_per_aa_scores.sum(dim=1) / score_mask.sum(dim=1)

casanovo/denovo/model_runner.py

@@ -394,7 +394,7 @@ def db_search(
     abs_experiment_dirpath = "/net/noble/vol2/home/vananth3/2023_vananth_denovo-dbsearch/results/2023-08-21_speedup"
     profiler = AdvancedProfiler(
         dirpath=abs_experiment_dirpath,
-        filename="test_remove_cms",
+        filename="placeholder",
     )
     trainer = pl.Trainer(
         accelerator="auto",