X-Git-Url: https://fleuret.org/cgi-bin/gitweb/gitweb.cgi?a=blobdiff_plain;f=tasks.py;h=0143ab20058de577a590f16ccfda2093adf174e4;hb=687d5b2d9f465577665991b84faec7c789685271;hp=b2f7d7dc5f750610333d03b7da6c183d83ff7a7a;hpb=16cb07f99cf770fb4e97824f874a68cbddd4c1cf;p=picoclvr.git

diff --git a/tasks.py b/tasks.py
index b2f7d7d..0143ab2 100755
--- a/tasks.py
+++ b/tasks.py
@@ -181,6 +181,39 @@ class SandBox(Task):
             f"accuracy_test {n_epoch} nb_total {test_nb_total} nb_correct {test_nb_correct} accuracy {(100.0*test_nb_correct)/test_nb_total:.02f}%"
         )
 
+        if save_attention_image is not None:
+            for k in range(10):
+                ns = torch.randint(self.test_input.size(0), (1,)).item()
+                input = self.test_input[ns : ns + 1].clone()
+
+                with torch.autograd.no_grad():
+                    t = model.training
+                    model.eval()
+                    model.record_attention(True)
+                    model(BracketedSequence(input))
+                    model.train(t)
+                    ram = model.retrieve_attention()
+                    model.record_attention(False)
+
+                tokens_output = [c for c in self.problem.seq2str(input[0])]
+                tokens_input = ["n/a"] + tokens_output[:-1]
+                for n_head in range(ram[0].size(1)):
+                    filename = os.path.join(
+                        result_dir, f"sandbox_attention_{k}_h{n_head}.pdf"
+                    )
+                    attention_matrices = [m[0, n_head] for m in ram]
+                    save_attention_image(
+                        filename,
+                        tokens_input,
+                        tokens_output,
+                        attention_matrices,
+                        k_top=10,
+                        # min_total_attention=0.9,
+                        token_gap=12,
+                        layer_gap=50,
+                    )
+                    logger(f"wrote {filename}")
+
 
 ######################################################################