X-Git-Url: https://fleuret.org/cgi-bin/gitweb/gitweb.cgi?a=blobdiff_plain;f=world.py;h=12c65535ff493700429167132494c5befc05be7e;hb=5366dfd7bd57ec3298d1030f7d5327ff26bc5aad;hp=3d6abbe07d7ac0eab712709e0b1bf834e61cde67;hpb=3dea181a5903a0e577e4830c66405b40f2a2df1d;p=picoclvr.git

diff --git a/world.py b/world.py
index 3d6abbe..12c6553 100755
--- a/world.py
+++ b/world.py
@@ -85,9 +85,9 @@ def loss_H(binary_logits, h_threshold=1):
 def train_encoder(
     train_input,
     test_input,
-    depth=2,
+    depth,
+    nb_bits_per_token,
     dim_hidden=48,
-    nb_bits_per_token=8,
     lambda_entropy=0.0,
     lr_start=1e-3,
     lr_end=1e-4,
@@ -96,8 +96,6 @@ def train_encoder(
     logger=None,
     device=torch.device("cpu"),
 ):
-    if logger is None:
-        logger = lambda s: print(s)
 
     mu, std = train_input.float().mean(), train_input.float().std()
 
@@ -157,7 +155,7 @@ def train_encoder(
 
     nb_parameters = sum(p.numel() for p in model.parameters())
 
-    logger(f"nb_parameters {nb_parameters}")
+    logger(f"vqae nb_parameters {nb_parameters}")
 
     model.to(device)
 
@@ -209,7 +207,7 @@ def train_encoder(
         train_loss = acc_train_loss / train_input.size(0)
         test_loss = acc_test_loss / test_input.size(0)
 
-        logger(f"train_ae {k} lr {lr} train_loss {train_loss} test_loss {test_loss}")
+        logger(f"vqae train {k} lr {lr} train_loss {train_loss} test_loss {test_loss}")
         sys.stdout.flush()
 
     return encoder, quantizer, decoder
@@ -366,6 +364,8 @@ def create_data_and_processors(
     nb_test_samples,
     mode,
     nb_steps,
+    depth=3,
+    nb_bits_per_token=8,
     nb_epochs=10,
     device=torch.device("cpu"),
     device_storage=torch.device("cpu"),
@@ -376,6 +376,9 @@ def create_data_and_processors(
     if mode == "first_last":
         steps = [True] + [False] * (nb_steps + 1) + [True]
 
+    if logger is None:
+        logger = lambda s: print(s)
+
     train_input, train_actions = generate_episodes(nb_train_samples, steps)
     train_input, train_actions = train_input.to(device_storage), train_actions.to(
         device_storage
@@ -388,6 +391,8 @@ def create_data_and_processors(
     encoder, quantizer, decoder = train_encoder(
         train_input,
         test_input,
+        depth=depth,
+        nb_bits_per_token=nb_bits_per_token,
         lambda_entropy=1.0,
         nb_epochs=nb_epochs,
         logger=logger,
@@ -401,6 +406,8 @@ def create_data_and_processors(
     pow2 = (2 ** torch.arange(z.size(1), device=device))[None, None, :]
     z_h, z_w = z.size(2), z.size(3)
 
+    logger(f"vqae input {train_input[0].size()} output {z[0].size()}")
+
     def frame2seq(input, batch_size=25):
         seq = []
         p = pow2.to(device)