Training using same example for meta

geoelements · Jul 15, 2024 · 48f6e32 · 48f6e32
1 parent b7fc8c5
commit 48f6e32
Showing 1 changed file with 27 additions and 47 deletions.
diff --git a/gns/train.py b/gns/train.py
@@ -709,6 +709,12 @@ def train(rank, cfg, world_size, device, verbose, use_dist):
     if use_dist:
         distribute.cleanup()
 
+import torch
+import torch.nn as nn
+import torch.optim as optim
+from tqdm import tqdm
+import copy
+from collections import defaultdict
 
 def train_maml(rank, cfg, world_size, device, verbose, use_dist):
     """Train the model using MAML.
@@ -817,7 +823,7 @@ def train_maml(rank, cfg, world_size, device, verbose, use_dist):
                     # Identify the material property and use the corresponding encoder
                     material_id = material_property[0].item()  # Assuming all materials in the batch are the same
                     task_encoder = task_encoders[material_id]
-                    task_optimizer = torch.optim.Adam(task_encoder.parameters(), lr=inner_lr)
+                    task_optimizer = optim.Adam(task_encoder.parameters(), lr=inner_lr)
 
                     # Perform a few steps of gradient descent in the inner loop
                     for _ in range(inner_steps):
@@ -851,56 +857,30 @@ def train_maml(rank, cfg, world_size, device, verbose, use_dist):
 
                     # Compute the meta-loss using the task-specific encoder parameters
                     meta_loss = 0.0
-                    for example in valid_dl:
-                        (
-                            position,
-                            particle_type,
-                            material_property,
-                            n_particles_per_example,
-                            labels,
-                        ) = prepare_data(example, device_id)
-
-                        n_particles_per_example = n_particles_per_example.to(device_id)
-                        labels = labels.to(device_id)
-
-                        sampled_noise = (
-                            noise_utils.get_random_walk_noise_for_position_sequence(
-                                position, noise_std_last_step=cfg.data.noise_std
-                            ).to(device_id)
-                        )
-                        non_kinematic_mask = (
-                            (particle_type != cfg.data.kinematic_particle_id)
-                            .clone()
-                            .detach()
-                            .to(device_id)
-                        )
-                        sampled_noise *= non_kinematic_mask.view(-1, 1, 1)
 
-                        # Replace the simulator's encoder with the task-specific encoder
-                        simulator._encode_process_decode._encoder = task_encoder
+                    # Replace the simulator's encoder with the task-specific encoder
+                    simulator._encode_process_decode._encoder = task_encoder
 
-                        pred_acc, target_acc = predict_fn(
-                            next_positions=labels.to(device_or_rank),
-                            position_sequence_noise=sampled_noise.to(device_or_rank),
-                            position_sequence=position.to(device_or_rank),
-                            nparticles_per_example=n_particles_per_example.to(
-                                device_or_rank
-                            ),
-                            particle_types=particle_type.to(device_or_rank),
-                            material_property=(
-                                material_property.to(device_or_rank)
-                                if n_features == 3
-                                else None
-                            ),
-                        )
-
-                        valid_loss = acceleration_loss(pred_acc, target_acc, non_kinematic_mask)
-                        meta_loss += valid_loss
+                    pred_acc, target_acc = predict_fn(
+                        next_positions=labels.to(device_or_rank),
+                        position_sequence_noise=sampled_noise.to(device_or_rank),
+                        position_sequence=position.to(device_or_rank),
+                        nparticles_per_example=n_particles_per_example.to(
+                            device_or_rank
+                        ),
+                        particle_types=particle_type.to(device_or_rank),
+                        material_property=(
+                            material_property.to(device_or_rank)
+                            if n_features == 3
+                            else None
+                        ),
+                    )
 
-                        # Restore the original encoder
-                        simulator._encode_process_decode._encoder = original_encoder
+                    valid_loss = acceleration_loss(pred_acc, target_acc, non_kinematic_mask)
+                    meta_loss += valid_loss
 
-                    meta_loss /= len(valid_dl)
+                    # Restore the original encoder
+                    simulator._encode_process_decode._encoder = original_encoder
 
                     # Outer loop: Update the main encoder using the meta-loss
                     optimizer.zero_grad()