update tracking

first results

GNN_particles_Ntype.py

@@ -45,7 +45,8 @@
     # config_list = ['boids_64_256']
     # config_list = ['boids_16_256_test_cell_division_c']
     # config_list = ['boids_16_256_division_death_model_2']
-    config_list = ['arbitrary_3_tracking']
+    # config_list = ['arbitrary_3_tracking']
+    config_list = ['boids_16_256_tracking']
     # config_list = ['arbitrary_64']
 
     seed_list = np.arange(10)

config/arbitrary_3_tracking.yaml

@@ -40,7 +40,7 @@ training:
   sparsity: 'replace_embedding_function'
   cluster_method: 'distance_plot'
   data_augmentation: True
-  data_augmentation_loop: 1
+  data_augmentation_loop: 200
   fix_cluster_embedding: True
   device: 'auto'
   coeff_loss1: 1

config/boids_16_256_tracking.yaml

@@ -0,0 +1,46 @@
+description: 'Boids 16 different types'
+dataset: 'boids_16_256'
+
+simulation:
+  params: [[27.6, 92.5, 48.2], [32.0, 51.8, 29.8], [23.6, 35.0, 13.5], [3.3, 76.4, 13.0], [94.0, 78.7, 30.8,], [81.4, 34.6, 2.2], [3.1, 40.5, 17.0], [88.8, 7.2, 29.8], [32.7, 76.8, 26.1], [14.5, 56.8, 27.6], [63.3, 99.9, 13.9], [32.5, 25.2, 24.4], [97.6, 56.5, 12.1], [62.2, 9.1, 28.3], [76.4, 52.9, 32.0], [48.1, 54.4, 30.7]]
+  min_radius: 0.001
+  max_radius: 0.04
+  n_particles: 1792
+  n_particle_types: 16
+  n_interactions: 16
+  has_cell_division: False
+  n_frames: 8000
+  sigma: 0.005
+  delta_t: 0.5
+  dpos_init: 5.0E-4
+  boundary: 'periodic'
+  start_frame: 0
+
+graph_model:
+  particle_model_name: 'PDE_B'
+  mesh_model_name: ''
+  prediction: '2nd_derivative'
+  input_size: 9
+  output_size: 2
+  hidden_dim: 256
+  n_mp_layers: 5
+  aggr_type: 'mean'
+  embedding_dim: 2
+  update_type: 'none'
+
+plotting:
+  colormap: 'tab20'
+  arrow_length: 5
+
+training:
+  n_epochs: 20
+  has_no_tracking: True
+  batch_size: 1
+  small_init_batch_size: True
+  n_runs: 2
+  noise_level: 0
+  data_augmentation: True
+  data_augmentation_loop: 100
+  fix_cluster_embedding: True
+  device: 'auto'
+

src/ParticleGraph/models/graph_trainer.py

@@ -226,7 +226,7 @@ def data_train_particles(config, config_file, device):
                     loss = ((pred - y_batch) / (y_batch)).norm(2) / 1E9
 
             visualize_embedding = True
-            if visualize_embedding & (((epoch < 3 ) & (N % 500 == 0)) | (N==0)):
+            if visualize_embedding & (((epoch < 3 ) & (N%(Niter//100) == 0)) | (N==0)):
                 plot_training(config=config, dataset_name=dataset_name, log_dir=log_dir,
                               epoch=epoch, N=N, x=x, model=model, n_nodes=0, n_node_types=0, index_nodes=0, dataset_num=1,
                               index_particles=index_particles, n_particles=n_particles,
@@ -609,7 +609,7 @@ def data_train_tracking(config, config_file, device):
             # loss2 = 0 * pred.norm(2) / (vnorm**2) * config.training.coeff_loss2
 
             visualize_embedding = True
-            if visualize_embedding & (((epoch < 3 ) & (N % 500 == 0)) | (N==0)):
+            if visualize_embedding & (((epoch < 3 ) & (N % (Niter//100) == 0)) | (N==0)):
                 print(N)
                 fig = plt.figure(figsize=(8, 8))
                 plt.scatter(to_numpy(x[:, 1]), to_numpy(x[:, 2]), s=10, c='k', alpha=0.05)
@@ -666,7 +666,7 @@ def data_train_tracking(config, config_file, device):
         plt.xticks([])
         plt.yticks([])
         plt.tight_layout()
-        plt.savefig(f"./{log_dir}/tmp_training/embedding/before_particle_{dataset_name}_{epoch}_{N}.tif",dpi=87)
+        plt.savefig(f"./{log_dir}/tmp_training/before_particle_{dataset_name}_{epoch}_{N}.tif",dpi=87)
         plt.close()
 
         if epoch%2 == 0:
@@ -739,7 +739,7 @@ def data_train_tracking(config, config_file, device):
         plt.xticks([])
         plt.yticks([])
         plt.tight_layout()
-        plt.savefig(f"./{log_dir}/tmp_training/embedding/after_particle_{dataset_name}_{epoch}_{N}.tif", dpi=87)
+        plt.savefig(f"./{log_dir}/tmp_training/after_particle_{dataset_name}_{epoch}_{N}.tif", dpi=87)
         plt.close()