Auto-upload sent_gpt2s_geom_drugs_0501-2151/config.yaml

xl_molecules/sent_gpt2s_geom_drugs_0501-2151/config.yaml

@@ -0,0 +1,78 @@
+tokenizer:
+  type: sent
+  max_length: -1
+  truncation_length: 1024
+  undirected: true
+  labeled_graph: true
+seed: 42
+resume: true
+model:
+  model_name: gpt2-s
+  learning_rate: 0.000849
+  weight_decay: 0.1
+  warmup_steps: 1414
+  max_steps: 200000
+data:
+  batch_size: 32
+  num_workers: 8
+  test_num_workers: 0
+  include_hydrogens: false
+  data_root: data
+  use_cache: true
+  cache_dir: data/cache
+  hierarchy_cache_dir: data/hierarchy_cache
+  use_hierarchy_cache: false
+  hf_repo_id: MOSAIC-UCSD/MOSAIC_data_cache
+  hf_ckpt_repo_id: MOSAIC-UCSD/MOSAIC_model_ckpt
+  dataset_name: geom_drugs
+  num_train: 350000
+  num_val: 1000
+  num_test: 1000
+  data_file: null
+  min_atoms: null
+  max_atoms: null
+  min_rings: null
+trainer:
+  target_samples_seen: null
+  max_epochs: -1
+  max_steps: 200000
+  val_checks_per_epoch: null
+  validate_every_n_epochs: null
+  val_check_interval: 5000
+  check_val_every_n_epoch: null
+  precision: 32
+  gradient_clip_val: 1.0
+  accelerator: auto
+  devices: 1
+  num_nodes: 1
+  strategy: auto
+  accumulate_grad_batches: 1
+sampling:
+  num_samples: 10000
+  batch_size: 32
+  top_k: 10
+  temperature: 1.0
+  max_length: 1024
+metrics:
+  compute_emd: false
+  orca_bin: null
+  motif_subsample_size: 1000
+precompute:
+  enabled: true
+  force: false
+logs:
+  base_dir: outputs/train_eval_nrp
+  run_name: sent_gpt2s_geom_drugs_0501-2151
+  path: /data/home/Andrew.Yang/StreamMUSE/MOSAIC/outputs/sent_gpt2s_sweep_0501-2151/geom_drugs
+wandb:
+  enabled: true
+  project: mosaic-round1
+  entity: null
+  name: null
+  tags: []
+  notes: null
+  log_model: true
+  log_graphs: true
+  max_logged_molecules: 12
+  eval_every_n_val: 250
+  eval_num_samples: 5