initial upload

fold0/config.yaml

@@ -0,0 +1,208 @@
+seed_everything: 42
+trainer:
+  accelerator: auto
+  strategy:
+    class_path: lightning.pytorch.strategies.DDPStrategy
+    init_args:
+      accelerator: null
+      parallel_devices: null
+      cluster_environment: null
+      checkpoint_io: null
+      precision_plugin: null
+      ddp_comm_state: null
+      ddp_comm_hook: null
+      ddp_comm_wrapper: null
+      model_averaging_period: null
+      process_group_backend: null
+      timeout: 0:30:00
+      start_method: popen
+      output_device: null
+      dim: 0
+      broadcast_buffers: true
+      process_group: null
+      bucket_cap_mb: 25
+      find_unused_parameters: false
+      check_reduction: false
+      gradient_as_bucket_view: false
+      static_graph: false
+      delay_all_reduce_named_params: null
+      param_to_hook_all_reduce: null
+      mixed_precision: null
+      device_mesh: null
+  devices: auto
+  num_nodes: 2
+  precision: 32
+  logger:
+    class_path: lightning.pytorch.loggers.WandbLogger
+    init_args:
+      name: BLAT_ECOLX_Gonzalez_2019_indels
+      save_dir: logs
+      version: null
+      offline: false
+      dir: null
+      id: null
+      anonymous: null
+      project: GBFT_PROTEINFM_DMS
+      log_model: false
+      experiment: null
+      prefix: ''
+      checkpoint_name: null
+      job_type: null
+      config: null
+      entity: null
+      reinit: null
+      tags: null
+      group: null
+      notes: null
+      magic: null
+      config_exclude_keys: null
+      config_include_keys: null
+      mode: null
+      allow_val_change: null
+      resume: null
+      force: null
+      tensorboard: null
+      sync_tensorboard: null
+      monitor_gym: null
+      save_code: true
+      settings: null
+  callbacks:
+  - class_path: lightning.pytorch.callbacks.LearningRateMonitor
+    init_args:
+      logging_interval: step
+      log_momentum: false
+      log_weight_decay: false
+  - class_path: lightning.pytorch.callbacks.ModelCheckpoint # save ckpt at the end of each epoch, and save the best val_mcc ckpt
+    init_args:
+      dirpath: null
+      filename: epoch_{epoch}-val_mcc:{val_spearman:.3f}
+      monitor: val_spearman
+      verbose: false
+      save_last: true
+      save_top_k: 1
+      save_weights_only: false
+      mode: max
+      auto_insert_metric_name: true
+      every_n_train_steps: null
+      train_time_interval: null
+      every_n_epochs: 1
+      save_on_train_epoch_end: null
+      enable_version_counter: true
+  - class_path: lightning.pytorch.callbacks.early_stopping.EarlyStopping
+    dict_kwargs:
+      monitor: val_spearman
+      mode: max
+      patience: 5
+  fast_dev_run: false
+  max_epochs: null
+  min_epochs: null
+  max_steps: 10000
+  min_steps: null
+  max_time: null
+  limit_train_batches: null
+  limit_val_batches: null
+  limit_test_batches: null
+  limit_predict_batches: null
+  overfit_batches: 0.0
+  val_check_interval: null
+  check_val_every_n_epoch: 1
+  num_sanity_val_steps: null
+  log_every_n_steps: 50
+  enable_checkpointing: null
+  enable_progress_bar: null
+  enable_model_summary: null
+  accumulate_grad_batches: 1
+  gradient_clip_val: 0.1
+  gradient_clip_algorithm: null
+  deterministic: null
+  benchmark: null
+  inference_mode: true
+  use_distributed_sampler: true
+  profiler:
+    class_path: lightning.pytorch.profilers.PyTorchProfiler
+    init_args:
+      dirpath: null
+      filename: null
+      group_by_input_shapes: false
+      emit_nvtx: false
+      export_to_chrome: true
+      row_limit: 20
+      sort_by_key: null
+      record_module_names: true
+      table_kwargs: null
+      record_shapes: false
+    dict_kwargs:
+      profile_memory: true
+  detect_anomaly: false
+  barebones: false
+  plugins: null
+  sync_batchnorm: false
+  reload_dataloaders_every_n_epochs: 0
+  default_root_dir: logs
+model:
+  class_path: genbio_finetune.tasks.SequenceRegression
+  init_args:
+    backbone:
+      class_path: genbio_finetune.models.proteinfm
+      init_args:
+        from_scratch: false
+        use_peft: true
+        save_peft_only: true
+        lora_r: 16
+        lora_alpha: 32
+        lora_dropout: 0.05
+        config_overwrites: null
+        model_init_args: null
+        max_length: 2048
+    adapter:
+      class_path: genbio_finetune.models.MLPPoolAdapter
+      init_args:
+        pooling: mean_pooling
+        hidden_sizes:
+        - 128
+        bias: true
+        dropout: 0.1
+        dropout_in_middle: false
+    optimizer:
+      class_path: torch.optim.AdamW
+      init_args:
+        lr: 0.0001
+        betas:
+        - 0.9
+        - 0.95
+        eps: 1.0e-08
+        weight_decay: 0.01
+        amsgrad: false
+        maximize: false
+        foreach: null
+        capturable: false
+        differentiable: false
+        fused: null
+    lr_scheduler:
+      class_path: genbio_finetune.lr_schedulers.CosineWithWarmup
+      init_args:
+        warmup_ratio: 0.05
+    strict_loading: true
+    reset_optimizer_states: false
+data:
+  class_path: genbio_finetune.data.DMSFitnessPrediction
+  init_args:
+    path: genbio-ai/ProteinGYM-DMS
+    mutation_type: indels
+    task: BLAT_ECOLX_Gonzalez_2019_indels
+    normalize: true
+    train_split_name: train
+    test_split_files: null
+    valid_split_files: null
+    random_seed: 42
+    batch_size: 2
+    shuffle: true
+    sampler: null
+    num_workers: 0
+    pin_memory: true
+    persistent_workers: false
+    cv_num_folds: 5
+    cv_test_fold_id: 0
+    cv_enable_val_fold: true
+    cv_fold_id_col: fold_id
+ckpt_path: null

fold0/model.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:444cce4d074864a21f4e5a7389e77cb278911182534318b732b87423ee629d21
+size 147655384

fold1/config.yaml

@@ -0,0 +1,208 @@
+seed_everything: 42
+trainer:
+  accelerator: auto
+  strategy:
+    class_path: lightning.pytorch.strategies.DDPStrategy
+    init_args:
+      accelerator: null
+      parallel_devices: null
+      cluster_environment: null
+      checkpoint_io: null
+      precision_plugin: null
+      ddp_comm_state: null
+      ddp_comm_hook: null
+      ddp_comm_wrapper: null
+      model_averaging_period: null
+      process_group_backend: null
+      timeout: 0:30:00
+      start_method: popen
+      output_device: null
+      dim: 0
+      broadcast_buffers: true
+      process_group: null
+      bucket_cap_mb: 25
+      find_unused_parameters: false
+      check_reduction: false
+      gradient_as_bucket_view: false
+      static_graph: false
+      delay_all_reduce_named_params: null
+      param_to_hook_all_reduce: null
+      mixed_precision: null
+      device_mesh: null
+  devices: auto
+  num_nodes: 2
+  precision: 32
+  logger:
+    class_path: lightning.pytorch.loggers.WandbLogger
+    init_args:
+      name: BLAT_ECOLX_Gonzalez_2019_indels
+      save_dir: logs
+      version: null
+      offline: false
+      dir: null
+      id: null
+      anonymous: null
+      project: GBFT_PROTEINFM_DMS
+      log_model: false
+      experiment: null
+      prefix: ''
+      checkpoint_name: null
+      job_type: null
+      config: null
+      entity: null
+      reinit: null
+      tags: null
+      group: null
+      notes: null
+      magic: null
+      config_exclude_keys: null
+      config_include_keys: null
+      mode: null
+      allow_val_change: null
+      resume: null
+      force: null
+      tensorboard: null
+      sync_tensorboard: null
+      monitor_gym: null
+      save_code: true
+      settings: null
+  callbacks:
+  - class_path: lightning.pytorch.callbacks.LearningRateMonitor
+    init_args:
+      logging_interval: step
+      log_momentum: false
+      log_weight_decay: false
+  - class_path: lightning.pytorch.callbacks.ModelCheckpoint # save ckpt at the end of each epoch, and save the best val_mcc ckpt
+    init_args:
+      dirpath: null
+      filename: epoch_{epoch}-val_mcc:{val_spearman:.3f}
+      monitor: val_spearman
+      verbose: false
+      save_last: true
+      save_top_k: 1
+      save_weights_only: false
+      mode: max
+      auto_insert_metric_name: true
+      every_n_train_steps: null
+      train_time_interval: null
+      every_n_epochs: 1
+      save_on_train_epoch_end: null
+      enable_version_counter: true
+  - class_path: lightning.pytorch.callbacks.early_stopping.EarlyStopping
+    dict_kwargs:
+      monitor: val_spearman
+      mode: max
+      patience: 5
+  fast_dev_run: false
+  max_epochs: null
+  min_epochs: null
+  max_steps: 10000
+  min_steps: null
+  max_time: null
+  limit_train_batches: null
+  limit_val_batches: null
+  limit_test_batches: null
+  limit_predict_batches: null
+  overfit_batches: 0.0
+  val_check_interval: null
+  check_val_every_n_epoch: 1
+  num_sanity_val_steps: null
+  log_every_n_steps: 50
+  enable_checkpointing: null
+  enable_progress_bar: null
+  enable_model_summary: null
+  accumulate_grad_batches: 1
+  gradient_clip_val: 0.1
+  gradient_clip_algorithm: null
+  deterministic: null
+  benchmark: null
+  inference_mode: true
+  use_distributed_sampler: true
+  profiler:
+    class_path: lightning.pytorch.profilers.PyTorchProfiler
+    init_args:
+      dirpath: null
+      filename: null
+      group_by_input_shapes: false
+      emit_nvtx: false
+      export_to_chrome: true
+      row_limit: 20
+      sort_by_key: null
+      record_module_names: true
+      table_kwargs: null
+      record_shapes: false
+    dict_kwargs:
+      profile_memory: true
+  detect_anomaly: false
+  barebones: false
+  plugins: null
+  sync_batchnorm: false
+  reload_dataloaders_every_n_epochs: 0
+  default_root_dir: logs
+model:
+  class_path: genbio_finetune.tasks.SequenceRegression
+  init_args:
+    backbone:
+      class_path: genbio_finetune.models.proteinfm
+      init_args:
+        from_scratch: false
+        use_peft: true
+        save_peft_only: true
+        lora_r: 16
+        lora_alpha: 32
+        lora_dropout: 0.05
+        config_overwrites: null
+        model_init_args: null
+        max_length: 2048
+    adapter:
+      class_path: genbio_finetune.models.MLPPoolAdapter
+      init_args:
+        pooling: mean_pooling
+        hidden_sizes:
+        - 128
+        bias: true
+        dropout: 0.1
+        dropout_in_middle: false
+    optimizer:
+      class_path: torch.optim.AdamW
+      init_args:
+        lr: 0.0001
+        betas:
+        - 0.9
+        - 0.95
+        eps: 1.0e-08
+        weight_decay: 0.01
+        amsgrad: false
+        maximize: false
+        foreach: null
+        capturable: false
+        differentiable: false
+        fused: null
+    lr_scheduler:
+      class_path: genbio_finetune.lr_schedulers.CosineWithWarmup
+      init_args:
+        warmup_ratio: 0.05
+    strict_loading: true
+    reset_optimizer_states: false
+data:
+  class_path: genbio_finetune.data.DMSFitnessPrediction
+  init_args:
+    path: genbio-ai/ProteinGYM-DMS
+    mutation_type: indels
+    task: BLAT_ECOLX_Gonzalez_2019_indels
+    normalize: true
+    train_split_name: train
+    test_split_files: null
+    valid_split_files: null
+    random_seed: 42
+    batch_size: 2
+    shuffle: true
+    sampler: null
+    num_workers: 0
+    pin_memory: true
+    persistent_workers: false
+    cv_num_folds: 5
+    cv_test_fold_id: 1
+    cv_enable_val_fold: true
+    cv_fold_id_col: fold_id
+ckpt_path: null

fold1/model.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09a7f6e5597b190f26e447ad584ee7b85a40bb21b278f85de1dac8d7e8bf2caf
+size 147655192

fold2/config.yaml

@@ -0,0 +1,208 @@
+seed_everything: 42
+trainer:
+  accelerator: auto
+  strategy:
+    class_path: lightning.pytorch.strategies.DDPStrategy
+    init_args:
+      accelerator: null
+      parallel_devices: null
+      cluster_environment: null
+      checkpoint_io: null
+      precision_plugin: null
+      ddp_comm_state: null
+      ddp_comm_hook: null
+      ddp_comm_wrapper: null
+      model_averaging_period: null
+      process_group_backend: null
+      timeout: 0:30:00
+      start_method: popen
+      output_device: null
+      dim: 0
+      broadcast_buffers: true
+      process_group: null
+      bucket_cap_mb: 25
+      find_unused_parameters: false
+      check_reduction: false
+      gradient_as_bucket_view: false
+      static_graph: false
+      delay_all_reduce_named_params: null
+      param_to_hook_all_reduce: null
+      mixed_precision: null
+      device_mesh: null
+  devices: auto
+  num_nodes: 2
+  precision: 32
+  logger:
+    class_path: lightning.pytorch.loggers.WandbLogger
+    init_args:
+      name: BLAT_ECOLX_Gonzalez_2019_indels
+      save_dir: logs
+      version: null
+      offline: false
+      dir: null
+      id: null
+      anonymous: null
+      project: GBFT_PROTEINFM_DMS
+      log_model: false
+      experiment: null
+      prefix: ''
+      checkpoint_name: null
+      job_type: null
+      config: null
+      entity: null
+      reinit: null
+      tags: null
+      group: null
+      notes: null
+      magic: null
+      config_exclude_keys: null
+      config_include_keys: null
+      mode: null
+      allow_val_change: null
+      resume: null
+      force: null
+      tensorboard: null
+      sync_tensorboard: null
+      monitor_gym: null
+      save_code: true
+      settings: null
+  callbacks:
+  - class_path: lightning.pytorch.callbacks.LearningRateMonitor
+    init_args:
+      logging_interval: step
+      log_momentum: false
+      log_weight_decay: false
+  - class_path: lightning.pytorch.callbacks.ModelCheckpoint # save ckpt at the end of each epoch, and save the best val_mcc ckpt
+    init_args:
+      dirpath: null
+      filename: epoch_{epoch}-val_mcc:{val_spearman:.3f}
+      monitor: val_spearman
+      verbose: false
+      save_last: true
+      save_top_k: 1
+      save_weights_only: false
+      mode: max
+      auto_insert_metric_name: true
+      every_n_train_steps: null
+      train_time_interval: null
+      every_n_epochs: 1
+      save_on_train_epoch_end: null
+      enable_version_counter: true
+  - class_path: lightning.pytorch.callbacks.early_stopping.EarlyStopping
+    dict_kwargs:
+      monitor: val_spearman
+      mode: max
+      patience: 5
+  fast_dev_run: false
+  max_epochs: null
+  min_epochs: null
+  max_steps: 10000
+  min_steps: null
+  max_time: null
+  limit_train_batches: null
+  limit_val_batches: null
+  limit_test_batches: null
+  limit_predict_batches: null
+  overfit_batches: 0.0
+  val_check_interval: null
+  check_val_every_n_epoch: 1
+  num_sanity_val_steps: null
+  log_every_n_steps: 50
+  enable_checkpointing: null
+  enable_progress_bar: null
+  enable_model_summary: null
+  accumulate_grad_batches: 1
+  gradient_clip_val: 0.1
+  gradient_clip_algorithm: null
+  deterministic: null
+  benchmark: null
+  inference_mode: true
+  use_distributed_sampler: true
+  profiler:
+    class_path: lightning.pytorch.profilers.PyTorchProfiler
+    init_args:
+      dirpath: null
+      filename: null
+      group_by_input_shapes: false
+      emit_nvtx: false
+      export_to_chrome: true
+      row_limit: 20
+      sort_by_key: null
+      record_module_names: true
+      table_kwargs: null
+      record_shapes: false
+    dict_kwargs:
+      profile_memory: true
+  detect_anomaly: false
+  barebones: false
+  plugins: null
+  sync_batchnorm: false
+  reload_dataloaders_every_n_epochs: 0
+  default_root_dir: logs
+model:
+  class_path: genbio_finetune.tasks.SequenceRegression
+  init_args:
+    backbone:
+      class_path: genbio_finetune.models.proteinfm
+      init_args:
+        from_scratch: false
+        use_peft: true
+        save_peft_only: true
+        lora_r: 16
+        lora_alpha: 32
+        lora_dropout: 0.05
+        config_overwrites: null
+        model_init_args: null
+        max_length: 2048
+    adapter:
+      class_path: genbio_finetune.models.MLPPoolAdapter
+      init_args:
+        pooling: mean_pooling
+        hidden_sizes:
+        - 128
+        bias: true
+        dropout: 0.1
+        dropout_in_middle: false
+    optimizer:
+      class_path: torch.optim.AdamW
+      init_args:
+        lr: 0.0001
+        betas:
+        - 0.9
+        - 0.95
+        eps: 1.0e-08
+        weight_decay: 0.01
+        amsgrad: false
+        maximize: false
+        foreach: null
+        capturable: false
+        differentiable: false
+        fused: null
+    lr_scheduler:
+      class_path: genbio_finetune.lr_schedulers.CosineWithWarmup
+      init_args:
+        warmup_ratio: 0.05
+    strict_loading: true
+    reset_optimizer_states: false
+data:
+  class_path: genbio_finetune.data.DMSFitnessPrediction
+  init_args:
+    path: genbio-ai/ProteinGYM-DMS
+    mutation_type: indels
+    task: BLAT_ECOLX_Gonzalez_2019_indels
+    normalize: true
+    train_split_name: train
+    test_split_files: null
+    valid_split_files: null
+    random_seed: 42
+    batch_size: 2
+    shuffle: true
+    sampler: null
+    num_workers: 0
+    pin_memory: true
+    persistent_workers: false
+    cv_num_folds: 5
+    cv_test_fold_id: 2
+    cv_enable_val_fold: true
+    cv_fold_id_col: fold_id
+ckpt_path: null

fold2/model.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f39d7d180a9996038632c1aac19de2f20e84de2377df4c8b9f0b1be8a5c57f0
+size 147655192

fold3/config.yaml

@@ -0,0 +1,208 @@
+seed_everything: 42
+trainer:
+  accelerator: auto
+  strategy:
+    class_path: lightning.pytorch.strategies.DDPStrategy
+    init_args:
+      accelerator: null
+      parallel_devices: null
+      cluster_environment: null
+      checkpoint_io: null
+      precision_plugin: null
+      ddp_comm_state: null
+      ddp_comm_hook: null
+      ddp_comm_wrapper: null
+      model_averaging_period: null
+      process_group_backend: null
+      timeout: 0:30:00
+      start_method: popen
+      output_device: null
+      dim: 0
+      broadcast_buffers: true
+      process_group: null
+      bucket_cap_mb: 25
+      find_unused_parameters: false
+      check_reduction: false
+      gradient_as_bucket_view: false
+      static_graph: false
+      delay_all_reduce_named_params: null
+      param_to_hook_all_reduce: null
+      mixed_precision: null
+      device_mesh: null
+  devices: auto
+  num_nodes: 2
+  precision: 32
+  logger:
+    class_path: lightning.pytorch.loggers.WandbLogger
+    init_args:
+      name: BLAT_ECOLX_Gonzalez_2019_indels
+      save_dir: logs
+      version: null
+      offline: false
+      dir: null
+      id: null
+      anonymous: null
+      project: GBFT_PROTEINFM_DMS
+      log_model: false
+      experiment: null
+      prefix: ''
+      checkpoint_name: null
+      job_type: null
+      config: null
+      entity: null
+      reinit: null
+      tags: null
+      group: null
+      notes: null
+      magic: null
+      config_exclude_keys: null
+      config_include_keys: null
+      mode: null
+      allow_val_change: null
+      resume: null
+      force: null
+      tensorboard: null
+      sync_tensorboard: null
+      monitor_gym: null
+      save_code: true
+      settings: null
+  callbacks:
+  - class_path: lightning.pytorch.callbacks.LearningRateMonitor
+    init_args:
+      logging_interval: step
+      log_momentum: false
+      log_weight_decay: false
+  - class_path: lightning.pytorch.callbacks.ModelCheckpoint # save ckpt at the end of each epoch, and save the best val_mcc ckpt
+    init_args:
+      dirpath: null
+      filename: epoch_{epoch}-val_mcc:{val_spearman:.3f}
+      monitor: val_spearman
+      verbose: false
+      save_last: true
+      save_top_k: 1
+      save_weights_only: false
+      mode: max
+      auto_insert_metric_name: true
+      every_n_train_steps: null
+      train_time_interval: null
+      every_n_epochs: 1
+      save_on_train_epoch_end: null
+      enable_version_counter: true
+  - class_path: lightning.pytorch.callbacks.early_stopping.EarlyStopping
+    dict_kwargs:
+      monitor: val_spearman
+      mode: max
+      patience: 5
+  fast_dev_run: false
+  max_epochs: null
+  min_epochs: null
+  max_steps: 10000
+  min_steps: null
+  max_time: null
+  limit_train_batches: null
+  limit_val_batches: null
+  limit_test_batches: null
+  limit_predict_batches: null
+  overfit_batches: 0.0
+  val_check_interval: null
+  check_val_every_n_epoch: 1
+  num_sanity_val_steps: null
+  log_every_n_steps: 50
+  enable_checkpointing: null
+  enable_progress_bar: null
+  enable_model_summary: null
+  accumulate_grad_batches: 1
+  gradient_clip_val: 0.1
+  gradient_clip_algorithm: null
+  deterministic: null
+  benchmark: null
+  inference_mode: true
+  use_distributed_sampler: true
+  profiler:
+    class_path: lightning.pytorch.profilers.PyTorchProfiler
+    init_args:
+      dirpath: null
+      filename: null
+      group_by_input_shapes: false
+      emit_nvtx: false
+      export_to_chrome: true
+      row_limit: 20
+      sort_by_key: null
+      record_module_names: true
+      table_kwargs: null
+      record_shapes: false
+    dict_kwargs:
+      profile_memory: true
+  detect_anomaly: false
+  barebones: false
+  plugins: null
+  sync_batchnorm: false
+  reload_dataloaders_every_n_epochs: 0
+  default_root_dir: logs
+model:
+  class_path: genbio_finetune.tasks.SequenceRegression
+  init_args:
+    backbone:
+      class_path: genbio_finetune.models.proteinfm
+      init_args:
+        from_scratch: false
+        use_peft: true
+        save_peft_only: true
+        lora_r: 16
+        lora_alpha: 32
+        lora_dropout: 0.05
+        config_overwrites: null
+        model_init_args: null
+        max_length: 2048
+    adapter:
+      class_path: genbio_finetune.models.MLPPoolAdapter
+      init_args:
+        pooling: mean_pooling
+        hidden_sizes:
+        - 128
+        bias: true
+        dropout: 0.1
+        dropout_in_middle: false
+    optimizer:
+      class_path: torch.optim.AdamW
+      init_args:
+        lr: 0.0001
+        betas:
+        - 0.9
+        - 0.95
+        eps: 1.0e-08
+        weight_decay: 0.01
+        amsgrad: false
+        maximize: false
+        foreach: null
+        capturable: false
+        differentiable: false
+        fused: null
+    lr_scheduler:
+      class_path: genbio_finetune.lr_schedulers.CosineWithWarmup
+      init_args:
+        warmup_ratio: 0.05
+    strict_loading: true
+    reset_optimizer_states: false
+data:
+  class_path: genbio_finetune.data.DMSFitnessPrediction
+  init_args:
+    path: genbio-ai/ProteinGYM-DMS
+    mutation_type: indels
+    task: BLAT_ECOLX_Gonzalez_2019_indels
+    normalize: true
+    train_split_name: train
+    test_split_files: null
+    valid_split_files: null
+    random_seed: 42
+    batch_size: 2
+    shuffle: true
+    sampler: null
+    num_workers: 0
+    pin_memory: true
+    persistent_workers: false
+    cv_num_folds: 5
+    cv_test_fold_id: 3
+    cv_enable_val_fold: true
+    cv_fold_id_col: fold_id
+ckpt_path: null

fold3/model.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b607eaff29156c240801bc9ed5840c8947f986c96cb2a71031bd2ebf87b5917
+size 147655192

fold4/config.yaml

@@ -0,0 +1,208 @@
+seed_everything: 42
+trainer:
+  accelerator: auto
+  strategy:
+    class_path: lightning.pytorch.strategies.DDPStrategy
+    init_args:
+      accelerator: null
+      parallel_devices: null
+      cluster_environment: null
+      checkpoint_io: null
+      precision_plugin: null
+      ddp_comm_state: null
+      ddp_comm_hook: null
+      ddp_comm_wrapper: null
+      model_averaging_period: null
+      process_group_backend: null
+      timeout: 0:30:00
+      start_method: popen
+      output_device: null
+      dim: 0
+      broadcast_buffers: true
+      process_group: null
+      bucket_cap_mb: 25
+      find_unused_parameters: false
+      check_reduction: false
+      gradient_as_bucket_view: false
+      static_graph: false
+      delay_all_reduce_named_params: null
+      param_to_hook_all_reduce: null
+      mixed_precision: null
+      device_mesh: null
+  devices: auto
+  num_nodes: 2
+  precision: 32
+  logger:
+    class_path: lightning.pytorch.loggers.WandbLogger
+    init_args:
+      name: BLAT_ECOLX_Gonzalez_2019_indels
+      save_dir: logs
+      version: null
+      offline: false
+      dir: null
+      id: null
+      anonymous: null
+      project: GBFT_PROTEINFM_DMS
+      log_model: false
+      experiment: null
+      prefix: ''
+      checkpoint_name: null
+      job_type: null
+      config: null
+      entity: null
+      reinit: null
+      tags: null
+      group: null
+      notes: null
+      magic: null
+      config_exclude_keys: null
+      config_include_keys: null
+      mode: null
+      allow_val_change: null
+      resume: null
+      force: null
+      tensorboard: null
+      sync_tensorboard: null
+      monitor_gym: null
+      save_code: true
+      settings: null
+  callbacks:
+  - class_path: lightning.pytorch.callbacks.LearningRateMonitor
+    init_args:
+      logging_interval: step
+      log_momentum: false
+      log_weight_decay: false
+  - class_path: lightning.pytorch.callbacks.ModelCheckpoint # save ckpt at the end of each epoch, and save the best val_mcc ckpt
+    init_args:
+      dirpath: null
+      filename: epoch_{epoch}-val_mcc:{val_spearman:.3f}
+      monitor: val_spearman
+      verbose: false
+      save_last: true
+      save_top_k: 1
+      save_weights_only: false
+      mode: max
+      auto_insert_metric_name: true
+      every_n_train_steps: null
+      train_time_interval: null
+      every_n_epochs: 1
+      save_on_train_epoch_end: null
+      enable_version_counter: true
+  - class_path: lightning.pytorch.callbacks.early_stopping.EarlyStopping
+    dict_kwargs:
+      monitor: val_spearman
+      mode: max
+      patience: 5
+  fast_dev_run: false
+  max_epochs: null
+  min_epochs: null
+  max_steps: 10000
+  min_steps: null
+  max_time: null
+  limit_train_batches: null
+  limit_val_batches: null
+  limit_test_batches: null
+  limit_predict_batches: null
+  overfit_batches: 0.0
+  val_check_interval: null
+  check_val_every_n_epoch: 1
+  num_sanity_val_steps: null
+  log_every_n_steps: 50
+  enable_checkpointing: null
+  enable_progress_bar: null
+  enable_model_summary: null
+  accumulate_grad_batches: 1
+  gradient_clip_val: 0.1
+  gradient_clip_algorithm: null
+  deterministic: null
+  benchmark: null
+  inference_mode: true
+  use_distributed_sampler: true
+  profiler:
+    class_path: lightning.pytorch.profilers.PyTorchProfiler
+    init_args:
+      dirpath: null
+      filename: null
+      group_by_input_shapes: false
+      emit_nvtx: false
+      export_to_chrome: true
+      row_limit: 20
+      sort_by_key: null
+      record_module_names: true
+      table_kwargs: null
+      record_shapes: false
+    dict_kwargs:
+      profile_memory: true
+  detect_anomaly: false
+  barebones: false
+  plugins: null
+  sync_batchnorm: false
+  reload_dataloaders_every_n_epochs: 0
+  default_root_dir: logs
+model:
+  class_path: genbio_finetune.tasks.SequenceRegression
+  init_args:
+    backbone:
+      class_path: genbio_finetune.models.proteinfm
+      init_args:
+        from_scratch: false
+        use_peft: true
+        save_peft_only: true
+        lora_r: 16
+        lora_alpha: 32
+        lora_dropout: 0.05
+        config_overwrites: null
+        model_init_args: null
+        max_length: 2048
+    adapter:
+      class_path: genbio_finetune.models.MLPPoolAdapter
+      init_args:
+        pooling: mean_pooling
+        hidden_sizes:
+        - 128
+        bias: true
+        dropout: 0.1
+        dropout_in_middle: false
+    optimizer:
+      class_path: torch.optim.AdamW
+      init_args:
+        lr: 0.0001
+        betas:
+        - 0.9
+        - 0.95
+        eps: 1.0e-08
+        weight_decay: 0.01
+        amsgrad: false
+        maximize: false
+        foreach: null
+        capturable: false
+        differentiable: false
+        fused: null
+    lr_scheduler:
+      class_path: genbio_finetune.lr_schedulers.CosineWithWarmup
+      init_args:
+        warmup_ratio: 0.05
+    strict_loading: true
+    reset_optimizer_states: false
+data:
+  class_path: genbio_finetune.data.DMSFitnessPrediction
+  init_args:
+    path: genbio-ai/ProteinGYM-DMS
+    mutation_type: indels
+    task: BLAT_ECOLX_Gonzalez_2019_indels
+    normalize: true
+    train_split_name: train
+    test_split_files: null
+    valid_split_files: null
+    random_seed: 42
+    batch_size: 2
+    shuffle: true
+    sampler: null
+    num_workers: 0
+    pin_memory: true
+    persistent_workers: false
+    cv_num_folds: 5
+    cv_test_fold_id: 4
+    cv_enable_val_fold: true
+    cv_fold_id_col: fold_id
+ckpt_path: null

fold4/model.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7062398fd8c838ea432c0949c8cf8c53fbd93ad36fc8ecb18a7bb2a5373e6a5a
+size 147655192