Training in progress, epoch 1

.ipynb_checkpoints/hyperparameters-checkpoint.yaml

@@ -0,0 +1,160 @@
+__cached__setup_devices: !!python/object/apply:torch.device
+- cuda
+- 0
+_n_gpu: 1
+accelerator_config: !!python/object:transformers.trainer_pt_utils.AcceleratorConfig
+  dispatch_batches: null
+  even_batches: true
+  gradient_accumulation_kwargs: null
+  split_batches: false
+  use_seedable_sampler: true
+adafactor: false
+adam_beta1: 0.9
+adam_beta2: 0.999
+adam_epsilon: 1.0e-08
+auto_find_batch_size: false
+bf16: false
+bf16_full_eval: false
+data_seed: null
+dataloader_drop_last: false
+dataloader_num_workers: 0
+dataloader_persistent_workers: false
+dataloader_pin_memory: true
+dataloader_prefetch_factor: null
+ddp_backend: null
+ddp_broadcast_buffers: null
+ddp_bucket_cap_mb: null
+ddp_find_unused_parameters: null
+ddp_timeout: 1800
+debug: []
+deepspeed: null
+deepspeed_plugin: null
+disable_tqdm: false
+dispatch_batches: null
+distributed_state: !!python/object:accelerate.state.PartialState
+  _cpu: false
+  backend: null
+  debug: false
+  device: !!python/object/apply:torch.device
+  - cuda
+  distributed_type: !!python/object/apply:accelerate.utils.dataclasses.DistributedType
+  - 'NO'
+  fork_launched: false
+  local_process_index: 0
+  num_processes: 1
+  process_index: 0
+do_eval: true
+do_predict: false
+do_train: false
+eval_accumulation_steps: null
+eval_delay: 0
+eval_do_concat_batches: true
+eval_steps: null
+evaluation_strategy: !!python/object/apply:transformers.trainer_utils.IntervalStrategy
+- epoch
+fp16: true
+fp16_backend: auto
+fp16_full_eval: false
+fp16_opt_level: O1
+fsdp: []
+fsdp_config:
+  min_num_params: 0
+  xla: false
+  xla_fsdp_grad_ckpt: false
+  xla_fsdp_v2: false
+fsdp_min_num_params: 0
+fsdp_transformer_layer_cls_to_wrap: null
+full_determinism: false
+generation_config: null
+generation_max_length: null
+generation_num_beams: null
+gradient_accumulation_steps: 2
+gradient_checkpointing: false
+gradient_checkpointing_kwargs: null
+greater_is_better: false
+group_by_length: false
+half_precision_backend: auto
+hub_always_push: false
+hub_model_id: donut_experiment_bayesian_3
+hub_private_repo: false
+hub_strategy: !!python/object/apply:transformers.trainer_utils.HubStrategy
+- every_save
+hub_token: null
+ignore_data_skip: false
+include_inputs_for_metrics: false
+include_num_input_tokens_seen: false
+include_tokens_per_second: false
+jit_mode_eval: false
+label_names: null
+label_smoothing_factor: 0.0
+learning_rate: 1.7686069311870188e-05
+length_column_name: length
+load_best_model_at_end: true
+local_rank: 0
+log_level: passive
+log_level_replica: warning
+log_on_each_node: true
+logging_dir: model_runs/donut_experiment_bayesian_3/runs/May20_15-53-57_ip-172-16-167-107.ec2.internal
+logging_first_step: false
+logging_nan_inf_filter: true
+logging_steps: 100
+logging_strategy: !!python/object/apply:transformers.trainer_utils.IntervalStrategy
+- steps
+lr_scheduler_kwargs: {}
+lr_scheduler_type: !!python/object/apply:transformers.trainer_utils.SchedulerType
+- linear
+max_grad_norm: 1.0
+max_steps: -1
+metric_for_best_model: loss
+mp_parameters: ''
+neftune_noise_alpha: null
+no_cuda: false
+num_train_epochs: 5
+optim: !!python/object/apply:transformers.training_args.OptimizerNames
+- adamw_torch
+optim_args: null
+optim_target_modules: null
+output_dir: model_runs/donut_experiment_bayesian_3
+overwrite_output_dir: false
+past_index: -1
+per_device_eval_batch_size: 1
+per_device_train_batch_size: 1
+per_gpu_eval_batch_size: null
+per_gpu_train_batch_size: null
+predict_with_generate: true
+prediction_loss_only: false
+push_to_hub: true
+push_to_hub_model_id: null
+push_to_hub_organization: null
+push_to_hub_token: null
+ray_scope: last
+remove_unused_columns: true
+report_to:
+- tensorboard
+resume_from_checkpoint: null
+run_name: model_runs/donut_experiment_bayesian_3
+save_on_each_node: false
+save_only_model: false
+save_safetensors: true
+save_steps: 500
+save_strategy: !!python/object/apply:transformers.trainer_utils.IntervalStrategy
+- epoch
+save_total_limit: 2
+seed: 42
+skip_memory_metrics: true
+sortish_sampler: false
+split_batches: null
+tf32: null
+torch_compile: false
+torch_compile_backend: null
+torch_compile_mode: null
+torchdynamo: null
+tpu_metrics_debug: false
+tpu_num_cores: null
+use_cpu: false
+use_ipex: false
+use_legacy_prediction_loop: false
+use_mps_device: false
+warmup_ratio: 0.0
+warmup_steps: 0
+weight_decay: 0.017348694927892504

hyperparameters.yaml

@@ -87,14 +87,14 @@ include_tokens_per_second: false
 jit_mode_eval: false
 label_names: null
 label_smoothing_factor: 0.0
-learning_rate: 6.160988528485566e-05
+learning_rate: 1.7686069311870188e-05
 length_column_name: length
 load_best_model_at_end: true
 local_rank: 0
 log_level: passive
 log_level_replica: warning
 log_on_each_node: true
-logging_dir: model_runs/donut_experiment_bayesian_3/runs/May20_15-19-23_ip-172-16-167-107.ec2.internal
+logging_dir: model_runs/donut_experiment_bayesian_3/runs/May20_15-53-57_ip-172-16-167-107.ec2.internal
 logging_first_step: false
 logging_nan_inf_filter: true
 logging_steps: 100
@@ -109,7 +109,7 @@ metric_for_best_model: loss
 mp_parameters: ''
 neftune_noise_alpha: null
 no_cuda: false
-num_train_epochs: 3
+num_train_epochs: 5
 optim: !!python/object/apply:transformers.training_args.OptimizerNames
 - adamw_torch
 optim_args: null
@@ -157,4 +157,4 @@ use_legacy_prediction_loop: false
 use_mps_device: false
 warmup_ratio: 0.0
 warmup_steps: 0
-weight_decay: 0.00204406383034946
+weight_decay: 0.017348694927892504

metrics.jsonl

@@ -2,3 +2,5 @@
 {"eval_loss": 0.6344355940818787, "eval_bleu": 0.052030851722171666, "eval_precisions": [0.7112970711297071, 0.5890736342042755, 0.5357142857142857, 0.4935064935064935], "eval_brevity_penalty": 0.0901881044262292, "eval_length_ratio": 0.2936117936117936, "eval_translation_length": 478, "eval_reference_length": 1628, "eval_cer": 0.7785648277776906, "eval_wer": 0.8579665473564559, "eval_runtime": 71.1013, "eval_samples_per_second": 0.802, "eval_steps_per_second": 0.802, "epoch": 2.0}
 {"eval_loss": 0.438687801361084, "eval_bleu": 0.07087177541785662, "eval_precisions": [0.8065173116089613, 0.7396313364055299, 0.6923076923076923, 0.64375], "eval_brevity_penalty": 0.09869882068848057, "eval_length_ratio": 0.3015970515970516, "eval_translation_length": 491, "eval_reference_length": 1628, "eval_cer": 0.7633511730869632, "eval_wer": 0.8388264265354226, "eval_runtime": 70.9255, "eval_samples_per_second": 0.804, "eval_steps_per_second": 0.804, "epoch": 1.0}
 {"eval_loss": 0.3932441771030426, "eval_bleu": 0.05886392815904822, "eval_precisions": [0.7965738758029979, 0.7195121951219512, 0.6798866855524079, 0.6418918918918919], "eval_brevity_penalty": 0.08323549742104484, "eval_length_ratio": 0.2868550368550369, "eval_translation_length": 467, "eval_reference_length": 1628, "eval_cer": 0.7598741859679926, "eval_wer": 0.8348911497469601, "eval_runtime": 70.6916, "eval_samples_per_second": 0.806, "eval_steps_per_second": 0.806, "epoch": 2.0}
+{"eval_loss": 0.3638775944709778, "eval_bleu": 0.06460375602417452, "eval_precisions": [0.8266384778012685, 0.7644230769230769, 0.7158774373259053, 0.6721854304635762], "eval_brevity_penalty": 0.08699884249108106, "eval_length_ratio": 0.2905405405405405, "eval_translation_length": 473, "eval_reference_length": 1628, "eval_cer": 0.7583714021313483, "eval_wer": 0.8242161658213439, "eval_runtime": 70.8539, "eval_samples_per_second": 0.804, "eval_steps_per_second": 0.804, "epoch": 3.0}
+{"eval_loss": 0.40527409315109253, "eval_bleu": 0.07198724134583591, "eval_precisions": [0.8254620123203286, 0.7697674418604651, 0.7265415549597856, 0.6835443037974683], "eval_brevity_penalty": 0.09604717412371518, "eval_length_ratio": 0.29914004914004916, "eval_translation_length": 487, "eval_reference_length": 1628, "eval_cer": 0.7516296550897489, "eval_wer": 0.8241301963267271, "eval_runtime": 70.9733, "eval_samples_per_second": 0.803, "eval_steps_per_second": 0.803, "epoch": 1.0}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cd21d511928720b97335148ac274287b90f4a3fef762037f81805cf57f941ee7
+oid sha256:1836cfa19c4b15ca1ec3abc38bd9fb029c9291f0b0b9688ef223a06df9d86f49
 size 809103512

runs/May20_15-19-23_ip-172-16-167-107.ec2.internal/events.out.tfevents.1716218363.ip-172-16-167-107.ec2.internal.15611.1

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:24086fe3228358b2e026b5fa042eafa0b7406a1028d3ca00a530e72a7ed40af8
-size 11556
+oid sha256:2f765de2e9259797fbb7941b061d56bda0b9d52dd867d988ed1326cf0fbf60b2
+size 12982

runs/May20_15-53-57_ip-172-16-167-107.ec2.internal/events.out.tfevents.1716220437.ip-172-16-167-107.ec2.internal.15611.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72fb9baff0e448b4fb49ef9c60534c1aed5e6747ea33d3386cafcc287a65ec5d
+size 10275

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b55a72069efb82ec2b205b11b999e8c15c0191de0e6921704d174d9b2e8113e7
+oid sha256:c934cde04a1ae229cb790139fc3f62ef304d419d10a553ecd3993a4e4216cb91
 size 5240