Datasets:

---

bigbossmonster

/

mylora

like 0

Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 1 new columns ({'stop_text_encoder_training_pct'}) and 47 missing columns ({'txt_mlp_dim', 'discrete_flow_shift', 't5xxl_lr', 'log_config', 'blocks_to_swap', 'mem_eff_save', 'img_mlp_dim', 't5xxl_max_token_length', 'txt_mod_dim', 'img_attn_dim', 'save_last_n_epochs', 'cpu_offload_checkpointing', 'train_blocks', 'loraplus_lr_ratio', 'save_last_n_epochs_state', 'loraplus_text_encoder_lr_ratio', 'highvram', 'lr_warmup_steps', 'flux1_cache_text_encoder_outputs_to_disk', 'lowvram', 'double_blocks_to_swap', 'guidance_scale', 'ae', 'flux1_checkbox', 'loraplus_unet_lr_ratio', 'single_mod_dim', 'model_prediction_type', 'apply_t5_attn_mask', 'stop_text_encoder_training', 'fp8_base_unet', 'enable_all_linear', 'single_blocks_to_swap', 'train_single_block_indices', 'split_qkv', 'train_double_block_indices', 'txt_attn_dim', 'split_mode', 'in_dims', 'lr_scheduler_type', 'skip_cache_check', 'img_mod_dim', 'train_t5xxl', 'clip_l', 'flux1_cache_text_encoder_outputs', 't5xxl', 'single_dim', 'timestep_sampling'}).

This happened while the json dataset builder was generating data using

hf://datasets/bigbossmonster/mylora/Kana_Yume/Kana_Yume_kk_20240629-195319.json (at revision a08b94a4e179473f0642bf17832bda4a36081977)

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/builder.py", line 1870, in _prepare_split_single
                  writer.write_table(table)
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/arrow_writer.py", line 622, in write_table
                  pa_table = table_cast(pa_table, self._schema)
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/table.py", line 2292, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/table.py", line 2240, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              LoRA_type: string
              LyCORIS_preset: string
              adaptive_noise_scale: int64
              additional_parameters: string
              async_upload: bool
              block_alphas: string
              block_dims: string
              block_lr_zero_threshold: string
              bucket_no_upscale: bool
              bucket_reso_steps: int64
              bypass_mode: bool
              cache_latents: bool
              cache_latents_to_disk: bool
              caption_dropout_every_n_epochs: int64
              caption_dropout_rate: int64
              caption_extension: string
              clip_skip: int64
              color_aug: bool
              constrain: int64
              conv_alpha: int64
              conv_block_alphas: string
              conv_block_dims: string
              conv_dim: int64
              dataset_config: string
              debiased_estimation_loss: bool
              decompose_both: bool
              dim_from_weights: bool
              dora_wd: bool
              down_lr_weight: string
              dynamo_backend: string
              dynamo_mode: string
              dynamo_use_dynamic: bool
              dynamo_use_fullgraph: bool
              enable_bucket: bool
              epoch: int64
              extra_accelerate_launch_args: string
              factor: int64
              flip_aug: bool
              fp8_base: bool
              full_bf16: bool
              full_fp16: bool
              gpu_ids: string
              gradient_accumulation_steps: int64
              gradient_checkpointing: bool
              huber_c: double
              huber_schedule: string
              huggingface_path_in_repo: string
              huggingface_repo_id: string
              huggingface_repo_type: string
              huggingface_repo_visibility: string
              huggingface_token: string
              ip_noise_gamma: int64
              ip_noise_gamma_random_strength: bool
              keep_tokens: int64
              learning_rate: double
              log_tracker_config: string
              log_tracker_name: string
              log_with: string
              logging_dir: string
              loss_type: string
              lr_scheduler: string
              lr_scheduler_args: string
              lr_scheduler_num_cycles: int64
              lr_scheduler_power: int64
              lr_warmup:
              ...
              work_dim: int64
              network_dropout: int64
              network_weights: string
              noise_offset: double
              noise_offset_random_strength: bool
              noise_offset_type: string
              num_cpu_threads_per_process: int64
              num_machines: int64
              num_processes: int64
              optimizer: string
              optimizer_args: string
              output_dir: string
              output_name: string
              persistent_data_loader_workers: bool
              pretrained_model_name_or_path: string
              prior_loss_weight: int64
              random_crop: bool
              rank_dropout: int64
              rank_dropout_scale: bool
              reg_data_dir: string
              rescaled: bool
              resume: string
              resume_from_huggingface: string
              sample_every_n_epochs: int64
              sample_every_n_steps: int64
              sample_prompts: string
              sample_sampler: string
              save_every_n_epochs: int64
              save_every_n_steps: int64
              save_last_n_steps: int64
              save_last_n_steps_state: int64
              save_model_as: string
              save_precision: string
              save_state: bool
              save_state_on_train_end: bool
              save_state_to_huggingface: bool
              scale_v_pred_loss_like_noise_pred: bool
              scale_weight_norms: int64
              sdxl: bool
              sdxl_cache_text_encoder_outputs: bool
              sdxl_no_half_vae: bool
              seed: int64
              shuffle_caption: bool
              stop_text_encoder_training_pct: int64
              text_encoder_lr: double
              train_batch_size: int64
              train_data_dir: string
              train_norm: bool
              train_on_input: bool
              training_comment: string
              unet_lr: double
              unit: int64
              up_lr_weight: string
              use_cp: bool
              use_scalar: bool
              use_tucker: bool
              v2: bool
              v_parameterization: bool
              v_pred_like_loss: int64
              vae: string
              vae_batch_size: int64
              wandb_api_key: string
              wandb_run_name: string
              weighted_captions: bool
              xformers: string
              to
              {'LoRA_type': Value(dtype='string', id=None), 'LyCORIS_preset': Value(dtype='string', id=None), 'adaptive_noise_scale': Value(dtype='int64', id=None), 'additional_parameters': Value(dtype='string', id=None), 'ae': Value(dtype='string', id=None), 'apply_t5_attn_mask': Value(dtype='bool', id=None), 'async_upload': Value(dtype='bool', id=None), 'block_alphas': Value(dtype='string', id=None), 'block_dims': Value(dtype='string', id=None), 'block_lr_zero_threshold': Value(dtype='string', id=None), 'blocks_to_swap': Value(dtype='int64', id=None), 'bucket_no_upscale': Value(dtype='bool', id=None), 'bucket_reso_steps': Value(dtype='int64', id=None), 'bypass_mode': Value(dtype='bool', id=None), 'cache_latents': Value(dtype='bool', id=None), 'cache_latents_to_disk': Value(dtype='bool', id=None), 'caption_dropout_every_n_epochs': Value(dtype='int64', id=None), 'caption_dropout_rate': Value(dtype='int64', id=None), 'caption_extension': Value(dtype='string', id=None), 'clip_l': Value(dtype='string', id=None), 'clip_skip': Value(dtype='int64', id=None), 'color_aug': Value(dtype='bool', id=None), 'constrain': Value(dtype='int64', id=None), 'conv_alpha': Value(dtype='int64', id=None), 'conv_block_alphas': Value(dtype='string', id=None), 'conv_block_dims': Value(dtype='string', id=None), 'conv_dim': Value(dtype='int64', id=None), 'cpu_offload_checkpointing': Value(dtype='bool', id=None), 'dataset_config': Value(dtype='string', id=None), 'debiased_estimation_loss': Value(dtype='bool', id=None),
              ...
              , id=None), 't5xxl_lr': Value(dtype='int64', id=None), 't5xxl_max_token_length': Value(dtype='int64', id=None), 'text_encoder_lr': Value(dtype='float64', id=None), 'timestep_sampling': Value(dtype='string', id=None), 'train_batch_size': Value(dtype='int64', id=None), 'train_blocks': Value(dtype='string', id=None), 'train_data_dir': Value(dtype='string', id=None), 'train_double_block_indices': Value(dtype='string', id=None), 'train_norm': Value(dtype='bool', id=None), 'train_on_input': Value(dtype='bool', id=None), 'train_single_block_indices': Value(dtype='string', id=None), 'train_t5xxl': Value(dtype='bool', id=None), 'training_comment': Value(dtype='string', id=None), 'txt_attn_dim': Value(dtype='string', id=None), 'txt_mlp_dim': Value(dtype='string', id=None), 'txt_mod_dim': Value(dtype='string', id=None), 'unet_lr': Value(dtype='float64', id=None), 'unit': Value(dtype='int64', id=None), 'up_lr_weight': Value(dtype='string', id=None), 'use_cp': Value(dtype='bool', id=None), 'use_scalar': Value(dtype='bool', id=None), 'use_tucker': Value(dtype='bool', id=None), 'v2': Value(dtype='bool', id=None), 'v_parameterization': Value(dtype='bool', id=None), 'v_pred_like_loss': Value(dtype='int64', id=None), 'vae': Value(dtype='string', id=None), 'vae_batch_size': Value(dtype='int64', id=None), 'wandb_api_key': Value(dtype='string', id=None), 'wandb_run_name': Value(dtype='string', id=None), 'weighted_captions': Value(dtype='bool', id=None), 'xformers': Value(dtype='string', id=None)}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1415, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 991, in stream_convert_to_parquet
                  builder._prepare_split(
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/builder.py", line 1741, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/builder.py", line 1872, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 1 new columns ({'stop_text_encoder_training_pct'}) and 47 missing columns ({'txt_mlp_dim', 'discrete_flow_shift', 't5xxl_lr', 'log_config', 'blocks_to_swap', 'mem_eff_save', 'img_mlp_dim', 't5xxl_max_token_length', 'txt_mod_dim', 'img_attn_dim', 'save_last_n_epochs', 'cpu_offload_checkpointing', 'train_blocks', 'loraplus_lr_ratio', 'save_last_n_epochs_state', 'loraplus_text_encoder_lr_ratio', 'highvram', 'lr_warmup_steps', 'flux1_cache_text_encoder_outputs_to_disk', 'lowvram', 'double_blocks_to_swap', 'guidance_scale', 'ae', 'flux1_checkbox', 'loraplus_unet_lr_ratio', 'single_mod_dim', 'model_prediction_type', 'apply_t5_attn_mask', 'stop_text_encoder_training', 'fp8_base_unet', 'enable_all_linear', 'single_blocks_to_swap', 'train_single_block_indices', 'split_qkv', 'train_double_block_indices', 'txt_attn_dim', 'split_mode', 'in_dims', 'lr_scheduler_type', 'skip_cache_check', 'img_mod_dim', 'train_t5xxl', 'clip_l', 'flux1_cache_text_encoder_outputs', 't5xxl', 'single_dim', 'timestep_sampling'}).
              
              This happened while the json dataset builder was generating data using
              
              hf://datasets/bigbossmonster/mylora/Kana_Yume/Kana_Yume_kk_20240629-195319.json (at revision a08b94a4e179473f0642bf17832bda4a36081977)
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)