wandb_version: 1 _wandb: desc: null value: python_version: 3.11.5 cli_version: 0.16.1 framework: huggingface huggingface_version: 4.36.2 is_jupyter_run: false is_kaggle_kernel: false start_time: 1706810330.083967 t: 1: - 1 - 2 - 3 - 5 - 11 - 49 - 51 - 53 - 55 - 71 - 84 - 98 2: - 1 - 2 - 3 - 5 - 11 - 49 - 51 - 53 - 55 - 71 - 84 - 98 3: - 7 - 23 4: 3.11.5 5: 0.16.1 6: 4.36.2 8: - 5 9: 1: transformers_trainer 13: linux-x86_64 m: - 1: train/global_step 6: - 3 - 1: train/loss 5: 1 6: - 1 - 1: train/learning_rate 5: 1 6: - 1 - 1: train/epoch 5: 1 6: - 1 - 1: eval/loss 5: 1 6: - 1 - 1: eval/runtime 5: 1 6: - 1 - 1: eval/samples_per_second 5: 1 6: - 1 - 1: eval/steps_per_second 5: 1 6: - 1 - 1: train/train_runtime 5: 1 6: - 1 - 1: train/train_samples_per_second 5: 1 6: - 1 - 1: train/train_steps_per_second 5: 1 6: - 1 - 1: train/total_flos 5: 1 6: - 1 - 1: train/train_loss 5: 1 6: - 1 vocab_size: desc: null value: 32000 max_position_embeddings: desc: null value: 32768 hidden_size: desc: null value: 4096 intermediate_size: desc: null value: 14336 num_hidden_layers: desc: null value: 6 num_attention_heads: desc: null value: 32 sliding_window: desc: null value: 4096 num_key_value_heads: desc: null value: 8 hidden_act: desc: null value: silu initializer_range: desc: null value: 0.02 rms_norm_eps: desc: null value: 1.0e-05 use_cache: desc: null value: false rope_theta: desc: null value: 10000.0 attention_dropout: desc: null value: 0.0 return_dict: desc: null value: true output_hidden_states: desc: null value: false output_attentions: desc: null value: false torchscript: desc: null value: false torch_dtype: desc: null value: bfloat16 use_bfloat16: desc: null value: false tf_legacy_loss: desc: null value: false pruned_heads: desc: null value: {} tie_word_embeddings: desc: null value: false is_encoder_decoder: desc: null value: false is_decoder: desc: null value: false cross_attention_hidden_size: desc: null value: null add_cross_attention: desc: null value: false tie_encoder_decoder: desc: null value: false max_length: desc: null value: 20 min_length: desc: null value: 0 do_sample: desc: null value: false early_stopping: desc: null value: false num_beams: desc: null value: 1 num_beam_groups: desc: null value: 1 diversity_penalty: desc: null value: 0.0 temperature: desc: null value: 1.0 top_k: desc: null value: 50 top_p: desc: null value: 1.0 typical_p: desc: null value: 1.0 repetition_penalty: desc: null value: 1.0 length_penalty: desc: null value: 1.0 no_repeat_ngram_size: desc: null value: 0 encoder_no_repeat_ngram_size: desc: null value: 0 bad_words_ids: desc: null value: null num_return_sequences: desc: null value: 1 chunk_size_feed_forward: desc: null value: 0 output_scores: desc: null value: false return_dict_in_generate: desc: null value: false forced_bos_token_id: desc: null value: null forced_eos_token_id: desc: null value: null remove_invalid_values: desc: null value: false exponential_decay_length_penalty: desc: null value: null suppress_tokens: desc: null value: null begin_suppress_tokens: desc: null value: null architectures: desc: null value: - MistralForCausalLM finetuning_task: desc: null value: null id2label: desc: null value: '0': LABEL_0 '1': LABEL_1 label2id: desc: null value: LABEL_0: 0 LABEL_1: 1 tokenizer_class: desc: null value: null prefix: desc: null value: null bos_token_id: desc: null value: 1 pad_token_id: desc: null value: null eos_token_id: desc: null value: 2 sep_token_id: desc: null value: null decoder_start_token_id: desc: null value: null task_specific_params: desc: null value: null problem_type: desc: null value: null _name_or_path: desc: null value: sanchit-gandhi/Mistral-7B-v0.1-6-layer transformers_version: desc: null value: 4.36.2 model_type: desc: null value: mistral output_dir: desc: null value: ./ overwrite_output_dir: desc: null value: true do_train: desc: null value: false do_eval: desc: null value: true do_predict: desc: null value: false evaluation_strategy: desc: null value: epoch prediction_loss_only: desc: null value: false per_device_train_batch_size: desc: null value: 64 per_device_eval_batch_size: desc: null value: 32 per_gpu_train_batch_size: desc: null value: null per_gpu_eval_batch_size: desc: null value: null gradient_accumulation_steps: desc: null value: 1 eval_accumulation_steps: desc: null value: null eval_delay: desc: null value: 0 learning_rate: desc: null value: 0.0003 weight_decay: desc: null value: 0.0 adam_beta1: desc: null value: 0.9 adam_beta2: desc: null value: 0.999 adam_epsilon: desc: null value: 1.0e-08 max_grad_norm: desc: null value: 1.0 num_train_epochs: desc: null value: 5 max_steps: desc: null value: -1 lr_scheduler_type: desc: null value: cosine lr_scheduler_kwargs: desc: null value: {} warmup_ratio: desc: null value: 0.1 warmup_steps: desc: null value: 0 log_level: desc: null value: info log_level_replica: desc: null value: warning log_on_each_node: desc: null value: true logging_dir: desc: null value: ./runs/Feb01_17-58-13_ip-26-0-165-24 logging_strategy: desc: null value: steps logging_first_step: desc: null value: true logging_steps: desc: null value: 10 logging_nan_inf_filter: desc: null value: true save_strategy: desc: null value: epoch save_steps: desc: null value: 500 save_total_limit: desc: null value: 1 save_safetensors: desc: null value: true save_on_each_node: desc: null value: false save_only_model: desc: null value: false no_cuda: desc: null value: false use_cpu: desc: null value: false use_mps_device: desc: null value: false seed: desc: null value: 42 data_seed: desc: null value: null jit_mode_eval: desc: null value: false use_ipex: desc: null value: false bf16: desc: null value: true fp16: desc: null value: false fp16_opt_level: desc: null value: O1 half_precision_backend: desc: null value: auto bf16_full_eval: desc: null value: false fp16_full_eval: desc: null value: false tf32: desc: null value: null local_rank: desc: null value: 0 ddp_backend: desc: null value: null tpu_num_cores: desc: null value: null tpu_metrics_debug: desc: null value: false debug: desc: null value: [] dataloader_drop_last: desc: null value: false eval_steps: desc: null value: null dataloader_num_workers: desc: null value: 0 past_index: desc: null value: -1 run_name: desc: null value: ./ disable_tqdm: desc: null value: false remove_unused_columns: desc: null value: true label_names: desc: null value: null load_best_model_at_end: desc: null value: false metric_for_best_model: desc: null value: null greater_is_better: desc: null value: null ignore_data_skip: desc: null value: false fsdp: desc: null value: [] fsdp_min_num_params: desc: null value: 0 fsdp_config: desc: null value: min_num_params: 0 xla: false xla_fsdp_grad_ckpt: false fsdp_transformer_layer_cls_to_wrap: desc: null value: null deepspeed: desc: null value: null label_smoothing_factor: desc: null value: 0.0 optim: desc: null value: adamw_torch optim_args: desc: null value: null adafactor: desc: null value: false group_by_length: desc: null value: false length_column_name: desc: null value: length report_to: desc: null value: - tensorboard - wandb ddp_find_unused_parameters: desc: null value: null ddp_bucket_cap_mb: desc: null value: null ddp_broadcast_buffers: desc: null value: null dataloader_pin_memory: desc: null value: true dataloader_persistent_workers: desc: null value: false skip_memory_metrics: desc: null value: true use_legacy_prediction_loop: desc: null value: false push_to_hub: desc: null value: true resume_from_checkpoint: desc: null value: null hub_model_id: desc: null value: null hub_strategy: desc: null value: every_save hub_token: desc: null value: hub_private_repo: desc: null value: false hub_always_push: desc: null value: false gradient_checkpointing: desc: null value: true gradient_checkpointing_kwargs: desc: null value: use_reentrant: false include_inputs_for_metrics: desc: null value: false fp16_backend: desc: null value: auto push_to_hub_model_id: desc: null value: null push_to_hub_organization: desc: null value: null push_to_hub_token: desc: null value: mp_parameters: desc: null value: '' auto_find_batch_size: desc: null value: false full_determinism: desc: null value: false torchdynamo: desc: null value: null ray_scope: desc: null value: last ddp_timeout: desc: null value: 1800 torch_compile: desc: null value: false torch_compile_backend: desc: null value: null torch_compile_mode: desc: null value: null dispatch_batches: desc: null value: null split_batches: desc: null value: false include_tokens_per_second: desc: null value: false include_num_input_tokens_seen: desc: null value: false neftune_noise_alpha: desc: null value: null max_seq_length: desc: null value: 2048