| --- |
| dataset_info: |
| - config_name: hyperparameters__sft |
| features: |
| - name: model_name_or_path |
| dtype: string |
| - name: trust_remote_code |
| dtype: bool |
| - name: stage |
| dtype: string |
| - name: do_train |
| dtype: bool |
| - name: finetuning_type |
| dtype: string |
| - name: deepspeed |
| dtype: string |
| - name: dataset |
| dtype: string |
| - name: template |
| dtype: string |
| - name: cutoff_len |
| dtype: int64 |
| - name: max_samples |
| dtype: int64 |
| - name: overwrite_cache |
| dtype: bool |
| - name: preprocessing_num_workers |
| dtype: int64 |
| - name: dataloader_num_workers |
| dtype: int64 |
| - name: disable_tqdm |
| dtype: bool |
| - name: output_dir |
| dtype: string |
| - name: logging_steps |
| dtype: int64 |
| - name: save_steps |
| dtype: int64 |
| - name: plot_loss |
| dtype: bool |
| - name: overwrite_output_dir |
| dtype: bool |
| - name: per_device_train_batch_size |
| dtype: int64 |
| - name: gradient_accumulation_steps |
| dtype: int64 |
| - name: learning_rate |
| dtype: float64 |
| - name: num_train_epochs |
| dtype: int64 |
| - name: lr_scheduler_type |
| dtype: string |
| - name: warmup_ratio |
| dtype: float64 |
| - name: weight_decay |
| dtype: float64 |
| - name: adam_beta1 |
| dtype: float64 |
| - name: adam_beta2 |
| dtype: float64 |
| - name: bf16 |
| dtype: bool |
| - name: ddp_timeout |
| dtype: int64 |
| - name: gradient_checkpointing |
| dtype: bool |
| - name: save_only_model |
| dtype: bool |
| - name: enable_masked_ranges |
| dtype: bool |
| - name: save_strategy |
| dtype: string |
| - name: save_total_limit |
| dtype: int64 |
| - name: sf_tracker_dataset_id |
| dtype: string |
| - name: sf_eval_before_training |
| dtype: bool |
| - name: sf_wandb_project |
| dtype: string |
| - name: sf_eval_steps |
| dtype: 'null' |
| - name: run_name |
| dtype: string |
| splits: |
| - name: train |
| num_bytes: 565 |
| num_examples: 1 |
| download_size: 17887 |
| dataset_size: 565 |
| - config_name: logs__llamafactory_sft |
| features: |
| - name: timestamp |
| dtype: string |
| - name: end_timestamp |
| dtype: string |
| - name: stage_name |
| dtype: string |
| - name: stage_number |
| dtype: int64 |
| - name: level |
| dtype: string |
| - name: message |
| dtype: string |
| - name: stdout_content |
| dtype: string |
| - name: stderr_content |
| dtype: string |
| - name: experiment_name |
| dtype: string |
| - name: elapsed_time_seconds |
| dtype: float64 |
| - name: stage_complete |
| dtype: bool |
| splits: |
| - name: train |
| num_bytes: 355473 |
| num_examples: 1 |
| download_size: 119262 |
| dataset_size: 355473 |
| - config_name: metadata |
| features: |
| - name: experiment_name |
| dtype: string |
| - name: start_time |
| dtype: string |
| - name: description |
| dtype: string |
| - name: base_org |
| dtype: string |
| - name: stage_number |
| dtype: string |
| - name: stage_type |
| dtype: string |
| - name: status |
| dtype: string |
| splits: |
| - name: train |
| num_bytes: 2970 |
| num_examples: 8 |
| download_size: 4581 |
| dataset_size: 2970 |
| - config_name: training_data__sft |
| features: |
| - name: split |
| dtype: string |
| - name: example_idx |
| dtype: int64 |
| - name: stage_name |
| dtype: string |
| - name: timestamp |
| dtype: string |
| - name: conversations |
| list: |
| - name: content |
| dtype: string |
| - name: role |
| dtype: string |
| - name: question |
| dtype: string |
| - name: source |
| dtype: 'null' |
| - name: id |
| dtype: 'null' |
| - name: prompt |
| list: |
| - name: content |
| dtype: string |
| - name: role |
| dtype: string |
| - name: answer |
| dtype: string |
| - name: prompt_type |
| dtype: string |
| - name: model_responses__example_annotation |
| sequence: string |
| - name: model_responses__example_annotation__finish_reason_length_flags |
| sequence: bool |
| - name: model_responses__example_annotation__length_partial_responses |
| sequence: string |
| - name: prompt__example_annotation__metadata |
| struct: |
| - name: api_url |
| dtype: string |
| - name: backend |
| dtype: string |
| - name: chat_template_applied |
| dtype: bool |
| - name: generation_params |
| struct: |
| - name: chat_template_applied |
| dtype: bool |
| - name: max_tokens |
| dtype: int64 |
| - name: n |
| dtype: int64 |
| - name: repetition_penalty |
| dtype: float64 |
| - name: temperature |
| dtype: float64 |
| - name: top_k |
| dtype: int64 |
| - name: top_p |
| dtype: float64 |
| - name: model_name |
| dtype: string |
| - name: prompt |
| list: |
| - name: content |
| dtype: string |
| - name: role |
| dtype: string |
| - name: model_responses__example_annotation__metadata |
| struct: |
| - name: backend |
| dtype: string |
| - name: model_name |
| dtype: string |
| - name: n_responses |
| dtype: int64 |
| - name: task_source |
| dtype: string |
| - name: task_config |
| dtype: string |
| - name: model_responses__example_annotation__eval_is_correct |
| sequence: bool |
| - name: model_responses__example_annotation__eval_extracted_answers |
| sequence: string |
| - name: model_responses__example_annotation__eval_extraction_metadata |
| dtype: string |
| - name: model_responses__example_annotation__eval_evaluation_metadata |
| dtype: string |
| - name: model_responses__example_annotation__internal_answers__eval_is_correct |
| sequence: |
| sequence: bool |
| - name: model_responses__example_annotation__internal_answers__eval_extracted_answers |
| sequence: |
| sequence: string |
| - name: model_responses__example_annotation__internal_answers__eval_extraction_metadata |
| dtype: string |
| - name: model_responses__example_annotation__internal_answers__eval_evaluation_metadata |
| dtype: string |
| - name: model_responses__example_annotation__metrics |
| struct: |
| - name: flips_by |
| sequence: int64 |
| - name: flips_total |
| dtype: int64 |
| - name: num_correct |
| dtype: int64 |
| - name: pass_at_n |
| dtype: int64 |
| - name: percent_correct |
| dtype: float64 |
| - name: total_responses |
| dtype: int64 |
| splits: |
| - name: train |
| num_bytes: 19268914 |
| num_examples: 1000 |
| download_size: 7334939 |
| dataset_size: 19268914 |
| - config_name: training_data__sft_metadata |
| features: |
| - name: stage_name |
| dtype: string |
| - name: stage_number |
| dtype: int64 |
| - name: timestamp |
| dtype: string |
| - name: original_dataset_id |
| dtype: string |
| - name: dataset_type |
| dtype: string |
| - name: usage |
| dtype: string |
| - name: dataset_info |
| struct: |
| - name: features |
| struct: |
| - name: train |
| dtype: string |
| - name: num_examples |
| struct: |
| - name: train |
| dtype: int64 |
| - name: splits |
| sequence: string |
| splits: |
| - name: train |
| num_bytes: 2945 |
| num_examples: 1 |
| download_size: 16690 |
| dataset_size: 2945 |
| configs: |
| - config_name: hyperparameters__sft |
| data_files: |
| - split: train |
| path: hyperparameters__sft/train-* |
| - config_name: logs__llamafactory_sft |
| data_files: |
| - split: train |
| path: logs__llamafactory_sft/train-* |
| - config_name: metadata |
| data_files: |
| - split: train |
| path: metadata/train-* |
| - config_name: training_data__sft |
| data_files: |
| - split: train |
| path: training_data__sft/train-* |
| - config_name: training_data__sft_metadata |
| data_files: |
| - split: train |
| path: training_data__sft_metadata/train-* |
| --- |
| # Experiment Tracker: OT_otdata_1k_SFT |
| |
| **Experiment Description:** Simple test experiment for Skill Factory workflows. |
| |
| **Start Time:** 2025-11-18T04:47:29.743724 |
| |
| **Tracker Dataset:** [SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1](https://huggingface.co/datasets/SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1) |
| |
| ## Stages Completed |
| |
| Total stages: 1 |
| |
| ## Models Created |
| |
| - **sft**: [SkillFactory/M-OT_otdata_1k_SFT-sft](https://huggingface.co/SkillFactory/M-OT_otdata_1k_SFT-sft) |
| |
| ## Dataset Configurations |
| |
| This tracker dataset contains the following configurations with **immediate upload** as stages complete: |
| |
| ### Training Data (Complete Datasets) |
| |
| |
| |
| ### Hyperparameters (Complete Configurations) |
| |
| |
| ### Logs (Stage-Specific) |
| |
| |
| ### Evaluation Results (Complete with Annotations) |
| |
| |
| ### Metadata |
| - **experiment_metadata**: Timeline and stage information |
| |
| ## Usage |
| |
| Load specific configurations with: |
| |
| ```python |
| from datasets import load_dataset |
|
|
| # Load experiment metadata |
| metadata = load_dataset('SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1', 'experiment_metadata') |
|
|
| # Load complete training datasets |
| sft_data = load_dataset('SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1', 'training_data__sft') |
| sft_metadata = load_dataset('SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1', 'training_data__sft_metadata') |
| |
| # Load complete configurations |
| sft_hyperparams = load_dataset('SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1', 'hyperparameters__sft') |
| rl_hyperparams = load_dataset('SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1', 'hyperparameters__rl') |
|
|
| # Load stage-specific logs |
| sft_logs = load_dataset('SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1', 'logs__sft') |
| rl_logs = load_dataset('SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1', 'logs__rl') |
| |
| # Load evaluation results with annotations |
| sft_eval_results = load_dataset('SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1', 'evals_eval_sft') |
| rl_eval_results = load_dataset('SkillFactory/D-ExpTracker__OT_otdata_1k_SFT__v1', 'evals_eval_rl') |
| ``` |
| |
| ## Models |
| |
| - [SkillFactory/M-OT_otdata_1k_SFT-sft](https://huggingface.co/SkillFactory/M-OT_otdata_1k_SFT-sft) |
| |
| ## Registry |
| |
| All models from this experiment are automatically registered in the [SkillFactory Model Registry](https://huggingface.co/datasets/TAUR-dev/SkillFactory-Registration) with: |
| |
| - **Complete training configuration** (hyperparameters, datasets, methods) |
| - **Experiment lineage** (links back to this tracker dataset) |
| - **Stage-specific metadata** (SFT vs RL training details) |
| - **Structured input data references** (training datasets and configurations) |
| |
| Registry entries follow the naming pattern: `Model - OT_otdata_1k_SFT - {stage_name} - {SFT/RL}` |
| |
| --- |
| *Generated by SkillFactory Experiment Management System* |
| *All artifacts uploaded immediately as stages complete with perfect data provenance* |
| |
