| onescience_manifest_version: "0.1" |
| resource_type: dataset |
| resource: |
| id: "OneScience/evo2_dataset" |
| name: "Evo2 mini genome dataset" |
| domain: bio |
| domain_tags: |
| - bio |
| - genomics |
| - dna_sequence |
| task: genome_sequence_modeling_data |
| task_tags: |
| - train_data |
| - validation_data |
| - test_data |
| - preflight |
| modalities: |
| - dna_sequence |
| input_formats: |
| - fasta |
| - fasta_gzip |
| - indexed_dataset_bin_idx |
| output_formats: [] |
| summary: "适配 OneScience/evo2/ 的 chr20/chr21/chr22 mini 基因组数据集，包含 FASTA 原始文件和 Byte-Level 预处理 bin/idx split。" |
|
|
| platform_resource: |
| primary: |
| platform: modelscope |
| repo_id: "OneScience/evo2_dataset" |
| repo_type: dataset |
| url: "https://modelscope.cn/datasets/OneScience/evo2_dataset" |
| revision: main |
| readme_path: README.md |
| manifest_path: manifest.yaml |
| mirrors: [] |
| access: |
| visibility: public |
| license: "Dataset license follows source genome data terms" |
|
|
| website_integration: |
| enabled: true |
| click_target: |
| platform: modelscope |
| resource_url: "https://modelscope.cn/datasets/OneScience/evo2_dataset" |
| llm_handoff: |
| readme_required: true |
| manifest_required: true |
| download_readme_first: true |
| resolve_related_models: true |
| default_run_goal: minimal_validation |
| cwd_rule: "如果使用 modelscope download --cache_dir 下载，运行 cwd 必须切换到实际下载后的数据集包根目录，即包含 README.md 和 manifest.yaml 的目录。" |
|
|
| runtime: |
| enabled: true |
| onescience_domain: bio |
| supported_execution: |
| - local |
| - local_slurm |
| - remote_slurm |
| - scnet_mcp |
|
|
| onescience: |
| repo: "https://gitee.com/onescience-ai/onescience" |
| official_links: |
| gitee: |
| doc: "https://gitee.com/onescience-ai/onescience-doc" |
| onescience: "https://gitee.com/onescience-ai/onescience" |
| skills: "https://gitee.com/onescience-ai/oneskills" |
| github: |
| doc: "https://github.com/onescience-ai/OneScience-doc" |
| onescience: "https://github.com/onescience-ai/OneScience" |
| skills: "https://github.com/onescience-ai/oneskills" |
| install: |
| required_by_default: false |
| command: "pip install -c constraints.txt .[bio]" |
| source_paths: [] |
| compatibility: |
| examples_path: "examples/biosciences/evo2" |
| status: examples_compatible |
|
|
| runtime_package: |
| kind: standard_runtime_package |
| package_root: . |
| standard_layout: |
| workdir: . |
| config_dir: null |
| checkpoint_dir: null |
| data_dir: data_mini |
| output_dir: outputs |
| apply_policy: |
| mode: direct_use |
| target: session_workdir |
| overwrite: true |
| protect_installed_onescience: true |
| entry_files: |
| - scripts/validate_evo2_dataset.py |
| - data_mini/genome_data/ |
| - metadata/data_mini.files.tsv |
| - metadata/data_mini.sha256 |
| entrypoints: |
| preflight: scripts/validate_evo2_dataset.py |
| inference: null |
| train: null |
| finetune: null |
| evaluate: null |
| visualize: null |
| deploy: null |
|
|
| files: |
| model_files: [] |
| dataset_files: |
| - id: data_mini |
| path: data_mini/ |
| role: train_eval_test_data |
| format: fasta_and_indexed_dataset |
| description_zh: "Evo2 mini 数据根目录，包含 genome_data 下的 chr20/chr21/chr22 FASTA、压缩 FASTA、合并 FASTA 和预处理 bin/idx split。" |
| required: true |
| required_for: |
| - preflight |
| - train |
| - predict |
| source: |
| platform: modelscope |
| repo_id: "OneScience/evo2_dataset" |
| repo_type: dataset |
| path: data_mini/ |
| revision: main |
| download_method: command_ref |
| command_ref: commands.download.download_dataset |
| local_path: data_mini/ |
| checksum_manifest: |
| sha256: metadata/data_mini.sha256 |
| sizes: metadata/data_mini.files.tsv |
| - id: genome_fasta |
| path: data_mini/genome_data/ |
| role: raw_fasta |
| format: fasta |
| description_zh: "chr20.fa、chr21.fa、chr22.fa 和 chr20_21_22.fa 原始序列文件。" |
| required: true |
| required_for: |
| - preflight |
| - preprocess |
| - predict |
| source: |
| platform: modelscope |
| repo_id: "OneScience/evo2_dataset" |
| repo_type: dataset |
| path: data_mini/genome_data/ |
| revision: main |
| download_method: command_ref |
| command_ref: commands.download.download_dataset |
| local_path: data_mini/genome_data/ |
| - id: indexed_splits |
| path: data_mini/genome_data/preprocessed_data/ |
| role: indexed_dataset |
| format: bin_idx |
| description_zh: "训练、验证、测试 split 的 chr20_21_22_uint8_distinct_byte-level_*.bin 和 *.idx。" |
| required: true |
| required_for: |
| - preflight |
| - train |
| source: |
| platform: modelscope |
| repo_id: "OneScience/evo2_dataset" |
| repo_type: dataset |
| path: data_mini/genome_data/preprocessed_data/ |
| revision: main |
| download_method: command_ref |
| command_ref: commands.download.download_dataset |
| local_path: data_mini/genome_data/preprocessed_data/ |
| config_files: [] |
| sample_files: [] |
|
|
| dataset: |
| format: |
| - fasta |
| - fasta_gzip |
| - bin_idx |
| sample_unit: "DNA 序列片段或 Byte-Level indexed dataset 样本。" |
| schema: |
| path: metadata/data_mini.files.tsv |
| fields: |
| - relative_path |
| - size_bytes |
| - sha256 |
| splits: |
| train: data_mini/genome_data/preprocessed_data/chr20_21_22_uint8_distinct_byte-level_train |
| validation: data_mini/genome_data/preprocessed_data/chr20_21_22_uint8_distinct_byte-level_val |
| test: data_mini/genome_data/preprocessed_data/chr20_21_22_uint8_distinct_byte-level_test |
|
|
| relations: |
| required_datasets: [] |
| optional_datasets: [] |
| compatible_models: |
| - id: "OneScience/evo2/" |
| role: train_data |
| required_for: |
| - preflight |
| - train |
| - predict |
| resource_ref: |
| platform: modelscope |
| repo_id: "OneScience/evo2/" |
| repo_type: model |
| url: "https://modelscope.cn/models/OneScience/evo2/" |
| revision: main |
| readme_path: README.md |
| manifest_path: manifest.yaml |
| expected_local_path: data/evo2_dataset/data_mini/genome_data/ |
| adapter: "模型包 config/genome_data_config.yaml 和 config/opengenome2.yml" |
|
|
| run_matrix: |
| scenarios: |
| - name: minimal_validation |
| default: true |
| capability: preflight |
| description_zh: "校验数据集目录结构、FASTA 可读性、bin/idx split、文件大小和 SHA256。" |
| required_dataset_files: |
| - data_mini |
| - genome_fasta |
| - indexed_splits |
| required_model_files: [] |
| required_datasets: [] |
| preconditions: |
| - "已下载数据集仓库 OneScience/evo2_dataset。" |
| - "cwd 位于数据集包根目录。" |
| command_refs: |
| - commands.preflight.validate_dataset |
| outputs: |
| - "终端输出 [OK] Evo2 mini dataset..." |
| - name: model_train_input |
| capability: train |
| description_zh: "作为 OneScience/evo2/ 训练或微调输入数据。" |
| required_dataset_files: |
| - indexed_splits |
| required_model_files: [] |
| required_datasets: [] |
| preconditions: |
| - "将 data_mini 复制或链接到模型包 data/evo2_dataset/data_mini。" |
| command_refs: |
| - commands.prepare.place_for_model |
| outputs: |
| - "<MODEL_REPO_ROOT>/data/evo2_dataset/data_mini" |
|
|
| capabilities: |
| inference: false |
| train: true |
| finetune: true |
| evaluate: false |
| visualize: false |
| deploy: false |
| preflight: true |
| data_reading: true |
|
|
| commands: |
| download: |
| - name: download_dataset |
| cwd: . |
| command: "modelscope download --dataset OneScience/evo2_dataset" |
| description_zh: "下载数据集标准包。若使用 --cache_dir，下载完成后必须 cd 到包含 README.md 和 manifest.yaml 的数据集包根目录。" |
| - name: download_model |
| cwd: . |
| command: "modelscope download --model OneScience/evo2/" |
| description_zh: "下载兼容模型 OneScience/evo2/。" |
| prepare: |
| - name: place_for_model |
| cwd: . |
| command: "mkdir -p <MODEL_REPO_ROOT>/data/evo2_dataset && cp -a data_mini <MODEL_REPO_ROOT>/data/evo2_dataset/data_mini" |
| description_zh: "把数据集放到模型包默认数据路径。" |
| preflight: |
| - name: validate_dataset |
| cwd: . |
| command: "python scripts/validate_evo2_dataset.py --package-root . --dataset-root data_mini" |
| description_zh: "检查数据结构、FASTA 可读性、预处理 split、大小和 SHA256。" |
| inference: [] |
| train: [] |
| finetune: [] |
| evaluate: [] |
| visualize: [] |
| deploy: [] |
|
|
| expected_outputs: |
| - id: dataset_validation_ok |
| path: stdout |
| description_zh: "数据集验证成功时输出 [OK] Evo2 mini dataset..." |
| produced_by: commands.preflight.validate_dataset |
| - id: placed_dataset |
| path: "<MODEL_REPO_ROOT>/data/evo2_dataset/data_mini" |
| description_zh: "供 OneScience/evo2/ 默认预检和训练命令读取的数据目录。" |
| produced_by: commands.prepare.place_for_model |
|
|
| diagnostics: |
| - id: missing_fasta |
| symptom: "missing FASTA file" |
| likely_cause_zh: "data_mini 未完整下载或目录位置不对。" |
| fix_zh: "重新运行 modelscope download --dataset OneScience/evo2_dataset，或用 --dataset-root 指向实际 data_mini。" |
| - id: missing_bin_idx |
| symptom: "missing preprocessed bin or idx" |
| likely_cause_zh: "预处理 split 文件缺失。" |
| fix_zh: "重新下载数据集，确认 data_mini/genome_data/preprocessed_data 下有 train/val/test 的 .bin 和 .idx。" |
| - id: sha256_mismatch |
| symptom: "sha256 mismatch" |
| likely_cause_zh: "上传、下载或复制过程中数据损坏。" |
| fix_zh: "删除损坏文件后重新下载 OneScience/evo2_dataset。" |
| - id: model_path_mismatch |
| symptom: "model preflight cannot find data/evo2_dataset/data_mini" |
| likely_cause_zh: "数据集没有放到模型包预期路径。" |
| fix_zh: "在模型包根目录执行 mkdir -p data/evo2_dataset && cp -a <DATASET_REPO_ROOT>/data_mini data/evo2_dataset/data_mini。" |
|
|
| domain_extension: |
| bio: |
| organism_scope_zh: "人类 chr20、chr21、chr22 mini 基因组数据。" |
| sequence_alphabet: "DNA / Byte-Level tokenizer" |
| raw_files: |
| - data_mini/genome_data/chr20.fa |
| - data_mini/genome_data/chr21.fa |
| - data_mini/genome_data/chr22.fa |
| - data_mini/genome_data/chr20_21_22.fa |
| indexed_dataset_dtype: uint8 |
| preprocessing_config_in_model: "OneScience/evo2/ config/genome_preprocess_config.yaml" |
|
|
