Quickstart¶
This guide walks through running your first experiment. Make sure you have installed pretrain-experiments and a training framework.
The config file¶
Experiments are defined in a single YAML file. The following example inserts ARC-Challenge benchmark questions into OLMo-3 7B midtraining data and evaluates how much the model overfits on them.
experiment: example-experiments
wandb:
name: olmo-3-midtrain
entity: your-entity
framework: olmo_core
model:
config: ${OLMO_CORE_REPO}/src/scripts/official/OLMo3/OLMo-3-1025-7B-midtrain.py
checkpoint_url: "https://olmo-checkpoints.org/ai2-llm/Olmo-3-1025-7B/stage2/"
checkpoint_step: 10000
training:
num_steps: 100
experiments:
experiments:
- type: add-texts-from-file
file: ${PRETRAIN_EXPERIMENTS}/resources/.../olmes_arc_challenge_test.jsonl
repetitions: 4
evaluation:
eval_on_load: true
evaluations:
- script: olmes.py
args:
task: arc_challenge::olmes
split: test
The config specifies a model checkpoint to continue training from, data interventions to apply, and evaluations to run. Environment variables (${...}) are substituted at runtime.
The data file¶
Texts to insert are stored as JSONL — one JSON object per line with a "text" field:
{"text": "Question: An astronomer observes that a planet rotates faster after a meteorite impact. Which is the most likely effect of this increase in rotation?\nAnswer: Planetary days will become shorter."}
{"text": "Question: The end result in the process of photosynthesis is the production of sugar and oxygen. Which step signals the beginning of photosynthesis?\nAnswer: Chlorophyll in the leaf captures light energy."}
Run the experiment¶
pretrain-experiments config/OLMo-3-1025-7B-midtrain.yaml
This will:
Download the checkpoint
Insert the texts into the training data
Train for 100 steps
Evaluate the result
CLI overrides¶
Any config parameter can be overridden from the command line using dot notation:
pretrain-experiments config/OLMo-3-1025-7B-midtrain.yaml --training.num_steps 50
Next steps¶
Concepts — understand the core abstractions
Configuration Reference — full config reference
Data Insertion — all insertion types and modes
Evaluation — evaluation scripts and benchmarks
See the
config/directory for more examples