Lightweight on/off-policy distillation framework with a single client interface.
If on_policy: false, a complete pipeline for generating distillation data will be set up. If on_policy: true, a complete pipeline for online training job will be set up.
spider takes care the whole workflow from dataset preparation, rollouts, kl supervision (on-policy only), and data post-processing in a few lines of code.
# install & launch a server
pip install -e ".[server]"
uvicorn server.app:app --host 0.0.0.0 --port 9000 --workers 1
# install a client
pip install -e ".[client]" (available for cpu machines)The following snippet showcases a complete job cycle for a distillation job.
from spider.client import SpiderClient
from spider.config import AppConfig
def filter_rows(records): # define custom data filtering logic
return records
config = AppConfig.load("config/test-remote-processor.yaml") # define rollout hyperparams
with SpiderClient(config=config, processor=filter_rows) as client:
submission = client.submit_job()
job_id = submission["job_id"]
status = client.poll_job(job_id, interval=5.0, timeout=600, wait_for_completion=True)
if status["status"] == "completed":
client.download_result(job_id, destionation="./artifacts/result.json") # return full data with metadata, optionally upload to HFThe following is the config file for an off-policy distillation job.
server:
base_url:
api_key:
request_timeout: 120
job:
model:
provider: vllm
name: "Qwen/Qwen2.5-7B-Instruct"
parameters:
tensor_parallel_size: 4
source:
dataset: "RiddleHe/OpenCodeReasoning-2-questions-dedup-34k-sample-1024"
split: "train[0:10]"
field: "question"
generation:
parameters:
temperature: 0.7
top_p: 0.9
max_tokens: 8196
output:
mode: "upload_hf"
hf:
repo_id: collinear-ai/spider-rollouts-qwen2.5-7b-instruct-ocr2-ast-filter
token:
private: falseThe following is the config file for an off-policy distillation job.
server:
base_url:
api_key:
request_timeout: 180
job:
model:
provider: tinker
name: "Qwen/Qwen3-8B"
source:
dataset: "RiddleHe/OpenCodeReasoning-2-questions-dedup-34k-sample-1024"
split: "train[0:8]"
field: "question"
generation:
on_policy: true
on_policy_options:
teacher: "Qwen/Qwen3-30B-A3B"
api_key:
learning_rate: 5e-5
groups_per_batch: 4
group_size: 2
max_tokens: 1024
lora_rank: 16
num_substeps: 1
kl_penalty_coef: 0.1
kl_discount_factor: 0.0
loss_fn: "importance_sampling"
compute_post_kl: true
eval_every: 5
save_every: 20
output:
mode: "return"