mirror of
https://github.com/wassname/alignment-handbook.git
synced 2026-06-27 17:29:09 +08:00
Edward Beeching
cbcb3f60fb
* adds configs and readme * cleaning config files * fix typos and removes things from config * updates text to use comparisons rather ablations * fix readme and adds launch script * fix launch script, adds blogpost link * bump release version, added missing dep, fixes configs * updates main readme file
39 lines
906 B
YAML
39 lines
906 B
YAML
# Model arguments
|
|
model_name_or_path: alignment-handbook/zephyr-7b-sft-full
|
|
torch_dtype: null
|
|
|
|
# Data training arguments
|
|
dataset_mixer:
|
|
HuggingFaceH4/ultrafeedback_binarized: 1.0
|
|
dataset_splits:
|
|
- train_prefs
|
|
- test_prefs
|
|
preprocessing_num_workers: 12
|
|
|
|
# Training arguments with sensible defaults
|
|
bf16: true
|
|
beta: 0.01
|
|
loss_type: sigmoid
|
|
do_eval: true
|
|
evaluation_strategy: steps
|
|
eval_steps: 100
|
|
gradient_accumulation_steps: 2
|
|
gradient_checkpointing: true
|
|
gradient_checkpointing_kwargs:
|
|
use_reentrant: False
|
|
hub_model_id: zephyr-7b-align-scan
|
|
hub_model_revision: dpo-beta-0.01
|
|
learning_rate: 5.0e-7
|
|
logging_steps: 10
|
|
lr_scheduler_type: cosine
|
|
max_prompt_length: 512
|
|
num_train_epochs: 1
|
|
optim: adamw_torch
|
|
output_dir: data/zephyr-7b-align-scan-dpo-beta-0.01
|
|
per_device_train_batch_size: 8
|
|
per_device_eval_batch_size: 8
|
|
save_strategy: "steps"
|
|
save_steps: 100
|
|
save_total_limit: 1
|
|
seed: 42
|
|
warmup_ratio: 0.1 |