update examples
Former-commit-id: 3b5f138155d96b346bda18e465cf60ec7d99e19c
This commit is contained in:
@@ -1,7 +1,7 @@
|
||||
# model
|
||||
### model
|
||||
model_name_or_path: models/llama3-8b-instruct-pro
|
||||
|
||||
# method
|
||||
### method
|
||||
stage: sft
|
||||
do_train: true
|
||||
finetuning_type: freeze
|
||||
@@ -9,7 +9,7 @@ freeze_trainable_layers: 8
|
||||
freeze_trainable_modules: all
|
||||
use_llama_pro: true
|
||||
|
||||
# dataset
|
||||
### dataset
|
||||
dataset: identity,alpaca_gpt4_en
|
||||
template: llama3
|
||||
cutoff_len: 1024
|
||||
@@ -17,14 +17,14 @@ max_samples: 1000
|
||||
overwrite_cache: true
|
||||
preprocessing_num_workers: 16
|
||||
|
||||
# output
|
||||
### output
|
||||
output_dir: saves/llama3-8b-instruct-pro/freeze/sft
|
||||
logging_steps: 10
|
||||
save_steps: 500
|
||||
plot_loss: true
|
||||
overwrite_output_dir: true
|
||||
|
||||
# train
|
||||
### train
|
||||
per_device_train_batch_size: 1
|
||||
gradient_accumulation_steps: 8
|
||||
learning_rate: 0.0001
|
||||
@@ -33,7 +33,7 @@ lr_scheduler_type: cosine
|
||||
warmup_steps: 0.1
|
||||
fp16: true
|
||||
|
||||
# eval
|
||||
### eval
|
||||
val_size: 0.1
|
||||
per_device_eval_batch_size: 1
|
||||
evaluation_strategy: steps
|
||||
|
||||
Reference in New Issue
Block a user