p04798526
/
LLaMA-Factory-310P3
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

fix: fix bf16

This commit is contained in:
wql 2024-09-05 15:49:32 +08:00
parent e754b62ccd
commit f15e37dfad
2 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
prepare_yaml_file.py
View File

@ -28,7 +28,7 @@ def main():
model_name_or_path = "../../models/llama-2-7b-ms" model_name_or_path = "../../models/llama-2-7b-ms"
template = "llama2" template = "llama2"
elif model == "Qwen-7B": elif model == "Qwen-7B":
model_name_or_path = "/root/models/Qwen-7B" model_name_or_path = "../../models/Qwen-7B"
template = "qwen" template = "qwen"
else: else:
print("ERROR: model not supported.") print("ERROR: model not supported.")

2
results/lora_sft_template.yaml
View File

@ -29,7 +29,7 @@ learning_rate: 1.0e-4
num_train_epochs: 10.0 num_train_epochs: 10.0
lr_scheduler_type: cosine lr_scheduler_type: cosine
warmup_ratio: 0.1 warmup_ratio: 0.1
bf16: true fp16: true
ddp_timeout: 180000000 ddp_timeout: 180000000
max_steps: 500 max_steps: 500
include_num_input_tokens_seen: true include_num_input_tokens_seen: true