423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-14 19:06:26 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix #4005 #4013

Browse Source 
Former-commit-id: eed33862bc
This commit is contained in:
hiyouga
2024-06-03 19:12:29 +08:00
parent eaab09fccb
commit e4ce59243b
24 changed files with 52 additions and 52 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
examples/lora_single_gpu/llama3_lora_dpo.yaml
View File

@@ -27,10 +27,10 @@ overwrite_output_dir: true
### train
per_device_train_batch_size: 1
gradient_accumulation_steps: 8
learning_rate: 0.000005
learning_rate: 5.0e-6
num_train_epochs: 3.0
lr_scheduler_type: cosine
warmup_steps: 0.1
warmup_ratio: 0.1
fp16: true
### eval

4
examples/lora_single_gpu/llama3_lora_kto.yaml
View File

@@ -25,10 +25,10 @@ overwrite_output_dir: true
### train
per_device_train_batch_size: 1
gradient_accumulation_steps: 8
learning_rate: 0.000005
learning_rate: 5.0e-6
num_train_epochs: 3.0
lr_scheduler_type: cosine
warmup_steps: 0.1
warmup_ratio: 0.1
fp16: true
### eval

4
examples/lora_single_gpu/llama3_lora_ppo.yaml
View File

@@ -26,10 +26,10 @@ overwrite_output_dir: true
### train
per_device_train_batch_size: 1
gradient_accumulation_steps: 8
learning_rate: 0.00001
learning_rate: 1.0e-5
num_train_epochs: 3.0
lr_scheduler_type: cosine
warmup_steps: 0.1
warmup_ratio: 0.1
fp16: true
### generate

4
examples/lora_single_gpu/llama3_lora_pretrain.yaml
View File

@@ -24,10 +24,10 @@ overwrite_output_dir: true
### train
per_device_train_batch_size: 1
gradient_accumulation_steps: 8
learning_rate: 0.0001
learning_rate: 1.0e-4
num_train_epochs: 3.0
lr_scheduler_type: cosine
warmup_steps: 0.1
warmup_ratio: 0.1
fp16: true
### eval

4
examples/lora_single_gpu/llama3_lora_reward.yaml
View File

@@ -25,10 +25,10 @@ overwrite_output_dir: true
### train
per_device_train_batch_size: 1
gradient_accumulation_steps: 8
learning_rate: 0.00001
learning_rate: 1.0e-5
num_train_epochs: 3.0
lr_scheduler_type: cosine
warmup_steps: 0.1
warmup_ratio: 0.1
fp16: true
### eval

4
examples/lora_single_gpu/llama3_lora_sft.yaml
View File

@@ -25,10 +25,10 @@ overwrite_output_dir: true
### train
per_device_train_batch_size: 1
gradient_accumulation_steps: 8
learning_rate: 0.0001
learning_rate: 1.0e-4
num_train_epochs: 3.0
lr_scheduler_type: cosine
warmup_steps: 0.1
warmup_ratio: 0.1
fp16: true
### eval

4
examples/lora_single_gpu/llava1_5_lora_sft.yaml
View File

@@ -26,10 +26,10 @@ overwrite_output_dir: true
### train
per_device_train_batch_size: 1
gradient_accumulation_steps: 8
learning_rate: 0.0001
learning_rate: 1.0e-4
num_train_epochs: 3.0
lr_scheduler_type: cosine
warmup_steps: 0.1
warmup_ratio: 0.1
fp16: true
### eval