423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-08-23 22:32:54 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #6246 from hiyouga/hiyouga/update_examples

Browse Source 
[examples] update examples

Former-commit-id: 967a6c12a72b97104bd262b0f2ed945a8b0cd81d
This commit is contained in:
hoshi-hiyouga 2024-12-05 16:49:30 +08:00 committed by GitHub
commit 7f8c59144e
7 changed files with 13 additions and 52 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
examples/README.md
View File

@ -126,14 +126,14 @@ llamafactory-cli train examples/train_qlora/llama3_lora_sft_aqlm.yaml
#### Supervised Fine-Tuning on Single Node
```bash
FORCE_TORCHRUN=1 llamafactory-cli train examples/train_full/llama3_full_sft_ds3.yaml
FORCE_TORCHRUN=1 llamafactory-cli train examples/train_full/llama3_full_sft.yaml
```
#### Supervised Fine-Tuning on Multiple Nodes
```bash
FORCE_TORCHRUN=1 NNODES=2 RANK=0 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft_ds3.yaml
FORCE_TORCHRUN=1 NNODES=2 RANK=1 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft_ds3.yaml
FORCE_TORCHRUN=1 NNODES=2 RANK=0 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft.yaml
FORCE_TORCHRUN=1 NNODES=2 RANK=1 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft.yaml
```
#### Multimodal Supervised Fine-Tuning

6
examples/README_zh.md
View File

@ -126,14 +126,14 @@ llamafactory-cli train examples/train_qlora/llama3_lora_sft_aqlm.yaml
#### 在单机上进行指令监督微调
```bash
FORCE_TORCHRUN=1 llamafactory-cli train examples/train_full/llama3_full_sft_ds3.yaml
FORCE_TORCHRUN=1 llamafactory-cli train examples/train_full/llama3_full_sft.yaml
```
#### 在多机上进行指令监督微调
```bash
FORCE_TORCHRUN=1 NNODES=2 RANK=0 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft_ds3.yaml
FORCE_TORCHRUN=1 NNODES=2 RANK=1 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft_ds3.yaml
FORCE_TORCHRUN=1 NNODES=2 RANK=0 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft.yaml
FORCE_TORCHRUN=1 NNODES=2 RANK=1 MASTER_ADDR=192.168.0.1 MASTER_PORT=29500 llamafactory-cli train examples/train_full/llama3_full_sft.yaml
```
#### 多模态指令监督微调

2
examples/train_full/llama3_full_sft_ds3.yaml → examples/train_full/llama3_full_sft.yaml
View File

@ -5,7 +5,7 @@ model_name_or_path: meta-llama/Meta-Llama-3-8B-Instruct
stage: sft
do_train: true
finetuning_type: full
deepspeed: examples/deepspeed/ds_z3_config.json
deepspeed: examples/deepspeed/ds_z3_config.json # choices: [ds_z0_config.json, ds_z2_config.json, ds_z3_config.json]
### dataset
dataset: identity,alpaca_en_demo

7
examples/train_full/qwen2vl_full_sft.yaml
View File

@ -5,10 +5,11 @@ model_name_or_path: Qwen/Qwen2-VL-7B-Instruct
stage: sft
do_train: true
finetuning_type: full
deepspeed: examples/deepspeed/ds_z3_config.json
freeze_vision_tower: true # choices: [true, false]
deepspeed: examples/deepspeed/ds_z3_config.json # choices: [ds_z0_config.json, ds_z2_config.json, ds_z3_config.json]
### dataset
dataset: mllm_demo,identity
dataset: mllm_demo,identity,alpaca_en_demo
template: qwen2_vl
cutoff_len: 2048
max_samples: 1000
@ -26,7 +27,7 @@ overwrite_output_dir: true
per_device_train_batch_size: 1
gradient_accumulation_steps: 2
learning_rate: 1.0e-5
num_train_epochs: 3.0
num_train_epochs: 30.0
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: true

40
examples/train_lora/llama3_lora_sft_ds0.yaml
View File

@ -1,40 +0,0 @@
### model
model_name_or_path: meta-llama/Meta-Llama-3-8B-Instruct
### method
stage: sft
do_train: true
finetuning_type: lora
lora_target: all
deepspeed: examples/deepspeed/ds_z0_config.json
### dataset
dataset: identity,alpaca_en_demo
template: llama3
cutoff_len: 2048
max_samples: 1000
overwrite_cache: true
preprocessing_num_workers: 16
### output
output_dir: saves/llama3-8b/lora/sft
logging_steps: 10
save_steps: 500
plot_loss: true
overwrite_output_dir: true
### train
per_device_train_batch_size: 1
gradient_accumulation_steps: 2
learning_rate: 1.0e-4
num_train_epochs: 3.0
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: true
ddp_timeout: 180000000
### eval
val_size: 0.1
per_device_eval_batch_size: 1
eval_strategy: steps
eval_steps: 500

2
examples/train_lora/llama3_lora_sft_ds3.yaml
View File

@ -6,7 +6,7 @@ stage: sft
do_train: true
finetuning_type: lora
lora_target: all
deepspeed: examples/deepspeed/ds_z3_config.json
deepspeed: examples/deepspeed/ds_z3_config.json # choices: [ds_z0_config.json, ds_z2_config.json, ds_z3_config.json]
### dataset
dataset: identity,alpaca_en_demo

2
examples/train_lora/qwen2vl_lora_sft.yaml
View File

@ -8,7 +8,7 @@ finetuning_type: lora
lora_target: all
### dataset
dataset: mllm_demo,identity # video: mllm_video_demo
dataset: mllm_demo,identity,alpaca_en_demo # video: mllm_video_demo
template: qwen2_vl
cutoff_len: 2048
max_samples: 1000