ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[misc] fix packing and eval plot (#7623)

Browse Source
This commit is contained in:
hoshi-hiyouga
2025-04-07 18:20:57 +08:00
committed by GitHub
parent 5115dc8c7f
commit c3c0efbaa0
70 changed files with 288 additions and 194 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
examples/train_full/llama3_full_sft.yaml
View File

@@ -24,6 +24,7 @@ save_steps: 500
plot_loss: true
overwrite_output_dir: true
save_only_model: false
report_to: none # choices: [none, wandb, tensorboard, swanlab, mlflow]
### train
per_device_train_batch_size: 1

9
examples/train_full/qwen2vl_full_sft.yaml
View File

@@ -8,10 +8,10 @@ trust_remote_code: true
stage: sft
do_train: true
finetuning_type: full
freeze_vision_tower: true # choices: [true, false]
freeze_multi_modal_projector: true # choices: [true, false]
freeze_language_model: false # choices: [true, false]
deepspeed: examples/deepspeed/ds_z3_config.json # choices: [ds_z0_config.json, ds_z2_config.json, ds_z3_config.json]
freeze_vision_tower: true
freeze_multi_modal_projector: true
freeze_language_model: false
deepspeed: examples/deepspeed/ds_z3_config.json
### dataset
dataset: mllm_demo,identity,alpaca_en_demo
@@ -29,6 +29,7 @@ save_steps: 500
plot_loss: true
overwrite_output_dir: true
save_only_model: false
report_to: none # choices: [none, wandb, tensorboard, swanlab, mlflow]
### train
per_device_train_batch_size: 1