Upload qwen2vl_lora_sft.yaml with huggingface_hub

qwen2vl_lora_sft.yaml

@@ -1,43 +1,13 @@
+### Note: DO NOT use quantized model or quantization_bit when merging lora adapters
+
 ### model
 model_name_or_path: /home/zbz5349/WorkSpace/aigeeks/Qwen2.5-VL/ckpt
-image_max_pixels: 262144
-video_max_pixels: 16384
-trust_remote_code: true
-
-### method
-stage: sft
-do_train: true
-finetuning_type: lora
-lora_rank: 8
-lora_target: all
-
-### dataset
-dataset: Percption
+adapter_name_or_path: /home/zbz5349/WorkSpace/aigeeks/Qwen2.5-VL/LLaMA-Factory/results/Qwen2.5-VL-3B_all
 template: qwen2_vl
-cutoff_len: 4096
-max_samples: 1000
-overwrite_cache: true
-preprocessing_num_workers: 16
-
-### output
-output_dir: results/Qwen2.5-VL-3B_all
-logging_steps: 10
-save_steps: 500
-plot_loss: true
-overwrite_output_dir: true
-
-### train
-per_device_train_batch_size: 1
-gradient_accumulation_steps: 8
-learning_rate: 1.0e-4
-num_train_epochs: 3.0
-lr_scheduler_type: cosine
-warmup_ratio: 0.1
-bf16: true
-ddp_timeout: 180000000
+trust_remote_code: true
 
-### eval
-# val_size: 0.1
-# per_device_eval_batch_size: 1
-# eval_strategy: steps
-# eval_steps: 500
+### export
+export_dir: output/Qwen2.5-VL-3B_all
+export_size: 5
+export_device: cpu
+export_legacy_format: false