Merge pull request #264 from LLaVA-VL/Luodian-patch-1

[script] add scripts for users to better understand training process.

Author: Luodian

scripts/train/finetune_clip.sh scripts/train/direct_finetune_clip.shscripts/train/finetune_clip.sh renamed to scripts/train/direct_finetune_clip.sh

@@ -62,4 +62,4 @@ ACCELERATE_CPU_AFFINITY=1 torchrun --nproc_per_node="${NUM_GPUS}" --nnodes="${NN
     --dataloader_drop_last True \
     --attn_implementation sdpa
 
-# You can delete the sdpa attn_implementation if you want to use flash attn
+# You can delete the sdpa attn_implementation if you want to use flash attn

scripts/train/finetune_siglip_a4.sh …ripts/train/direct_finetune_siglip_a4.shscripts/train/finetune_siglip_a4.sh renamed to scripts/train/direct_finetune_siglip_a4.sh scripts/train/finetune_siglip_a4.sh renamed to scripts/train/direct_finetune_siglip_a4.sh

@@ -64,4 +64,4 @@ ACCELERATE_CPU_AFFINITY=1 torchrun --nproc_per_node="${NUM_GPUS}" --nnodes="${NN
     --dataloader_drop_last True \
     --attn_implementation sdpa
 
-# You can delete the sdpa attn_implementation if you want to use flash attn
+# You can delete the sdpa attn_implementation if you want to use flash attn

scripts/train/finetune_si.sh

@@ -30,7 +30,7 @@ ACCELERATE_CPU_AFFINITY=1 torchrun --nproc_per_node="${NUM_GPUS}" --nnodes="${NN
     --deepspeed scripts/zero3.json \
     --model_name_or_path $PREV_STAGE_CHECKPOINT \
     --version $PROMPT_VERSION \
-    --data_path /mnt/bn/vl-research/workspace/boli01/projects/LLaVA_Next/scripts/i18n/scale_llms/next_ov_stage_july21.yaml \
+    --data_path /mnt/bn/vl-research/workspace/boli01/projects/LLaVA_Next/scripts/i18n/scale_llms/next_3p2m_single_image.yaml \
     --image_folder /mnt/bn/vl-research/data/llava_data \
     --video_folder /mnt/bn/vl-research/data/llava_video \
     --mm_tunable_parts="mm_vision_tower,mm_mlp_adapter,mm_language_model" \