--caption_column to be path error

[kjosh925]

I am gettng following error while training Hunyuan or LTX-Video models

Also, how can one differentiate a training between t2v and i2v during training via bash script provided?

Kindly advice.

[a-r-r-o-w]

Could you share how you're launching the training?

[kjosh925]

I am using the following script provided in the docs:

`#!/bin/bash

export WANDB_MODE="offline"
export NCCL_P2P_DISABLE=1
export TORCH_NCCL_ENABLE_MONITORING=0
export FINETRAINERS_LOG_LEVEL=DEBUG

GPU_IDS="0,1"

DATA_ROOT="/path/to/dataset"
CAPTION_COLUMN="prompts.txt"
VIDEO_COLUMN="videos.txt"
OUTPUT_DIR="/path/to/models/hunyuan-video/"

ID_TOKEN="afkx"

Model arguments

model_cmd="--model_name hunyuan_video
--pretrained_model_name_or_path hunyuanvideo-community/HunyuanVideo"

Dataset arguments

dataset_cmd="--data_root $DATA_ROOT
--video_column $VIDEO_COLUMN
--caption_column $CAPTION_COLUMN
--id_token $ID_TOKEN
--video_resolution_buckets 17x512x768 49x512x768 61x512x768
--caption_dropout_p 0.05"

Dataloader arguments

dataloader_cmd="--dataloader_num_workers 0"

Diffusion arguments

diffusion_cmd=""

Training arguments

training_cmd="--training_type lora
--seed 42
--batch_size 1
--train_steps 500
--rank 128
--lora_alpha 128
--target_modules to_q to_k to_v to_out.0
--gradient_accumulation_steps 1
--gradient_checkpointing
--checkpointing_steps 500
--checkpointing_limit 2
--enable_slicing
--enable_tiling"

Optimizer arguments

optimizer_cmd="--optimizer adamw
--lr 2e-5
--lr_scheduler constant_with_warmup
--lr_warmup_steps 100
--lr_num_cycles 1
--beta1 0.9
--beta2 0.95
--weight_decay 1e-4
--epsilon 1e-8
--max_grad_norm 1.0"

Miscellaneous arguments

miscellaneous_cmd="--tracker_name finetrainers-hunyuan-video
--output_dir $OUTPUT_DIR
--nccl_timeout 1800
--report_to wandb"

cmd="accelerate launch --config_file accelerate_configs/uncompiled_8.yaml --gpu_ids $GPU_IDS train.py
$model_cmd
$dataset_cmd
$dataloader_cmd
$diffusion_cmd
$training_cmd
$optimizer_cmd
$miscellaneous_cmd"

echo "Running command: $cmd"
eval $cmd
echo -ne "-------------------- Finished executing script --------------------\n\n"`

I am running the script above and getting the --caption_column error

[dorpxam]

I think you use latest version with legacy script. In this case, you need to set a path to your dataset file (--dataset_file) as well the --data_root. But the best way is to migrate to the new scripts version, see the example folder.

[a-r-r-o-w]

Hi, it seems like you're using the examples from version v0.0.1 (https://github.com/a-r-r-o-w/finetrainers/tree/v0.0.1) with the current main branch. These are incompatible. For using the main branch, please follow this:

• https://github.com/a-r-r-o-w/finetrainers/tree/main/examples/training/sft

otherwise checkout to the specific tag for the release.

The latest examples are unlikely to change across multiple future versions so you can expect stable support going forward

[kjosh925]

I am already checked out at v0.0.1 branch while running the above mentioned script.

When using main branch, how to start a training on Disney dataset across LTXVideo or Hunyuan models? How can we differentiate the training between image2video and text2video training?

Kindly advice.