nodes/hao-ai-lab-FastVideo
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

update

Browse Source
This commit is contained in:
rlsu9
2024-12-01 21:56:36 +04:00
parent 892324ae59
commit ebfbf6046a
9 changed files with 26 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+2 -2
View File
@@ -60,9 +60,9 @@ Scripts are located at scripts/experiment_N.sh
10. pcm_linear_quadratic， euler_steps 50, 0.05, batchsize 32
11. pcm_linear_quadratic， euler_steps 50, learning rate,1e-7
12. shift1, euler_steps 50
13. pcm_linear_quadratic， euler_steps 50, 0.1, linear_range 0.75 ema 0.98, decay 0.1
13. pcm_linear_quadratic， euler_steps 50, 0.1, linear_range 0.75 ema 0.98
14. 4.5 cfg, validation no cfg, pcm_linear_quadratic， euler_steps 50, 0.1, linear_range 0.75
15. pcm_linear_quadratic， euler_steps 50, 0.15, linear_range 0.75
16. 8pcm_linear_quadratic， euler_steps 50, 0.2, linear_range 0.75
16. pcm_linear_quadratic， euler_steps 50, 0.1, linear_range 0.75 ema 0.95, decay 0.0
17. no cfg, validation no cfg, pcm_linear_quadratic， euler_steps 50, 0.1, linear_range 0.75
18. shift16, euler_steps 50
fastvideo/train.py
+2 -1
View File
@@ -227,7 +227,7 @@ def main(args):
params_to_optimize,
lr=args.learning_rate,
betas=(0.9,0.999),
weight_decay=0.01,
weight_decay=args.weight_decay,
eps=1e-8,
)
@@ -477,5 +477,6 @@ if __name__ == "__main__":
)
parser.add_argument("--lr_num_cycles", type=int, default=1, help="Number of cycles in the learning rate scheduler.")
parser.add_argument("--lr_power", type=float, default=1.0, help="Power factor of the polynomial scheduler.",)
parser.add_argument("--weight_decay", type=float, default=0.01, help="Weight decay to apply.")
args = parser.parse_args()
main(args)
scripts/experiment_12.sh
+1 -1
View File
@@ -36,7 +36,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--cfg 0.0\
--ema_decay 0.999\
--log_validation\
--output_dir="data/outputs/shift1_euler_50"\
--output_dir="$DATA_DIR/outputs/shift1_euler_50"\
--tracker_project_name PCM \
--num_frames 163 \
--shift 1 \
scripts/experiment_13.sh
+2 -2
View File
@@ -17,7 +17,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--rdzv_endpoint=$IP:29500 \
fastvideo/distill.py\
--seed 42\
--pretrained_model_name_or_path data/mochi\
--pretrained_model_name_or_path $DATA_DIR/mochi\
--cache_dir "data/.cache"\
--data_json_path "$DATA_DIR/Merge-30k-Data/video2caption.json"\
--validation_prompt_dir "$DATA_DIR/validation_embeddings/validation_prompt_embed_mask"\
@@ -39,7 +39,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--ema_start_step 0\
--cfg 0.0\
--log_validation\
--output_dir="data/outputs/lq_euler_50_thres0.1_lrg_0.75_ema_0.98"\
--output_dir="$DATA_DIR/outputs/lq_euler_50_thres0.1_lrg_0.75_ema_0.98"\
--tracker_project_name PCM \
--num_frames 163 \
--scheduler_type pcm_linear_quadratic \
scripts/experiment_14.sh
+2 -2
View File
@@ -17,7 +17,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--rdzv_endpoint=$IP:29500 \
fastvideo/distill.py\
--seed 42\
--pretrained_model_name_or_path data/mochi\
--pretrained_model_name_or_path $DATA_DIR/mochi\
--cache_dir "data/.cache"\
--data_json_path "$DATA_DIR/Merge-30k-Data/video2caption.json"\
--validation_prompt_dir "$DATA_DIR/validation_embeddings/validation_prompt_embed_mask"\
@@ -40,7 +40,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--cfg 0.0\
--ema_decay 0.999\
--log_validation\
--output_dir="data/outputs/lq_euler_50_thres0.1_lrg_0.75_cfg_4.5"\
--output_dir="$DATA_DIR/outputs/lq_euler_50_thres0.1_lrg_0.75_cfg_4.5"\
--tracker_project_name PCM \
--num_frames 163 \
--scheduler_type pcm_linear_quadratic \
scripts/experiment_15.sh
+2 -2
View File
@@ -17,7 +17,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--rdzv_endpoint=$IP:29500 \
fastvideo/distill.py\
--seed 42\
--pretrained_model_name_or_path data/mochi\
--pretrained_model_name_or_path $DATA_DIR/mochi\
--cache_dir "data/.cache"\
--data_json_path "$DATA_DIR/Merge-30k-Data/video2caption.json"\
--validation_prompt_dir "$DATA_DIR/validation_embeddings/validation_prompt_embed_mask"\
@@ -40,7 +40,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--cfg 0.0\
--ema_decay 0.999\
--log_validation\
--output_dir="data/outputs/lq_euler_50_thres0.15_lrg_0.75"\
--output_dir="$DATA_DIR/outputs/lq_euler_50_thres0.15_lrg_0.75"\
--tracker_project_name PCM \
--num_frames 163 \
--scheduler_type pcm_linear_quadratic \
scripts/experiment_16.sh
+8 -5
View File
@@ -17,7 +17,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--rdzv_endpoint=$IP:29500 \
fastvideo/distill.py\
--seed 42\
--pretrained_model_name_or_path data/mochi\
--pretrained_model_name_or_path $DATA_DIR/mochi\
--cache_dir "data/.cache"\
--data_json_path "$DATA_DIR/Merge-30k-Data/video2caption.json"\
--validation_prompt_dir "$DATA_DIR/validation_embeddings/validation_prompt_embed_mask"\
@@ -38,14 +38,17 @@ torchrun --nnodes 2 --nproc_per_node 8\
--allow_tf32\
--ema_start_step 0\
--cfg 0.0\
--ema_decay 0.999\
--log_validation\
--output_dir="data/outputs/lq_euler_50_thres0.15_lrg_0.75"\
--output_dir="$DATA_DIR/outputs/lq_euler_50_thres0.1_lrg_0.75_ema_0.95_decay0.0"\
--tracker_project_name PCM \
--num_frames 163 \
--scheduler_type pcm_linear_quadratic \
--validation_guidance_scale "0.5,1.5,2.5,4.5" \
--num_euler_timesteps 50 \
--linear_quadratic_threshold 0.15 \
--linear_range 0.75
--linear_quadratic_threshold 0.1 \
--linear_range 0.75 \
--use_ema \
--ema_decay 0.95 \
--weight_decay 0.0
scripts/experiment_17.sh
+5 -4
View File
@@ -17,7 +17,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--rdzv_endpoint=$IP:29500 \
fastvideo/distill.py\
--seed 42\
--pretrained_model_name_or_path data/mochi\
--pretrained_model_name_or_path $DATA_DIR/mochi\
--cache_dir "data/.cache"\
--data_json_path "$DATA_DIR/Merge-30k-Data/video2caption.json"\
--validation_prompt_dir "$DATA_DIR/validation_embeddings/validation_prompt_embed_mask"\
@@ -40,12 +40,13 @@ torchrun --nnodes 2 --nproc_per_node 8\
--cfg 0.0\
--ema_decay 0.999\
--log_validation\
--output_dir="data/outputs/lq_euler_50_thres0.2_lrg_0.75"\
--output_dir="$DATA_DIR/outputs/lq_euler_50_thres0.1_lrg_0.75_cfg0.0"\
--tracker_project_name PCM \
--num_frames 163 \
--scheduler_type pcm_linear_quadratic \
--validation_guidance_scale "0.5,1.5,2.5,4.5" \
--num_euler_timesteps 50 \
--linear_quadratic_threshold 0.2 \
--linear_range 0.75
--linear_quadratic_threshold 0.1 \
--linear_range 0.75 \
--not_apply_cfg_solver
scripts/experiment_18.sh
+2 -2
View File
@@ -17,7 +17,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--rdzv_endpoint=$IP:29500 \
fastvideo/distill.py\
--seed 42\
--pretrained_model_name_or_path data/mochi\
--pretrained_model_name_or_path $DATA_DIR/mochi\
--cache_dir "data/.cache"\
--data_json_path "$DATA_DIR/Merge-30k-Data/video2caption.json"\
--validation_prompt_dir "$DATA_DIR/validation_embeddings/validation_prompt_embed_mask"\
@@ -40,7 +40,7 @@ torchrun --nnodes 2 --nproc_per_node 8\
--cfg 0.0\
--ema_decay 0.999\
--log_validation\
--output_dir="data/outputs/shift16_euler_50"\
--output_dir="$DATA_DIR/outputs/shift16_euler_50"\
--tracker_project_name PCM \
--num_frames 163 \
--shift 16 \