Aryanvs's workspace
Runs
1
Name
1 visualized
State
Notes
User
Tags
Created
Runtime
Sweep
dataloader_arguments.dataloader_num_workers
dataloader_arguments.pin_memory
dataset_arguments.dataset_config
dataset_arguments.dataset_shuffle_buffer_size
dataset_arguments.precomputation_dir
dataset_arguments.precomputation_items
dataset_arguments.precomputation_once
miscellaneous_arguments.allow_tf32
miscellaneous_arguments.init_timeout
miscellaneous_arguments.logging_dir
miscellaneous_arguments.logging_steps
miscellaneous_arguments.nccl_timeout
miscellaneous_arguments.output_dir
miscellaneous_arguments.push_to_hub
miscellaneous_arguments.report_to
miscellaneous_arguments.tracker_name
miscellaneous_arguments.verbose
model_arguments.layerwise_upcasting_modules
model_arguments.layerwise_upcasting_skip_modules_pattern
model_arguments.layerwise_upcasting_storage_dtype
model_arguments.model_name
model_arguments.pretrained_model_name_or_path
model_arguments.text_encoder_2_dtype
model_arguments.text_encoder_3_dtype
model_arguments.text_encoder_dtype
model_arguments.transformer_dtype
model_arguments.vae_dtype
optimizer_arguments.beta1
optimizer_arguments.beta2
optimizer_arguments.epsilon
optimizer_arguments.lr
optimizer_arguments.lr_num_cycles
optimizer_arguments.lr_power
optimizer_arguments.lr_scheduler
optimizer_arguments.lr_warmup_steps
optimizer_arguments.max_grad_norm
optimizer_arguments.optimizer
optimizer_arguments.weight_decay
parallel_arguments.cp_degree
parallel_arguments.dp_degree
parallel_arguments.dp_shards
parallel_arguments.pp_degree
parallel_arguments.tp_degree
training_arguments.batch_size
Finished
-
aryanvs
7h 22m 19s
-
0
false
examples/training/sft/cogvideox/crush_smol_lora/training.json
10
/raid/aryan/cogvideox/precomputed
25
true
false
600
logs
1
600
/raid/aryan/cogvideox
false
wandb
finetrainers-cogvideox
0
[]
["patch_embed","pos_embed","x_embedder","context_embedder","^proj_in$","^proj_out$","norm"]
torch.float8_e4m3fn
cogvideox
THUDM/CogVideoX1.5-5B
torch.bfloat16
torch.bfloat16
torch.bfloat16
torch.bfloat16
torch.bfloat16
0.9
0.99
1.0000e-8
0.00005
1
1
constant_with_warmup
1000
1
adamw
0.0001
1
2
1
1
1
1
1-1
of 1