Jiakai's workspace
Runs
1
State
Notes
User
Tags
Created
Runtime
Sweep
_name_or_path
accelerator_config.even_batches
accelerator_config.non_blocking
accelerator_config.split_batches
accelerator_config.use_seedable_sampler
adafactor
adam_beta1
adam_beta2
adam_epsilon
add_cross_attention
architectures
attention_bias
attention_dropout
auto_find_batch_size
average_tokens_across_devices
batch_eval_metrics
bf16
bf16_full_eval
bos_token_id
chunk_size_feed_forward
data_args.buffer_size
data_args.cutoff_len
data_args.data_shared_file_system
data_args.dataset
data_args.dataset_dir
data_args.enable_thinking
data_args.eval_on_each_dataset
data_args.ignore_pad_token_for_loss
data_args.mask_history
data_args.media_dir
data_args.mix_strategy
data_args.neat_packing
data_args.overwrite_cache
data_args.packing
data_args.preprocessing_batch_size
data_args.preprocessing_num_workers
data_args.streaming
data_args.template
data_args.train_on_prompt
data_args.val_size
dataloader_drop_last
dataloader_num_workers
dataloader_persistent_workers
dataloader_pin_memory
Finished
-
jiakai
5h 13m 52s
-
/root/autodl-tmp/modelscope/models/Qwen/Qwen3-14B
true
false
false
true
false
0.9
0.999
1.0000e-8
false
["Qwen3ForCausalLM"]
false
0
false
false
false
true
false
151643
0
16384
2048
false
["ai-expert-alpaca-25-8-8"]
data
false
false
true
false
data
concat
false
true
false
1000
16
false
qwen3
false
0
false
4
false
true
1-1
of 1