Asap-zzhou's group workspace
Group: sft-reproduction
Name
0 visualized
State
Notes
User
Tags
Created
Runtime
Sweep
_attn_implementation_autoset
_name_or_path
accelerator_config.dispatch_batches
accelerator_config.even_batches
accelerator_config.non_blocking
accelerator_config.split_batches
accelerator_config.use_seedable_sampler
activation_type
adafactor
adam_beta1
adam_beta2
adam_epsilon
add_cross_attention
alibi
alibi_bias_max
architectures
attention_bias
attention_dropout
attention_layer_norm
attention_layer_norm_with_affine
attention_probs_dropout_prob
auto_find_batch_size
auto_map.AutoConfig
auto_map.AutoModel
auto_map.AutoModelForCausalLM
auto_map.AutoModelForMaskedLM
average_tokens_across_devices
batch_eval_metrics
bf16
bf16_full_eval
bias
bias_for_layer_norm
block_group_size
block_type
bos_token_id
chunk_size_feed_forward
d_model
dataloader_drop_last
dataloader_num_workers
dataloader_persistent_workers
dataloader_pin_memory
ddp_timeout
debug
decoder_sparse_step
Finished
asap-zzhou
16h 45m 36s
-
-
/mnt/lustrenew/mllm_aligned/shared/models/huggingface/Dream-org/Dream-v0-Base-7B
-
true
false
false
true
-
false
0.9
0.999
1.0000e-8
false
-
-
["DreamModel"]
-
0
-
-
-
false
configuration_dream.DreamConfig
modeling_dream.DreamModel
-
-
true
false
true
false
none
-
-
-
151665
0
-
false
0
false
true
1800
[]
-
Finished
asap-zzhou
7h 33m 14s
-
-
/mnt/lustrenew/mllm_aligned/shared/models/huggingface/GSAI-ML/LLaDA-8B-Base
-
true
false
false
true
silu
false
0.9
0.999
1.0000e-8
false
false
8
["LLaDAModelLM"]
-
0
false
true
-
false
configuration_llada.LLaDAConfig
modeling_llada.LLaDAModelLM
modeling_llada.LLaDAModelLM
-
true
false
true
false
none
false
1
llama
126080
0
4096
false
0
false
true
1800
[]
-
Crashed
asap-zzhou
7h 40m 48s
-
-
/mnt/lustrenew/mllm_aligned/shared/models/huggingface/GSAI-ML/LLaDA-8B-Base
-
true
false
false
true
silu
false
0.9
0.999
1.0000e-8
false
false
8
["LLaDAModelLM"]
-
0
false
true
-
false
configuration_llada.LLaDAConfig
modeling_llada.LLaDAModelLM
modeling_llada.LLaDAModelLM
-
true
false
true
false
none
false
1
llama
126080
0
4096
false
0
false
true
1800
[]
-
1-3
of 3