Keisuke-kamata's workspace
Runs
4
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
activation_offloading
adafactor
adam_beta1
adam_beta2
adam_epsilon
add_cross_attention
architectures
assistant_only_loss
auto_find_batch_size
average_tokens_across_devices
batch_eval_metrics
bf16
bf16_full_eval
block_auto_adjust_ff_dim
block_dim
block_ff_dim
block_ffn_dim_multiplier
block_mlp_init_scale
block_multiple_of
block_norm_eps
block_out_init_scale
block_use_swiglu
block_use_xavier_init
bos_token_id
chunk_size_feed_forward
conv_L_cache
conv_bias
conv_dim
conv_dim_out
conv_use_xavier_init
data_seed
dataloader_drop_last
dataloader_num_workers
dataloader_persistent_workers
dataloader_pin_memory
dataset_num_proc
dataset_text_field
ddp_timeout
debug
Finished
yuya-yamamoto
periodic-evaluation
weave-integration
33m 7s
-
unsloth/LFM2-350M-unsloth-bnb-4bit
true
false
false
true
false
false
0.9
0.999
1.0000e-8
false
["Lfm2ForCausalLM"]
false
false
false
false
true
true
true
1024
6656
1
1
256
0.00001
1
true
true
1
0
3
false
1024
1024
true
3407
false
0
false
true
16
text
1800
[]
Finished
yuya-yamamoto
periodic-evaluation
weave-integration
10m 59s
-
unsloth/LFM2-350M-unsloth-bnb-4bit
true
false
false
true
false
false
0.9
0.999
1.0000e-8
false
["Lfm2ForCausalLM"]
false
false
false
false
true
true
true
1024
6656
1
1
256
0.00001
1
true
true
1
0
3
false
1024
1024
true
3407
false
0
false
true
16
text
1800
[]
Crashed
yuya-yamamoto
periodic-evaluation
weave-integration
20m 15s
-
unsloth/LFM2-350M-unsloth-bnb-4bit
true
false
false
true
false
false
0.9
0.999
1.0000e-8
false
["Lfm2ForCausalLM"]
false
false
false
false
true
true
true
1024
6656
1
1
256
0.00001
1
true
true
1
0
3
false
1024
1024
true
3407
false
0
false
true
16
text
1800
[]
Crashed
yuya-yamamoto
periodic-evaluation
weave-integration
16m 31s
-
unsloth/LFM2-350M-unsloth-bnb-4bit
true
false
false
true
false
false
0.9
0.999
1.0000e-8
false
["Lfm2ForCausalLM"]
false
false
false
false
true
true
true
1024
6656
1
1
256
0.00001
1
true
true
1
0
3
false
1024
1024
true
3407
false
0
false
true
16
text
1800
[]
1-4
of 4