Andrewhinh's workspace
Runs
6
Name
2 visualized
State
Notes
User
Tags
Created
Runtime
Sweep
_attn_implementation_autoset
_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_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.dataset
data_args.dataset_dir
data_args.ignore_pad_token_for_loss
data_args.mask_history
data_args.max_samples
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
ddp_timeout
debug
Finished
-
andrewhinh
15m 37s
-
true
Qwen/Qwen2.5-VL-3B-Instruct
true
false
false
true
false
0.9
0.999
1.0000e-8
false
["Qwen2_5_VLForConditionalGeneration"]
0
false
false
false
true
false
151643
0
16384
32768
["sft"]
data
true
false
1000
data
concat
false
true
false
1000
16
false
qwen2_vl
false
0.1
false
0
false
true
180000000
[]
Finished
-
andrewhinh
7m 30s
-
true
Qwen/Qwen2.5-VL-3B-Instruct
true
false
false
true
false
0.9
0.999
1.0000e-8
false
["Qwen2_5_VLForConditionalGeneration"]
0
false
false
false
true
false
151643
0
16384
32768
["sft"]
data
true
false
1000
data
concat
false
true
false
1000
16
false
qwen2_vl
false
0.1
false
0
false
true
180000000
[]
Finished
-
andrewhinh
27m 41s
-
true
Qwen/Qwen2.5-VL-3B-Instruct
true
false
false
true
false
0.9
0.999
1.0000e-8
false
["Qwen2_5_VLForConditionalGeneration"]
0
false
false
false
true
false
151643
0
16384
32768
["sft"]
data
true
false
1000
data
concat
false
true
false
1000
16
false
qwen2_vl
false
0.1
false
0
false
true
180000000
[]
Crashed
-
andrewhinh
41m 1s
-
true
Qwen/Qwen2.5-VL-3B-Instruct
true
false
false
true
false
0.9
0.999
1.0000e-8
false
["Qwen2_5_VLForConditionalGeneration"]
0
false
false
false
true
false
151643
0
16384
32768
["sft"]
data
true
false
1000
data
concat
false
true
false
1000
16
false
qwen2_vl
false
0.1
false
0
false
true
180000000
[]
Finished
-
andrewhinh
5m 43s
-
true
Qwen/Qwen2.5-VL-3B-Instruct
true
false
false
true
false
0.9
0.999
1.0000e-8
false
["Qwen2_5_VLForConditionalGeneration"]
0
false
false
false
true
false
151643
0
16384
32768
["sft"]
data
true
false
1000
data
concat
false
true
false
1000
16
false
qwen2_vl
false
0.1
false
0
false
true
180000000
[]
Finished
-
andrewhinh
21m 46s
-
true
Qwen/Qwen2.5-VL-3B-Instruct
true
false
false
true
false
0.9
0.999
1.0000e-8
false
["Qwen2_5_VLForConditionalGeneration"]
0
false
false
false
true
false
151643
0
16384
32768
["sft"]
data
true
false
1000
data
concat
false
true
false
1000
16
false
qwen2_vl
false
0.1
false
0
false
true
180000000
[]
1-6
of 6